[ `gemma`] Adds support for Gemma 💎 (#29167)

* inital commit

* update

* update conversion checkpoint

* update conversion script

* nits

* some fixes

* nits

* merge

* fix permute

* nits

* fix

* nits

* nits

* nits

* fix rope

* fix both rope

* nites

* style

* make sure flax works

* fix flax init code

* fix foward

* nits

* print flax generation out

* current code

* nits

* SIIIIIIIIIIIIIIIIIII

* update

* add new tokenizer

* correct fast tokenizer

* fix conversion

* more comments

* fix modeling and conversion

* nits and nits

* nits testing

* add some tokenization tests

* add some edge cases

* add slow tests and fix them

* fixup

* fix copies for modeling

* fix copies

* add 7B slow tests

* fix

* fix

* fix tests

* make tokenizer cis go green

* styling

* last tokenizer nits

* update jax tests

* fix flax for 7b

* add jit testing 🤗



* cleanups

* isolated nit, inv_freq for rotary_emb.inv_freq

* propagate to jax

* Apply suggestions from code review
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>

* adjust test

* fix conversion script

* change name

* correct file names

* update conversion script

* Fix bos and eos token ids in the model configuration (#3)

* update modelling

* update conversion script

* add static cache for gemma

* fix sdpa generate

* fix batched

* multiple fixes

* fix FA2

* final fix

* Rename a few missing strings and filenames (#4)

* merge with upstream main

* fix copies

* fix copies

* fix fixup

* fix fixup

* fix

* fix

* final tests

* fix fx gemma tests

* fix fx bf16/fp16 tests

* update slow fx tests

* fx slow tests: one logits, one generation

* move jit test standalone

* Apply suggestions from code review

* nits

* tokenizer updates

* more tokenization updates: custom GemmaSentencepieceExtrator

* style

* Update src/transformers/cache_utils.py

* Update src/transformers/models/gemma/__init__.py

* Update tests/models/gemma/test_modeling_flax_gemma.py

* small nits

* style

* update tokenization test

* fix the rotary embedding

* with style

* fix slow tests

* WARNING this commit might be very important for precisions

* Update tests/models/gemma/test_modeling_flax_gemma.py

* Update src/transformers/models/gemma/configuration_gemma.py
Co-authored-by: Lysandre Debut <hi@lysand.re>

* Update src/transformers/models/gemma/modeling_flax_gemma.py
Co-authored-by: Lysandre Debut <hi@lysand.re>

* small nits here and there!

* forgotten nit

* remove on the fly computation of inv_freq

* revert previous change, let's be safe and for now re-compute freq cis to make sure it's in float

* Apply suggestions from code review
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update src/transformers/models/gemma/convert_gemma_weights_to_hf.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update src/transformers/models/gemma/convert_gemma_weights_to_hf.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_flax_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_tokenization_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_tokenization_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_tokenization_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_tokenization_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* Update tests/models/gemma/test_modeling_gemma.py
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* nit conversion script link

* fix some tests

* add not doctest and pr doctest

* repo consistency

* fix last CIs 🚀



* update all readmes

---------
Co-authored-by: younesbelkada <younesbelkada@gmail.com>
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>
Co-authored-by: sanchit-gandhi <sanchit@huggingface.co>
Co-authored-by: Lysandre Debut <hi@lysand.re>

src/transformers/models/auto/modeling_auto.py

@@ -103,6 +103,7 @@ MODEL_MAPPING_NAMES = OrderedDict(
         ("focalnet", "FocalNetModel"),
         ("fsmt", "FSMTModel"),
         ("funnel", ("FunnelModel", "FunnelBaseModel")),
+        ("gemma", "GemmaModel"),
         ("git", "GitModel"),
         ("glpn", "GLPNModel"),
         ("gpt-sw3", "GPT2Model"),
@@ -426,6 +427,7 @@ MODEL_FOR_CAUSAL_LM_MAPPING_NAMES = OrderedDict(
         ("ernie", "ErnieForCausalLM"),
         ("falcon", "FalconForCausalLM"),
         ("fuyu", "FuyuForCausalLM"),
+        ("gemma", "GemmaForCausalLM"),
         ("git", "GitForCausalLM"),
         ("gpt-sw3", "GPT2LMHeadModel"),
         ("gpt2", "GPT2LMHeadModel"),
@@ -764,6 +766,7 @@ MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES = OrderedDict(
         ("flaubert", "FlaubertForSequenceClassification"),
         ("fnet", "FNetForSequenceClassification"),
         ("funnel", "FunnelForSequenceClassification"),
+        ("gemma", "GemmaForSequenceClassification"),
         ("gpt-sw3", "GPT2ForSequenceClassification"),
         ("gpt2", "GPT2ForSequenceClassification"),
         ("gpt_bigcode", "GPTBigCodeForSequenceClassification"),

src/transformers/models/auto/modeling_flax_auto.py

@@ -39,6 +39,7 @@ FLAX_MODEL_MAPPING_NAMES = OrderedDict(
         ("clip", "FlaxCLIPModel"),
         ("distilbert", "FlaxDistilBertModel"),
         ("electra", "FlaxElectraModel"),
+        ("gemma", "FlaxGemmaModel"),
         ("gpt-sw3", "FlaxGPT2Model"),
         ("gpt2", "FlaxGPT2Model"),
         ("gpt_neo", "FlaxGPTNeoModel"),
@@ -144,6 +145,7 @@ FLAX_MODEL_FOR_CAUSAL_LM_MAPPING_NAMES = OrderedDict(
         ("big_bird", "FlaxBigBirdForCausalLM"),
         ("bloom", "FlaxBloomForCausalLM"),
         ("electra", "FlaxElectraForCausalLM"),
+        ("gemma", "FlaxGemmaForCausalLM"),
         ("gpt-sw3", "FlaxGPT2LMHeadModel"),
         ("gpt2", "FlaxGPT2LMHeadModel"),
         ("gpt_neo", "FlaxGPTNeoForCausalLM"),

src/transformers/models/auto/tokenization_auto.py

@@ -178,6 +178,13 @@ else:
             ("fnet", ("FNetTokenizer", "FNetTokenizerFast" if is_tokenizers_available() else None)),
             ("fsmt", ("FSMTTokenizer", None)),
             ("funnel", ("FunnelTokenizer", "FunnelTokenizerFast" if is_tokenizers_available() else None)),
+            (
+                "gemma",
+                (
+                    "GemmaTokenizer" if is_sentencepiece_available() else None,
+                    "GemmaTokenizerFast" if is_tokenizers_available() else None,
+                ),
+            ),
             ("git", ("BertTokenizer", "BertTokenizerFast" if is_tokenizers_available() else None)),
             ("gpt-sw3", ("GPTSw3Tokenizer" if is_sentencepiece_available() else None, None)),
             ("gpt2", ("GPT2Tokenizer", "GPT2TokenizerFast" if is_tokenizers_available() else None)),

src/transformers/models/gemma/__init__.py

+# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+from typing import TYPE_CHECKING
+
+from ...utils import (
+    OptionalDependencyNotAvailable,
+    _LazyModule,
+    is_flax_available,
+    is_sentencepiece_available,
+    is_tokenizers_available,
+    is_torch_available,
+)
+
+
+_import_structure = {
+    "configuration_gemma": ["GEMMA_PRETRAINED_CONFIG_ARCHIVE_MAP", "GemmaConfig"],
+}
+
+try:
+    if not is_sentencepiece_available():
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    pass
+else:
+    _import_structure["tokenization_gemma"] = ["GemmaTokenizer"]
+
+try:
+    if not is_tokenizers_available():
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    pass
+else:
+    _import_structure["tokenization_gemma_fast"] = ["GemmaTokenizerFast"]
+
+
+try:
+    if not is_torch_available():
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    pass
+else:
+    _import_structure["modeling_gemma"] = [
+        "GemmaForCausalLM",
+        "GemmaModel",
+        "GemmaPreTrainedModel",
+        "GemmaForSequenceClassification",
+    ]
+
+try:
+    if not is_flax_available():
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    pass
+else:
+    _import_structure["modeling_flax_gemma"] = [
+        "FlaxGemmaForCausalLM",
+        "FlaxGemmaModel",
+        "FlaxGemmaPreTrainedModel",
+    ]
+
+
+if TYPE_CHECKING:
+    from .configuration_gemma import GEMMA_PRETRAINED_CONFIG_ARCHIVE_MAP, GemmaConfig
+
+    try:
+        if not is_sentencepiece_available():
+            raise OptionalDependencyNotAvailable()
+    except OptionalDependencyNotAvailable:
+        pass
+    else:
+        from .tokenization_gemma import GemmaTokenizer
+
+    try:
+        if not is_tokenizers_available():
+            raise OptionalDependencyNotAvailable()
+    except OptionalDependencyNotAvailable:
+        pass
+    else:
+        from .tokenization_gemma_fast import GemmaTokenizerFast
+
+    try:
+        if not is_torch_available():
+            raise OptionalDependencyNotAvailable()
+    except OptionalDependencyNotAvailable:
+        pass
+    else:
+        from .modeling_gemma import (
+            GemmaForCausalLM,
+            GemmaForSequenceClassification,
+            GemmaModel,
+            GemmaPreTrainedModel,
+        )
+
+    try:
+        if not is_flax_available():
+            raise OptionalDependencyNotAvailable()
+    except OptionalDependencyNotAvailable:
+        pass
+    else:
+        from .modeling_flax_gemma import (
+            FlaxGemmaForCausalLM,
+            FlaxGemmaModel,
+            FlaxGemmaPreTrainedModel,
+        )
+
+
+else:
+    import sys
+
+    sys.modules[__name__] = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)

src/transformers/models/gemma/configuration_gemma.py

+# coding=utf-8
+# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+""" Gemma model configuration"""
+
+from ...configuration_utils import PretrainedConfig
+from ...utils import logging
+
+
+logger = logging.get_logger(__name__)
+
+GEMMA_PRETRAINED_CONFIG_ARCHIVE_MAP = {}
+
+
+class GemmaConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`GemmaModel`]. It is used to instantiate an Gemma
+    model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
+    defaults will yield a similar configuration to that of the Gemma-7B.
+
+    e.g. [google/gemma-7b](https://huggingface.co/google/gemma-7b)
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+
+    Args:
+        vocab_size (`int`, *optional*, defaults to 256000):
+            Vocabulary size of the Gemma model. Defines the number of different tokens that can be represented by the
+            `inputs_ids` passed when calling [`GemmaModel`]
+        hidden_size (`int`, *optional*, defaults to 3072):
+            Dimension of the hidden representations.
+        intermediate_size (`int`, *optional*, defaults to 24576):
+            Dimension of the MLP representations.
+        num_hidden_layers (`int`, *optional*, defaults to 28):
+            Number of hidden layers in the Transformer decoder.
+        num_attention_heads (`int`, *optional*, defaults to 16):
+            Number of attention heads for each attention layer in the Transformer decoder.
+        num_key_value_heads (`int`, *optional*, defaults to 16):
+            This is the number of key_value heads that should be used to implement Grouped Query Attention. If
+            `num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
+            `num_key_value_heads=1 the model will use Multi Query Attention (MQA) otherwise GQA is used. When
+            converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
+            by meanpooling all the original heads within that group. For more details checkout [this
+            paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to
+            `num_attention_heads`.
+        head_dim (`int`, *optional*, defaults to 256):
+            The attention head dimension.
+        hidden_act (`str` or `function`, *optional*, defaults to `"gelu"`):
+            The non-linear activation function (function or string) in the decoder.
+        max_position_embeddings (`int`, *optional*, defaults to 8192):
+            The maximum sequence length that this model might ever be used with.
+        initializer_range (`float`, *optional*, defaults to 0.02):
+            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
+        rms_norm_eps (`float`, *optional*, defaults to 1e-06):
+            The epsilon used by the rms normalization layers.
+        use_cache (`bool`, *optional*, defaults to `True`):
+            Whether or not the model should return the last key/values attentions (not used by all models). Only
+            relevant if `config.is_decoder=True`.
+        pad_token_id (`int`, *optional*, defaults to 0):
+            Padding token id.
+        eos_token_id (`int`, *optional*, defaults to 1):
+            End of stream token id.
+        bos_token_id (`int`, *optional*, defaults to 2):
+            Beginning of stream token id.
+        tie_word_embeddings (`bool`, *optional*, defaults to `True`):
+            Whether to tie weight embeddings
+        rope_theta (`float`, *optional*, defaults to 10000.0):
+            The base period of the RoPE embeddings.
+        attention_bias (`bool`, defaults to `False`, *optional*, defaults to `False`):
+            Whether to use a bias in the query, key, value and output projection layers during self-attention.
+        attention_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio for the attention probabilities.
+
+    ```python
+    >>> from transformers import GemmaModel, GemmaConfig
+
+    >>> # Initializing a Gemma gemma-7b style configuration
+    >>> configuration = GemmaConfig()
+
+    >>> # Initializing a model from the gemma-7b style configuration
+    >>> model = GemmaModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "gemma"
+    keys_to_ignore_at_inference = ["past_key_values"]
+
+    def __init__(
+        self,
+        vocab_size=256000,
+        hidden_size=3072,
+        intermediate_size=24576,
+        num_hidden_layers=28,
+        num_attention_heads=16,
+        num_key_value_heads=16,
+        head_dim=256,
+        hidden_act="gelu",
+        max_position_embeddings=8192,
+        initializer_range=0.02,
+        rms_norm_eps=1e-6,
+        use_cache=True,
+        pad_token_id=0,
+        eos_token_id=1,
+        bos_token_id=2,
+        tie_word_embeddings=True,
+        rope_theta=10000.0,
+        attention_bias=False,
+        attention_dropout=0.0,
+        **kwargs,
+    ):
+        self.vocab_size = vocab_size
+        self.max_position_embeddings = max_position_embeddings
+        self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.head_dim = head_dim
+        self.num_key_value_heads = num_key_value_heads
+        self.hidden_act = hidden_act
+        self.initializer_range = initializer_range
+        self.rms_norm_eps = rms_norm_eps
+        self.use_cache = use_cache
+        self.rope_theta = rope_theta
+        self.attention_bias = attention_bias
+        self.attention_dropout = attention_dropout
+
+        super().__init__(
+            pad_token_id=pad_token_id,
+            bos_token_id=bos_token_id,
+            eos_token_id=eos_token_id,
+            tie_word_embeddings=tie_word_embeddings,
+            **kwargs,
+        )

src/transformers/models/gemma/convert_gemma_weights_to_hf.py

+# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import argparse
+import os
+import warnings
+
+import torch
+from accelerate import init_empty_weights
+
+from transformers import GemmaConfig, GemmaForCausalLM, GemmaTokenizer
+
+
+try:
+    from transformers import GemmaTokenizerFast
+except ImportError as e:
+    warnings.warn(e)
+    warnings.warn(
+        "The converted tokenizer will be the `slow` tokenizer. To use the fast, update your `tokenizers` library and re-run the tokenizer conversion"
+    )
+    GemmaTokenizerFast = None
+
+"""
+Sample usage:
+
+```
+python src/transformers/models/gemma/convert_gemma_weights_to_hf.py \
+    --input_dir /path/to/downloaded/gemma/weights --model_size 7B --output_dir /output/path
+```
+
+Thereafter, models can be loaded via:
+
+```py
+from transformers import GemmaForCausalLM, GemmaTokenizerFast
+
+model = GemmaForCausalLM.from_pretrained("/output/path")
+tokenizer = GemmaTokenizerFast.from_pretrained("/output/path")
+```
+
+Important note: you need to be able to host the whole model in RAM to execute this script (even if the biggest versions
+come in several checkpoints they each contain a part of each weight of the model, so we need to load them all in RAM).
+"""
+
+gemma_2b_config = GemmaConfig(
+    num_hidden_layers=18,
+    num_attention_heads=8,
+    num_key_value_heads=1,
+    hidden_size=2048,
+    intermediate_size=16384,
+)
+
+gemma_7b_config = GemmaConfig()
+
+CONFIG_MAPPING = {"2B": gemma_2b_config, "7B": gemma_7b_config}
+LAYER_NAME_MAPPING = {"embedder.weight": "model.embed_tokens.weight"}
+
+
+def write_model(save_path, input_base_path, config, safe_serialization=True, push_to_hub=False):
+    num_attn_heads = config.num_attention_heads
+    hidden_size = config.hidden_size
+    num_kv_heads = config.num_key_value_heads
+    head_dim = config.head_dim
+
+    print(f"Fetching all parameters from the checkpoint at '{input_base_path}'")
+    model_state_dict = torch.load(input_base_path, map_location="cpu")["model_state_dict"]
+    model_state_dict.pop("freqs_cis")
+
+    state_dict = {}
+    for k, v in model_state_dict.items():
+        if "qkv_proj" in k:
+            if num_kv_heads == 1:
+                v = v.reshape(num_attn_heads + num_kv_heads * 2, head_dim, hidden_size)
+                q_proj = v[:num_attn_heads, ...]
+                k_proj = v[num_attn_heads : num_attn_heads + num_kv_heads, ...].repeat(num_kv_heads, 1, 1)
+                v_proj = v[-num_kv_heads:, ...].repeat(num_kv_heads, 1, 1)
+
+                state_dict[k.replace("qkv_proj", "q_proj")] = q_proj.reshape(
+                    num_attn_heads * head_dim, hidden_size
+                ).clone()
+                state_dict[k.replace("qkv_proj", "k_proj")] = k_proj.reshape(
+                    num_kv_heads * head_dim, hidden_size
+                ).clone()
+                state_dict[k.replace("qkv_proj", "v_proj")] = v_proj[0].clone()
+            else:
+                q_proj, k_proj, v_proj = torch.split(v, v.shape[0] // 3, 0)
+                state_dict[k.replace("qkv_proj", "q_proj")] = q_proj.reshape(
+                    num_attn_heads * head_dim, hidden_size
+                ).clone()
+                state_dict[k.replace("qkv_proj", "k_proj")] = k_proj.reshape(
+                    num_kv_heads * head_dim, hidden_size
+                ).clone()
+                state_dict[k.replace("qkv_proj", "v_proj")] = v_proj.clone()
+
+        elif k == "embedder.weight":
+            state_dict[LAYER_NAME_MAPPING[k]] = v
+            state_dict["lm_head.weight"] = v
+        else:
+            state_dict[k] = v
+
+    print("Loading the checkpoint in a Gemma model.")
+    with init_empty_weights():
+        model = GemmaForCausalLM(config)
+    model.load_state_dict(state_dict, assign=True, strict=False)
+
+    model.config.torch_dtype = torch.float32
+    del model.config._name_or_path
+    print("Saving in the Transformers format.")
+
+    if push_to_hub:
+        print(f"pushing the model to {save_path}")
+        model.push_to_hub(save_path, safe_serialization=safe_serialization, private=True)
+    else:
+        model.save_pretrained(save_path, safe_serialization=safe_serialization)
+
+
+def write_tokenizer(input_tokenizer_path, save_path, push_to_hub=False):
+    # Initialize the tokenizer based on the `spm` model
+    tokenizer_class = GemmaTokenizer if GemmaTokenizerFast is None else GemmaTokenizerFast
+    print(f"Saving a {tokenizer_class.__name__} to {save_path}.")
+    tokenizer = tokenizer_class(input_tokenizer_path)
+    if push_to_hub:
+        tokenizer.push_to_hub(save_path)
+    else:
+        tokenizer.save_pretrained(save_path)
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--input_checkpoint",
+        help="Absolute path to the target Gemma weights.",
+        required=True,
+    )
+    parser.add_argument(
+        "--tokenizer_checkpoint",
+        help="Location of Gemma tokenizer model",
+    )
+    parser.add_argument(
+        "--model_size",
+        default="7B",
+        choices=["2B", "7B", "tokenizer_only"],
+        help="'f' models correspond to the finetuned versions, and are specific to the Gemma2 official release. For more details on Gemma2, checkout the original repo: https://huggingface.co/google/gemma-7b",
+    )
+    parser.add_argument(
+        "--output_dir",
+        default="google/gemma-7b",
+        help="Location to write HF model and tokenizer",
+    )
+    parser.add_argument(
+        "--pickle_serialization",
+        help="Whether or not to save using `safetensors`.",
+        action="store_true",
+        default=False,
+    )
+    parser.add_argument(
+        "--convert_tokenizer",
+        help="Whether or not to convert the tokenizer as well.",
+        action="store_true",
+        default=False,
+    )
+    parser.add_argument(
+        "--push_to_hub",
+        help="Whether or not to push the model to the hub at `output_dir` instead of saving it locally.",
+        action="store_true",
+        default=False,
+    )
+    args = parser.parse_args()
+
+    if args.convert_tokenizer:
+        if args.tokenizer_checkpoint is None:
+            raise ValueError("Path to the tokenizer is required when passing --convert_tokenizer")
+
+        spm_path = os.path.join(args.tokenizer_checkpoint)
+        write_tokenizer(spm_path, args.output_dir, args.push_to_hub)
+
+    config = CONFIG_MAPPING[args.model_size]
+    write_model(
+        config=config,
+        input_base_path=args.input_checkpoint,
+        save_path=args.output_dir,
+        safe_serialization=not args.pickle_serialization,
+        push_to_hub=args.push_to_hub,
+    )
+
+
+if __name__ == "__main__":
+    main()

src/transformers/models/gemma/modeling_flax_gemma.py

+# coding=utf-8
+# Copyright 2024 Google Inc., and the HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Flax Gemma model."""
+from typing import Optional, Tuple
+
+import flax.linen as nn
+import jax
+import jax.numpy as jnp
+import numpy as np
+from flax.core.frozen_dict import FrozenDict, freeze, unfreeze
+from flax.linen import combine_masks, make_causal_mask
+from flax.linen.attention import dot_product_attention_weights
+from flax.traverse_util import flatten_dict, unflatten_dict
+from jax import lax
+
+from ...modeling_flax_outputs import FlaxBaseModelOutput, FlaxCausalLMOutput
+from ...modeling_flax_utils import ACT2FN, FlaxPreTrainedModel, append_call_sample_docstring
+from ...utils import add_start_docstrings, add_start_docstrings_to_model_forward, logging
+from .configuration_gemma import GemmaConfig
+
+
+logger = logging.get_logger(__name__)
+
+_CONFIG_FOR_DOC = "GemmaConfig"
+_CHECKPOINT_FOR_DOC = "google/gemma-2b"
+_REAL_CHECKPOINT_FOR_DOC = "openlm-research/open_llama_3b_v2"
+
+GEMMA_START_DOCSTRING = r"""
+
+    This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
+    library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
+    etc.)
+
+    This model is also a Flax Linen
+    [flax.nn.Module](https://flax.readthedocs.io/en/latest/_autosummary/flax.nn.module.html) subclass. Use it as a
+    regular Flax Module and refer to the Flax documentation for all matter related to general usage and behavior.
+
+    Finally, this model supports inherent JAX features such as:
+
+    - [Just-In-Time (JIT) compilation](https://jax.readthedocs.io/en/latest/jax.html#just-in-time-compilation-jit)
+    - [Automatic Differentiation](https://jax.readthedocs.io/en/latest/jax.html#automatic-differentiation)
+    - [Vectorization](https://jax.readthedocs.io/en/latest/jax.html#vectorization-vmap)
+    - [Parallelization](https://jax.readthedocs.io/en/latest/jax.html#parallelization-pmap)
+
+    Parameters:
+        config ([`GemmaConfig`]): Model configuration class with all the parameters of the model.
+            Initializing with a config file does not load the weights associated with the model, only the
+            configuration. Check out the [`~FlaxPreTrainedModel.from_pretrained`] method to load the model weights.
+        dtype (`jax.numpy.dtype`, *optional*, defaults to `jax.numpy.float32`):
+            The data type of the computation. Can be one of `jax.numpy.float32`, `jax.numpy.float16`, or
+            `jax.numpy.bfloat16`.
+
+            This can be used to enable mixed-precision training or half-precision inference on GPUs or TPUs. If
+            specified all the computation will be performed with the given `dtype`.
+
+            **Note that this only specifies the dtype of the computation and does not influence the dtype of model
+            parameters.**
+
+            If you wish to change the dtype of the model parameters, see [`~FlaxPreTrainedModel.to_fp16`] and
+            [`~FlaxPreTrainedModel.to_bf16`].
+"""
+
+GEMMA_INPUTS_DOCSTRING = r"""
+    Args:
+        input_ids (`numpy.ndarray` of shape `(batch_size, input_ids_length)`):
+            Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
+            it.
+
+            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            [What are input IDs?](../glossary#input-ids)
+        attention_mask (`numpy.ndarray` of shape `(batch_size, sequence_length)`, *optional*):
+            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
+
+            - 1 for tokens that are **not masked**,
+            - 0 for tokens that are **masked**.
+
+            [What are attention masks?](../glossary#attention-mask)
+
+            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            If `past_key_values` is used, optionally only the last `decoder_input_ids` have to be input (see
+            `past_key_values`).
+
+            If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`]
+            and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more
+            information on the default strategy.
+
+            - 1 indicates the head is **not masked**,
+            - 0 indicates the head is **masked**.
+        position_ids (`numpy.ndarray` of shape `(batch_size, sequence_length)`, *optional*):
+            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
+            config.n_positions - 1]`.
+
+            [What are position IDs?](../glossary#position-ids)
+        past_key_values (`Dict[str, np.ndarray]`, *optional*, returned by `init_cache` or when passing previous `past_key_values`):
+            Dictionary of pre-computed hidden-states (key and values in the attention blocks) that can be used for fast
+            auto-regressive decoding. Pre-computed key and value hidden-states are of shape *[batch_size, max_length]*.
+        output_attentions (`bool`, *optional*):
+            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
+            tensors for more detail.
+        output_hidden_states (`bool`, *optional*):
+            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
+            more detail.
+        return_dict (`bool`, *optional*):
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+"""
+
+
+def create_sinusoidal_positions(num_pos, dim):
+    inv_freq = 1.0 / (10000 ** (np.arange(0, dim, 2)[: (dim // 2)] / dim))
+    freqs = np.einsum("i , j -> i j", np.arange(num_pos), inv_freq).astype("float32")
+
+    emb = np.concatenate((freqs, freqs), axis=-1)
+    out = np.concatenate((np.sin(emb)[:, None, :], np.cos(emb)[:, None, :]), axis=-1)
+    return jnp.array(out[:, :, :num_pos])
+
+
+# Copied from transformers.models.llama.modeling_flax_llama.rotate_half
+def rotate_half(tensor):
+    """Rotates half the hidden dims of the input."""
+    rotate_half_tensor = jnp.concatenate(
+        (-tensor[..., tensor.shape[-1] // 2 :], tensor[..., : tensor.shape[-1] // 2]), axis=-1
+    )
+    return rotate_half_tensor
+
+
+# Copied from transformers.models.llama.modeling_flax_llama.apply_rotary_pos_emb
+def apply_rotary_pos_emb(tensor, sin_pos, cos_pos):
+    return (tensor * cos_pos) + (rotate_half(tensor) * sin_pos)
+
+
+class FlaxGemmaRMSNorm(nn.Module):
+    config: GemmaConfig
+    dtype: jnp.dtype = jnp.float32
+
+    def setup(self):
+        self.epsilon = self.config.rms_norm_eps
+        self.weight = self.param("weight", lambda _, shape: jnp.ones(shape), self.config.hidden_size)
+
+    def __call__(self, hidden_states):
+        variance = jnp.asarray(hidden_states, dtype=jnp.float32)
+        variance = jnp.power(variance, 2)
+        variance = variance.mean(-1, keepdims=True)
+        # use `jax.numpy.sqrt` as `jax.lax.rsqrt` does not match `torch.rsqrt`
+        hidden_states = hidden_states / jnp.sqrt(variance + self.epsilon)
+
+        return (1 + self.weight) * jnp.asarray(hidden_states, dtype=self.dtype)
+
+
+# Copied from transformers.models.llama.modeling_flax_llama.FlaxLlamaRotaryEmbedding with Llama->Gemma
+class FlaxGemmaRotaryEmbedding(nn.Module):
+    config: GemmaConfig
+    dtype: jnp.dtype = jnp.float32
+
+    # Ignore copy
+    def setup(self):
+        head_dim = self.config.head_dim
+        self.sincos = create_sinusoidal_positions(self.config.max_position_embeddings, head_dim)
+
+    def __call__(self, key, query, position_ids):
+        sincos = self.sincos[position_ids]
+        sin_pos, cos_pos = jnp.split(sincos, 2, axis=-1)
+
+        key = apply_rotary_pos_emb(key, sin_pos, cos_pos)
+        query = apply_rotary_pos_emb(query, sin_pos, cos_pos)
+
+        key = jnp.asarray(key, dtype=self.dtype)
+        query = jnp.asarray(query, dtype=self.dtype)
+
+        return key, query
+
+
+class FlaxGemmaAttention(nn.Module):
+    config: GemmaConfig
+    dtype: jnp.dtype = jnp.float32
+    causal: bool = True
+    is_cross_attention: bool = False
+
+    def setup(self):
+        config = self.config
+        self.embed_dim = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = config.head_dim
+        self.attention_softmax_in_fp32 = self.dtype is not jnp.float32
+
+        self.num_key_value_heads = config.num_key_value_heads
+        self.num_key_value_groups = self.num_heads // self.num_key_value_heads
+
+        kernel = jax.nn.initializers.normal(self.config.initializer_range)
+        self.q_proj = nn.Dense(
+            self.num_heads * self.head_dim, use_bias=config.attention_bias, dtype=self.dtype, kernel_init=kernel
+        )
+        self.k_proj = nn.Dense(
+            self.num_key_value_heads * self.head_dim,
+            use_bias=config.attention_bias,
+            dtype=self.dtype,
+            kernel_init=kernel,
+        )
+        self.v_proj = nn.Dense(
+            self.num_key_value_heads * self.head_dim,
+            use_bias=config.attention_bias,
+            dtype=self.dtype,
+            kernel_init=kernel,
+        )
+        self.o_proj = nn.Dense(self.embed_dim, use_bias=config.attention_bias, dtype=self.dtype, kernel_init=kernel)
+
+        self.causal_mask = make_causal_mask(jnp.ones((1, config.max_position_embeddings), dtype="bool"), dtype="bool")
+        self.rotary_emb = FlaxGemmaRotaryEmbedding(config, dtype=self.dtype)
+
+    def _split_heads(self, hidden_states, num_heads):
+        return hidden_states.reshape(hidden_states.shape[:2] + (num_heads, self.head_dim))
+
+    def _merge_heads(self, hidden_states):
+        return hidden_states.reshape(hidden_states.shape[:2] + (self.num_heads * self.head_dim,))
+
+    @nn.compact
+    # Copied from transformers.models.gpt_neo.modeling_flax_gpt_neo.FlaxGPTNeoSelfAttention._concatenate_to_cache
+    def _concatenate_to_cache(self, key, value, query, attention_mask):
+        """
+        This function takes projected key, value states from a single input token and concatenates the states to cached
+        states from previous steps. This function is slighly adapted from the official Flax repository:
+        https://github.com/google/flax/blob/491ce18759622506588784b4fca0e4bf05f8c8cd/flax/linen/attention.py#L252
+        """
+        # detect if we're initializing by absence of existing cache data.
+        is_initialized = self.has_variable("cache", "cached_key")
+        cached_key = self.variable("cache", "cached_key", jnp.zeros, key.shape, key.dtype)
+        cached_value = self.variable("cache", "cached_value", jnp.zeros, value.shape, value.dtype)
+        cache_index = self.variable("cache", "cache_index", lambda: jnp.array(0, dtype=jnp.int32))
+
+        if is_initialized:
+            *batch_dims, max_length, num_heads, depth_per_head = cached_key.value.shape
+            # update key, value caches with our new 1d spatial slices
+            cur_index = cache_index.value
+            indices = (0,) * len(batch_dims) + (cur_index, 0, 0)
+            key = lax.dynamic_update_slice(cached_key.value, key, indices)
+            value = lax.dynamic_update_slice(cached_value.value, value, indices)
+            cached_key.value = key
+            cached_value.value = value
+            num_updated_cache_vectors = query.shape[1]
+            cache_index.value = cache_index.value + num_updated_cache_vectors
+            # causal mask for cached decoder self-attention: our single query position should only attend to those key positions that have already been generated and cached, not the remaining zero elements.
+            pad_mask = jnp.broadcast_to(
+                jnp.arange(max_length) < cur_index + num_updated_cache_vectors,
+                tuple(batch_dims) + (1, num_updated_cache_vectors, max_length),
+            )
+            attention_mask = combine_masks(pad_mask, attention_mask)
+        return key, value, attention_mask
+
+    def __call__(
+        self,
+        hidden_states,
+        attention_mask,
+        position_ids,
+        deterministic: bool = True,
+        init_cache: bool = False,
+        output_attentions: bool = False,
+    ):
+        query = self.q_proj(hidden_states)
+        key = self.k_proj(hidden_states)
+        value = self.v_proj(hidden_states)
+
+        query = self._split_heads(query, self.num_heads)
+        key = self._split_heads(key, self.num_key_value_heads)
+        value = self._split_heads(value, self.num_key_value_heads)
+
+        key, query = self.rotary_emb(key, query, position_ids)
+
+        query_length, key_length = query.shape[1], key.shape[1]
+
+        if self.has_variable("cache", "cached_key"):
+            mask_shift = self.variables["cache"]["cache_index"]
+            max_decoder_length = self.variables["cache"]["cached_key"].shape[1]
+            causal_mask = lax.dynamic_slice(
+                self.causal_mask, (0, 0, mask_shift, 0), (1, 1, query_length, max_decoder_length)
+            )
+        else:
+            causal_mask = self.causal_mask[:, :, :query_length, :key_length]
+
+        batch_size = hidden_states.shape[0]
+        causal_mask = jnp.broadcast_to(causal_mask, (batch_size,) + causal_mask.shape[1:])
+
+        attention_mask = jnp.broadcast_to(jnp.expand_dims(attention_mask, axis=(-3, -2)), causal_mask.shape)
+        attention_mask = combine_masks(attention_mask, causal_mask)
+
+        dropout_rng = None
+        if not deterministic and self.config.attention_dropout > 0.0:
+            dropout_rng = self.make_rng("dropout")
+
+        # During fast autoregressive decoding, we feed one position at a time,
+        # and cache the keys and values step by step.
+        if self.has_variable("cache", "cached_key") or init_cache:
+            key, value, attention_mask = self._concatenate_to_cache(key, value, query, attention_mask)
+
+        # transform boolean mask into float mask
+        attention_bias = lax.select(
+            attention_mask > 0,
+            jnp.full(attention_mask.shape, 0.0).astype(self.dtype),
+            jnp.full(attention_mask.shape, jnp.finfo(self.dtype).min).astype(self.dtype),
+        )
+
+        key = jnp.repeat(key, repeats=self.num_key_value_groups, axis=2)
+        value = jnp.repeat(value, repeats=self.num_key_value_groups, axis=2)
+
+        # usual dot product attention
+        attention_dtype = jnp.float32 if self.attention_softmax_in_fp32 else self.dtype
+        attn_weights = dot_product_attention_weights(
+            query,
+            key,
+            bias=attention_bias,
+            dropout_rng=dropout_rng,
+            dropout_rate=self.config.attention_dropout,
+            deterministic=deterministic,
+            dtype=attention_dtype,
+        )
+
+        if self.attention_softmax_in_fp32:
+            attn_weights = attn_weights.astype(self.dtype)
+
+        attn_output = jnp.einsum("...hqk,...khd->...qhd", attn_weights, value)
+        attn_output = self._merge_heads(attn_output)
+        attn_output = self.o_proj(attn_output)
+
+        outputs = (attn_output, attn_weights) if output_attentions else (attn_output,)
+        return outputs
+
+
+# Copied from transformers.models.llama.modeling_flax_llama.FlaxLlamaMLP with Llama->Gemma
+class FlaxGemmaMLP(nn.Module):
+    config: GemmaConfig
+    dtype: jnp.dtype = jnp.float32
+
+    def setup(self):
+        embed_dim = self.config.hidden_size
+        inner_dim = self.config.intermediate_size if self.config.intermediate_size is not None else 4 * embed_dim
+
+        kernel_init = jax.nn.initializers.normal(self.config.initializer_range)
+        self.act = ACT2FN[self.config.hidden_act]
+
+        self.gate_proj = nn.Dense(inner_dim, use_bias=False, dtype=self.dtype, kernel_init=kernel_init)
+        self.down_proj = nn.Dense(embed_dim, use_bias=False, dtype=self.dtype, kernel_init=kernel_init)
+        self.up_proj = nn.Dense(inner_dim, use_bias=False, dtype=self.dtype, kernel_init=kernel_init)
+
+    def __call__(self, hidden_states):
+        up_proj_states = self.up_proj(hidden_states)
+        gate_states = self.act(self.gate_proj(hidden_states))
+
+        hidden_states = self.down_proj(up_proj_states * gate_states)
+        return hidden_states
+
+
+# Copied from transformers.models.llama.modeling_flax_llama.FlaxLlamaDecoderLayer with Llama->Gemma
+class FlaxGemmaDecoderLayer(nn.Module):
+    config: GemmaConfig
+    dtype: jnp.dtype = jnp.float32
+
+    def setup(self):
+        self.input_layernorm = FlaxGemmaRMSNorm(self.config, dtype=self.dtype)
+        self.self_attn = FlaxGemmaAttention(self.config, dtype=self.dtype)
+        self.post_attention_layernorm = FlaxGemmaRMSNorm(self.config, dtype=self.dtype)
+        self.mlp = FlaxGemmaMLP(self.config, dtype=self.dtype)
+
+    def __call__(
+        self,
+        hidden_states,
+        attention_mask=None,
+        position_ids=None,
+        deterministic: bool = True,
+        init_cache: bool = False,
+        output_attentions: bool = False,
+    ):
+        residual = hidden_states
+        hidden_states = self.input_layernorm(hidden_states)
+        outputs = self.self_attn(
+            hidden_states,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            deterministic=deterministic,
+            init_cache=init_cache,
+            output_attentions=output_attentions,
+        )
+        # residual connection
+        attn_output = outputs[0]
+        hidden_states = residual + attn_output
+
+        residual = hidden_states
+        hidden_states = self.post_attention_layernorm(hidden_states)
+        hidden_states = self.mlp(hidden_states)
+        # residual connection
+        hidden_states = residual + hidden_states
+
+        return (hidden_states,) + outputs[1:]
+
+
+# Copied from transformers.models.gpt_neo.modeling_flax_gpt_neo.FlaxGPTNeoPreTrainedModel with GPTNeo->Gemma, GPT_NEO->GEMMA, transformer->model
+class FlaxGemmaPreTrainedModel(FlaxPreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
+    models.
+    """
+
+    config_class = GemmaConfig
+    base_model_prefix = "model"
+    module_class: nn.Module = None
+
+    def __init__(
+        self,
+        config: GemmaConfig,
+        input_shape: Tuple = (1, 1),
+        seed: int = 0,
+        dtype: jnp.dtype = jnp.float32,
+        _do_init: bool = True,
+        **kwargs,
+    ):
+        module = self.module_class(config=config, dtype=dtype, **kwargs)
+        super().__init__(config, module, input_shape=input_shape, seed=seed, dtype=dtype, _do_init=_do_init)
+
+    def init_weights(self, rng: jax.random.PRNGKey, input_shape: Tuple, params: FrozenDict = None) -> FrozenDict:
+        # init input tensors
+        input_ids = jnp.zeros(input_shape, dtype="i4")
+        attention_mask = jnp.ones_like(input_ids)
+        position_ids = jnp.broadcast_to(jnp.arange(jnp.atleast_2d(input_ids).shape[-1]), input_shape)
+        params_rng, dropout_rng = jax.random.split(rng)
+        rngs = {"params": params_rng, "dropout": dropout_rng}
+
+        random_params = self.module.init(rngs, input_ids, attention_mask, position_ids, return_dict=False)["params"]
+
+        if params is not None:
+            random_params = flatten_dict(unfreeze(random_params))
+            params = flatten_dict(unfreeze(params))
+            for missing_key in self._missing_keys:
+                params[missing_key] = random_params[missing_key]
+            self._missing_keys = set()
+            return freeze(unflatten_dict(params))
+        else:
+            return random_params
+
+    def init_cache(self, batch_size, max_length):
+        r"""
+        Args:
+            batch_size (`int`):
+                batch_size used for fast auto-regressive decoding. Defines the batch size of the initialized cache.
+            max_length (`int`):
+                maximum possible length for auto-regressive decoding. Defines the sequence length of the initialized
+                cache.
+        """
+        # init input variables to retrieve cache
+        input_ids = jnp.ones((batch_size, max_length))
+        attention_mask = jnp.ones_like(input_ids)
+        position_ids = jnp.broadcast_to(jnp.arange(jnp.atleast_2d(input_ids).shape[-1]), input_ids.shape)
+
+        init_variables = self.module.init(
+            jax.random.PRNGKey(0), input_ids, attention_mask, position_ids, return_dict=False, init_cache=True
+        )
+        return unfreeze(init_variables["cache"])
+
+    @add_start_docstrings_to_model_forward(GEMMA_INPUTS_DOCSTRING)
+    def __call__(
+        self,
+        input_ids,
+        attention_mask=None,
+        position_ids=None,
+        params: dict = None,
+        past_key_values: dict = None,
+        dropout_rng: jax.random.PRNGKey = None,
+        train: bool = False,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ):
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.return_dict
+
+        batch_size, sequence_length = input_ids.shape
+
+        if position_ids is None:
+            if past_key_values is not None:
+                raise ValueError("Make sure to provide `position_ids` when passing `past_key_values`.")
+
+            position_ids = jnp.broadcast_to(jnp.arange(sequence_length)[None, :], (batch_size, sequence_length))
+
+        if attention_mask is None:
+            attention_mask = jnp.ones((batch_size, sequence_length))
+
+        # Handle any PRNG if needed
+        rngs = {}
+        if dropout_rng is not None:
+            rngs["dropout"] = dropout_rng
+
+        inputs = {"params": params or self.params}
+
+        # if past_key_values are passed then cache is already initialized a private flag init_cache has to be passed down to ensure cache is used. It has to be made sure that cache is marked as mutable so that it can be changed by FlaxGemmaAttention module
+        if past_key_values:
+            inputs["cache"] = past_key_values
+            mutable = ["cache"]
+        else:
+            mutable = False
+
+        outputs = self.module.apply(
+            inputs,
+            jnp.array(input_ids, dtype="i4"),
+            jnp.array(attention_mask, dtype="i4"),
+            jnp.array(position_ids, dtype="i4"),
+            not train,
+            False,
+            output_attentions,
+            output_hidden_states,
+            return_dict,
+            rngs=rngs,
+            mutable=mutable,
+        )
+
+        # add updated cache to model output
+        if past_key_values is not None and return_dict:
+            outputs, past_key_values = outputs
+            outputs["past_key_values"] = unfreeze(past_key_values["cache"])
+            return outputs
+        elif past_key_values is not None and not return_dict:
+            outputs, past_key_values = outputs
+            outputs = outputs[:1] + (unfreeze(past_key_values["cache"]),) + outputs[1:]
+
+        return outputs
+
+
+# Copied from transformers.models.llama.modeling_flax_llama.FlaxLlamaLayerCollection with Llama->Gemma
+class FlaxGemmaLayerCollection(nn.Module):
+    config: GemmaConfig
+    dtype: jnp.dtype = jnp.float32
+
+    def setup(self):
+        self.blocks = [
+            FlaxGemmaDecoderLayer(self.config, dtype=self.dtype, name=str(i))
+            for i in range(self.config.num_hidden_layers)
+        ]
+
+    def __call__(
+        self,
+        hidden_states,
+        attention_mask=None,
+        position_ids=None,
+        deterministic: bool = True,
+        init_cache: bool = False,
+        output_attentions: bool = False,
+        output_hidden_states: bool = False,
+        return_dict: bool = False,
+    ):
+        all_attentions = () if output_attentions else None
+        all_hidden_states = () if output_hidden_states else None
+
+        for block in self.blocks:
+            if output_hidden_states:
+                all_hidden_states += (hidden_states,)
+            layer_outputs = block(
+                hidden_states,
+                attention_mask=attention_mask,
+                position_ids=position_ids,
+                deterministic=deterministic,
+                init_cache=init_cache,
+                output_attentions=output_attentions,
+            )
+            hidden_states = layer_outputs[0]
+
+            if output_attentions:
+                all_attentions += (layer_outputs[1],)
+
+        # this contains possible `None` values - `FlaxGemmaModule` will filter them out
+        outputs = (hidden_states, all_hidden_states, all_attentions)
+
+        return outputs
+
+
+# Copied from transformers.models.llama.modeling_flax_llama.FlaxLlamaModule with Llama->Gemma
+class FlaxGemmaModule(nn.Module):
+    config: GemmaConfig
+    dtype: jnp.dtype = jnp.float32
+
+    def setup(self):
+        self.hidden_size = self.config.hidden_size
+        embedding_init = jax.nn.initializers.normal(stddev=self.config.initializer_range)
+        self.embed_tokens = nn.Embed(
+            self.config.vocab_size,
+            self.hidden_size,
+            embedding_init=embedding_init,
+            dtype=self.dtype,
+        )
+        self.layers = FlaxGemmaLayerCollection(self.config, dtype=self.dtype)
+        self.norm = FlaxGemmaRMSNorm(self.config, dtype=self.dtype)
+
+    # Ignore copy
+    def __call__(
+        self,
+        input_ids,
+        attention_mask=None,
+        position_ids=None,
+        deterministic=True,
+        init_cache: bool = False,
+        output_attentions: bool = False,
+        output_hidden_states: bool = False,
+        return_dict: bool = True,
+    ):
+        input_embeds = self.embed_tokens(input_ids.astype("i4"))
+
+        input_embeds = input_embeds * (self.config.hidden_size**0.5)
+
+        outputs = self.layers(
+            input_embeds,
+            position_ids=position_ids,
+            attention_mask=attention_mask,
+            deterministic=deterministic,
+            init_cache=init_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        hidden_states = outputs[0]
+        hidden_states = self.norm(hidden_states)
+
+        if output_hidden_states:
+            all_hidden_states = outputs[1] + (hidden_states,)
+            outputs = (hidden_states, all_hidden_states) + outputs[2:]
+        else:
+            outputs = (hidden_states,) + outputs[1:]
+
+        if not return_dict:
+            return tuple(v for v in outputs if v is not None)
+
+        return FlaxBaseModelOutput(
+            last_hidden_state=hidden_states,
+            hidden_states=outputs[1],
+            attentions=outputs[-1],
+        )
+
+
+@add_start_docstrings(
+    "The bare Gemma Model transformer outputting raw hidden-states without any specific head on top.",
+    GEMMA_START_DOCSTRING,
+)
+# Copied from transformers.models.llama.modeling_flax_llama.FlaxLlamaModel with Llama->Gemma
+class FlaxGemmaModel(FlaxGemmaPreTrainedModel):
+    module_class = FlaxGemmaModule
+
+
+append_call_sample_docstring(
+    FlaxGemmaModel,
+    _CHECKPOINT_FOR_DOC,
+    FlaxBaseModelOutput,
+    _CONFIG_FOR_DOC,
+    real_checkpoint=_REAL_CHECKPOINT_FOR_DOC,
+)
+
+
+# Copied from transformers.models.llama.modeling_flax_llama.FlaxLlamaForCausalLMModule with Llama->Gemma
+class FlaxGemmaForCausalLMModule(nn.Module):
+    config: GemmaConfig
+    dtype: jnp.dtype = jnp.float32
+
+    def setup(self):
+        self.model = FlaxGemmaModule(self.config, dtype=self.dtype)
+        self.lm_head = nn.Dense(
+            self.config.vocab_size,
+            use_bias=False,
+            dtype=self.dtype,
+            kernel_init=jax.nn.initializers.normal(stddev=self.config.initializer_range),
+        )
+
+    # Ignore copy
+    def __call__(
+        self,
+        input_ids,
+        attention_mask=None,
+        position_ids=None,
+        deterministic: bool = True,
+        init_cache: bool = False,
+        output_attentions: bool = False,
+        output_hidden_states: bool = False,
+        return_dict: bool = True,
+    ):
+        outputs = self.model(
+            input_ids,
+            position_ids=position_ids,
+            attention_mask=attention_mask,
+            deterministic=deterministic,
+            init_cache=init_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        hidden_states = outputs[0]
+        if self.config.tie_word_embeddings:
+            shared_kernel = self.model.variables["params"]["embed_tokens"]["embedding"].T
+            lm_logits = self.lm_head.apply({"params": {"kernel": shared_kernel}}, hidden_states)
+        else:
+            lm_logits = self.lm_head(hidden_states)
+
+        if not return_dict:
+            return (lm_logits,) + outputs[1:]
+
+        return FlaxCausalLMOutput(logits=lm_logits, hidden_states=outputs.hidden_states, attentions=outputs.attentions)
+
+
+@add_start_docstrings(
+    """
+    The Gemma Model transformer with a language modeling head (linear layer) on top.
+    """,
+    GEMMA_START_DOCSTRING,
+)
+# Copied from transformers.models.gptj.modeling_flax_gptj.FlaxGPTJForCausalLM with GPTJ->Gemma
+class FlaxGemmaForCausalLM(FlaxGemmaPreTrainedModel):
+    module_class = FlaxGemmaForCausalLMModule
+
+    def prepare_inputs_for_generation(self, input_ids, max_length, attention_mask: Optional[jax.Array] = None):
+        # initializing the cache
+        batch_size, seq_length = input_ids.shape
+
+        past_key_values = self.init_cache(batch_size, max_length)
+        # Note that usually one would have to put 0's in the attention_mask for x > input_ids.shape[-1] and x < cache_length.
+        # But since Gemma uses a causal mask, those positions are masked anyways.
+        # Thus we can create a single static attention_mask here, which is more efficient for compilation
+        extended_attention_mask = jnp.ones((batch_size, max_length), dtype="i4")
+        if attention_mask is not None:
+            position_ids = attention_mask.cumsum(axis=-1) - 1
+            extended_attention_mask = lax.dynamic_update_slice(extended_attention_mask, attention_mask, (0, 0))
+        else:
+            position_ids = jnp.broadcast_to(jnp.arange(seq_length, dtype="i4")[None, :], (batch_size, seq_length))
+
+        return {
+            "past_key_values": past_key_values,
+            "attention_mask": extended_attention_mask,
+            "position_ids": position_ids,
+        }
+
+    def update_inputs_for_generation(self, model_outputs, model_kwargs):
+        model_kwargs["past_key_values"] = model_outputs.past_key_values
+        model_kwargs["position_ids"] = model_kwargs["position_ids"][:, -1:] + 1
+        return model_kwargs
+
+
+append_call_sample_docstring(
+    FlaxGemmaForCausalLM,
+    _CHECKPOINT_FOR_DOC,
+    FlaxCausalLMOutput,
+    _CONFIG_FOR_DOC,
+    real_checkpoint=_REAL_CHECKPOINT_FOR_DOC,
+)

src/transformers/models/gemma/modeling_gemma.py

+# coding=utf-8
+# Copyright 2024 Google Inc. HuggingFace Inc. team. All rights reserved.
+#
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+""" PyTorch Gemma model."""
+import math
+import warnings
+from typing import List, Optional, Tuple, Union
+
+import torch
+import torch.nn.functional as F
+import torch.utils.checkpoint
+from torch import nn
+from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
+
+from ...activations import ACT2FN
+from ...cache_utils import Cache, DynamicCache, StaticCache
+from ...modeling_attn_mask_utils import (
+    _prepare_4d_causal_attention_mask,
+)
+from ...modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast, SequenceClassifierOutputWithPast
+from ...modeling_utils import PreTrainedModel
+from ...pytorch_utils import ALL_LAYERNORM_LAYERS, is_torch_greater_or_equal_than_1_13
+from ...utils import (
+    add_start_docstrings,
+    add_start_docstrings_to_model_forward,
+    is_flash_attn_2_available,
+    is_flash_attn_greater_or_equal_2_10,
+    logging,
+    replace_return_docstrings,
+)
+from ...utils.import_utils import is_torch_fx_available
+from .configuration_gemma import GemmaConfig
+
+
+if is_flash_attn_2_available():
+    from flash_attn import flash_attn_func, flash_attn_varlen_func
+    from flash_attn.bert_padding import index_first_axis, pad_input, unpad_input  # noqa
+
+
+# This makes `_prepare_4d_causal_attention_mask` a leaf function in the FX graph.
+# It means that the function will not be traced through and simply appear as a node in the graph.
+if is_torch_fx_available():
+    if not is_torch_greater_or_equal_than_1_13:
+        import torch.fx
+
+    _prepare_4d_causal_attention_mask = torch.fx.wrap(_prepare_4d_causal_attention_mask)
+
+
+logger = logging.get_logger(__name__)
+
+_CONFIG_FOR_DOC = "GemmaConfig"
+
+
+def _get_unpad_data(attention_mask):
+    seqlens_in_batch = attention_mask.sum(dim=-1, dtype=torch.int32)
+    indices = torch.nonzero(attention_mask.flatten(), as_tuple=False).flatten()
+    max_seqlen_in_batch = seqlens_in_batch.max().item()
+    cu_seqlens = F.pad(torch.cumsum(seqlens_in_batch, dim=0, dtype=torch.torch.int32), (1, 0))
+    return (
+        indices,
+        cu_seqlens,
+        max_seqlen_in_batch,
+    )
+
+
+class GemmaRMSNorm(nn.Module):
+    def __init__(self, dim: int, eps: float = 1e-6):
+        super().__init__()
+        self.eps = eps
+        self.weight = nn.Parameter(torch.zeros(dim))
+
+    def _norm(self, x):
+        return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps)
+
+    def forward(self, x):
+        output = self._norm(x.float()).type_as(x)
+        return output * (1 + self.weight)
+
+
+ALL_LAYERNORM_LAYERS.append(GemmaRMSNorm)
+
+
+class GemmaRotaryEmbedding(nn.Module):
+    def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):
+        super().__init__()
+
+        self.dim = dim
+        self.max_position_embeddings = max_position_embeddings
+        self.base = base
+        self.register_buffer("inv_freq", None, persistent=False)
+
+    def forward(self, x, position_ids, seq_len=None):
+        # x: [bs, num_attention_heads, seq_len, head_size]
+        if self.inv_freq is None:
+            self.inv_freq = 1.0 / (
+                self.base ** (torch.arange(0, self.dim, 2, dtype=torch.int64, device=x.device).float() / self.dim)
+            )
+
+        inv_freq_expanded = self.inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1)
+        position_ids_expanded = position_ids[:, None, :].float()
+        freqs = (inv_freq_expanded @ position_ids_expanded).transpose(1, 2)
+        emb = torch.cat((freqs, freqs), dim=-1)
+        return emb.cos().to(dtype=x.dtype), emb.sin().to(dtype=x.dtype)
+
+
+# Copied from transformers.models.llama.modeling_llama.rotate_half
+def rotate_half(x):
+    """Rotates half the hidden dims of the input."""
+    x1 = x[..., : x.shape[-1] // 2]
+    x2 = x[..., x.shape[-1] // 2 :]
+    return torch.cat((-x2, x1), dim=-1)
+
+
+# Copied from transformers.models.llama.modeling_llama.apply_rotary_pos_emb
+def apply_rotary_pos_emb(q, k, cos, sin, position_ids, unsqueeze_dim=1):
+    """Applies Rotary Position Embedding to the query and key tensors.
+
+    Args:
+        q (`torch.Tensor`): The query tensor.
+        k (`torch.Tensor`): The key tensor.
+        cos (`torch.Tensor`): The cosine part of the rotary embedding.
+        sin (`torch.Tensor`): The sine part of the rotary embedding.
+        position_ids (`torch.Tensor`):
+            The position indices of the tokens corresponding to the query and key tensors. For example, this can be
+            used to pass offsetted position ids when working with a KV-cache.
+        unsqueeze_dim (`int`, *optional*, defaults to 1):
+            The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and
+            sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. For example, note
+            that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. Then, if q and
+            k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes
+            cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. Similarly, if q and k have
+            the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2.
+    Returns:
+        `tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding.
+    """
+    cos = cos.unsqueeze(unsqueeze_dim)
+    sin = sin.unsqueeze(unsqueeze_dim)
+    q_embed = (q * cos) + (rotate_half(q) * sin)
+    k_embed = (k * cos) + (rotate_half(k) * sin)
+    return q_embed, k_embed
+
+
+# Copied from transformers.models.mistral.modeling_mistral.MistralMLP with Mistral->Gemma
+class GemmaMLP(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.hidden_size = config.hidden_size
+        self.intermediate_size = config.intermediate_size
+        self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
+        self.act_fn = ACT2FN[config.hidden_act]
+
+    def forward(self, x):
+        return self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
+
+
+# Copied from transformers.models.llama.modeling_llama.repeat_kv
+def repeat_kv(hidden_states: torch.Tensor, n_rep: int) -> torch.Tensor:
+    """
+    This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
+    num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
+    """
+    batch, num_key_value_heads, slen, head_dim = hidden_states.shape
+    if n_rep == 1:
+        return hidden_states
+    hidden_states = hidden_states[:, :, None, :, :].expand(batch, num_key_value_heads, n_rep, slen, head_dim)
+    return hidden_states.reshape(batch, num_key_value_heads * n_rep, slen, head_dim)
+
+
+class GemmaAttention(nn.Module):
+    """Multi-headed attention from 'Attention Is All You Need' paper"""
+
+    # Ignore copy
+    def __init__(self, config: GemmaConfig, layer_idx: Optional[int] = None):
+        super().__init__()
+        self.config = config
+        self.layer_idx = layer_idx
+        if layer_idx is None:
+            logger.warning_once(
+                f"Instantiating {self.__class__.__name__} without passing a `layer_idx` is not recommended and will "
+                "lead to errors during the forward call if caching is used. Please make sure to provide a `layer_idx` "
+                "when creating this class."
+            )
+
+        self.attention_dropout = config.attention_dropout
+        self.hidden_size = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = config.head_dim
+        self.num_key_value_heads = config.num_key_value_heads
+        self.num_key_value_groups = self.num_heads // self.num_key_value_heads
+        self.max_position_embeddings = config.max_position_embeddings
+        self.rope_theta = config.rope_theta
+        self.is_causal = True
+
+        if self.hidden_size % self.num_heads != 0:
+            raise ValueError(
+                f"hidden_size must be divisible by num_heads (got `hidden_size`: {self.hidden_size}"
+                f" and `num_heads`: {self.num_heads})."
+            )
+
+        self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=config.attention_bias)
+        self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=config.attention_bias)
+        self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=config.attention_bias)
+        self.o_proj = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=config.attention_bias)
+        self.rotary_emb = GemmaRotaryEmbedding(
+            self.head_dim,
+            max_position_embeddings=self.max_position_embeddings,
+            base=self.rope_theta,
+        )
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_value: Optional[Cache] = None,
+        output_attentions: bool = False,
+        use_cache: bool = False,
+        cache_position: Optional[torch.LongTensor] = None,
+        **kwargs,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        bsz, q_len, _ = hidden_states.size()
+
+        query_states = self.q_proj(hidden_states)
+        key_states = self.k_proj(hidden_states)
+        value_states = self.v_proj(hidden_states)
+
+        query_states = query_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        key_states = key_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+        value_states = value_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+
+        past_key_value = getattr(self, "past_key_value", past_key_value)
+        cos, sin = self.rotary_emb(value_states, position_ids, seq_len=None)
+        query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, None)
+
+        if past_key_value is not None:
+            # sin and cos are specific to RoPE models; position_ids needed for the static cache
+            cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
+            key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)
+
+        key_states = repeat_kv(key_states, self.num_key_value_groups)
+        value_states = repeat_kv(value_states, self.num_key_value_groups)
+
+        attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) / math.sqrt(self.head_dim)
+
+        if attention_mask is not None:  # no matter the length, we just slice it
+            if cache_position is not None:
+                causal_mask = attention_mask[:, :, cache_position, : key_states.shape[-2]]
+            else:
+                causal_mask = attention_mask
+            attn_weights = attn_weights + causal_mask
+
+        # upcast attention to fp32
+        attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)
+        attn_weights = nn.functional.dropout(attn_weights, p=self.attention_dropout, training=self.training)
+        attn_output = torch.matmul(attn_weights, value_states)
+
+        if attn_output.size() != (bsz, self.num_heads, q_len, self.head_dim):
+            raise ValueError(
+                f"`attn_output` should be of size {(bsz, self.num_heads, q_len, self.head_dim)}, but is"
+                f" {attn_output.size()}"
+            )
+
+        attn_output = attn_output.transpose(1, 2).contiguous()
+
+        attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
+        attn_output = self.o_proj(attn_output)
+
+        if not output_attentions:
+            attn_weights = None
+
+        return attn_output, attn_weights, past_key_value
+
+
+# Copied from transformers.models.llama.modeling_llama.LlamaFlashAttention2 with Llama->Gemma
+class GemmaFlashAttention2(GemmaAttention):
+    """
+    Gemma flash attention module. This module inherits from `GemmaAttention` as the weights of the module stays
+    untouched. The only required change would be on the forward pass where it needs to correctly call the public API of
+    flash attention and deal with padding tokens in case the input contains any of them.
+    """
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+
+        # TODO: Should be removed once Flash Attention for RoCm is bumped to 2.1.
+        # flash_attn<2.1 generates top-left aligned causal mask, while what is needed here is bottom-right alignement, that was made default for flash_attn>=2.1. This attribute is used to handle this difference. Reference: https://github.com/Dao-AILab/flash-attention/releases/tag/v2.1.0.
+        # Beware that with flash_attn<2.1, using q_seqlen != k_seqlen (except for the case q_seqlen == 1) produces a wrong mask (top-left).
+        self._flash_attn_uses_top_left_mask = not is_flash_attn_greater_or_equal_2_10()
+
+    # Ignore copy
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_value: Optional[Cache] = None,
+        output_attentions: bool = False,
+        use_cache: bool = False,
+        cache_position: Optional[torch.LongTensor] = None,
+        **kwargs,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        output_attentions = False
+
+        bsz, q_len, _ = hidden_states.size()
+
+        query_states = self.q_proj(hidden_states)
+        key_states = self.k_proj(hidden_states)
+        value_states = self.v_proj(hidden_states)
+
+        # Flash attention requires the input to have the shape
+        # batch_size x seq_length x head_dim x hidden_dim
+        # therefore we just need to keep the original shape
+        query_states = query_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        key_states = key_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+        value_states = value_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+
+        cos, sin = self.rotary_emb(value_states, position_ids, seq_len=None)
+        query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, None)
+
+        past_key_value = getattr(self, "past_key_value", past_key_value)
+
+        if past_key_value is not None:
+            # sin and cos are specific to RoPE models; position_ids needed for the static cache
+            cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
+            key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)
+
+        # TODO: These transpose are quite inefficient but Flash Attention requires the layout [batch_size, sequence_length, num_heads, head_dim]. We would need to refactor the KV cache
+        # to be able to avoid many of these transpose/reshape/view.
+        query_states = query_states.transpose(1, 2)
+        key_states = key_states.transpose(1, 2)
+        value_states = value_states.transpose(1, 2)
+
+        dropout_rate = self.attention_dropout if self.training else 0.0
+
+        # In PEFT, usually we cast the layer norms in float32 for training stability reasons
+        # therefore the input hidden states gets silently casted in float32. Hence, we need
+        # cast them back in the correct dtype just to be sure everything works as expected.
+        # This might slowdown training & inference so it is recommended to not cast the LayerNorms
+        # in fp32. (GemmaRMSNorm handles it correctly)
+
+        input_dtype = query_states.dtype
+        if input_dtype == torch.float32:
+            if torch.is_autocast_enabled():
+                target_dtype = torch.get_autocast_gpu_dtype()
+            # Handle the case where the model is quantized
+            elif hasattr(self.config, "_pre_quantization_dtype"):
+                target_dtype = self.config._pre_quantization_dtype
+            else:
+                target_dtype = self.q_proj.weight.dtype
+
+            logger.warning_once(
+                f"The input hidden states seems to be silently casted in float32, this might be related to"
+                f" the fact you have upcasted embedding or layer norm layers in float32. We will cast back the input in"
+                f" {target_dtype}."
+            )
+
+            query_states = query_states.to(target_dtype)
+            key_states = key_states.to(target_dtype)
+            value_states = value_states.to(target_dtype)
+
+        attn_output = self._flash_attention_forward(
+            query_states, key_states, value_states, attention_mask, q_len, dropout=dropout_rate
+        )
+
+        attn_output = attn_output.reshape(bsz, q_len, -1).contiguous()
+        attn_output = self.o_proj(attn_output)
+
+        if not output_attentions:
+            attn_weights = None
+
+        return attn_output, attn_weights, past_key_value
+
+    def _flash_attention_forward(
+        self, query_states, key_states, value_states, attention_mask, query_length, dropout=0.0, softmax_scale=None
+    ):
+        """
+        Calls the forward method of Flash Attention - if the input hidden states contain at least one padding token
+        first unpad the input, then computes the attention scores and pad the final attention scores.
+
+        Args:
+            query_states (`torch.Tensor`):
+                Input query states to be passed to Flash Attention API
+            key_states (`torch.Tensor`):
+                Input key states to be passed to Flash Attention API
+            value_states (`torch.Tensor`):
+                Input value states to be passed to Flash Attention API
+            attention_mask (`torch.Tensor`):
+                The padding mask - corresponds to a tensor of size `(batch_size, seq_len)` where 0 stands for the
+                position of padding tokens and 1 for the position of non-padding tokens.
+            dropout (`int`, *optional*):
+                Attention dropout
+            softmax_scale (`float`, *optional*):
+                The scaling of QK^T before applying softmax. Default to 1 / sqrt(head_dim)
+        """
+        if not self._flash_attn_uses_top_left_mask:
+            causal = self.is_causal
+        else:
+            # TODO: Remove the `query_length != 1` check once Flash Attention for RoCm is bumped to 2.1. For details, please see the comment in GemmaFlashAttention2 __init__.
+            causal = self.is_causal and query_length != 1
+
+        # Contains at least one padding token in the sequence
+        if attention_mask is not None:
+            batch_size = query_states.shape[0]
+            query_states, key_states, value_states, indices_q, cu_seq_lens, max_seq_lens = self._upad_input(
+                query_states, key_states, value_states, attention_mask, query_length
+            )
+
+            cu_seqlens_q, cu_seqlens_k = cu_seq_lens
+            max_seqlen_in_batch_q, max_seqlen_in_batch_k = max_seq_lens
+
+            attn_output_unpad = flash_attn_varlen_func(
+                query_states,
+                key_states,
+                value_states,
+                cu_seqlens_q=cu_seqlens_q,
+                cu_seqlens_k=cu_seqlens_k,
+                max_seqlen_q=max_seqlen_in_batch_q,
+                max_seqlen_k=max_seqlen_in_batch_k,
+                dropout_p=dropout,
+                softmax_scale=softmax_scale,
+                causal=causal,
+            )
+
+            attn_output = pad_input(attn_output_unpad, indices_q, batch_size, query_length)
+        else:
+            attn_output = flash_attn_func(
+                query_states, key_states, value_states, dropout, softmax_scale=softmax_scale, causal=causal
+            )
+
+        return attn_output
+
+    def _upad_input(self, query_layer, key_layer, value_layer, attention_mask, query_length):
+        indices_k, cu_seqlens_k, max_seqlen_in_batch_k = _get_unpad_data(attention_mask)
+        batch_size, kv_seq_len, num_key_value_heads, head_dim = key_layer.shape
+
+        key_layer = index_first_axis(
+            key_layer.reshape(batch_size * kv_seq_len, num_key_value_heads, head_dim), indices_k
+        )
+        value_layer = index_first_axis(
+            value_layer.reshape(batch_size * kv_seq_len, num_key_value_heads, head_dim), indices_k
+        )
+        if query_length == kv_seq_len:
+            query_layer = index_first_axis(
+                query_layer.reshape(batch_size * kv_seq_len, self.num_heads, head_dim), indices_k
+            )
+            cu_seqlens_q = cu_seqlens_k
+            max_seqlen_in_batch_q = max_seqlen_in_batch_k
+            indices_q = indices_k
+        elif query_length == 1:
+            max_seqlen_in_batch_q = 1
+            cu_seqlens_q = torch.arange(
+                batch_size + 1, dtype=torch.int32, device=query_layer.device
+            )  # There is a memcpy here, that is very bad.
+            indices_q = cu_seqlens_q[:-1]
+            query_layer = query_layer.squeeze(1)
+        else:
+            # The -q_len: slice assumes left padding.
+            attention_mask = attention_mask[:, -query_length:]
+            query_layer, indices_q, cu_seqlens_q, max_seqlen_in_batch_q = unpad_input(query_layer, attention_mask)
+
+        return (
+            query_layer,
+            key_layer,
+            value_layer,
+            indices_q,
+            (cu_seqlens_q, cu_seqlens_k),
+            (max_seqlen_in_batch_q, max_seqlen_in_batch_k),
+        )
+
+
+# Copied from transformers.models.llama.modeling_llama.LlamaSdpaAttention with Llama->Gemma
+class GemmaSdpaAttention(GemmaAttention):
+    """
+    Gemma attention module using torch.nn.functional.scaled_dot_product_attention. This module inherits from
+    `GemmaAttention` as the weights of the module stays untouched. The only changes are on the forward pass to adapt to
+    SDPA API.
+    """
+
+    # Ignore copy
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_value: Optional[Cache] = None,
+        output_attentions: bool = False,
+        use_cache: bool = False,
+        cache_position: Optional[torch.LongTensor] = None,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        if output_attentions:
+            # TODO: Improve this warning with e.g. `model.config.attn_implementation = "manual"` once this is implemented.
+            logger.warning_once(
+                "GemmaModel is using GemmaSdpaAttention, but `torch.nn.functional.scaled_dot_product_attention` does not support `output_attentions=True`. Falling back to the manual attention implementation, "
+                'but specifying the manual implementation will be required from Transformers version v5.0.0 onwards. This warning can be removed using the argument `attn_implementation="eager"` when loading the model.'
+            )
+            return super().forward(
+                hidden_states=hidden_states,
+                attention_mask=attention_mask,
+                position_ids=position_ids,
+                past_key_value=past_key_value,
+                output_attentions=output_attentions,
+                use_cache=use_cache,
+                cache_position=cache_position,
+            )
+
+        bsz, q_len, _ = hidden_states.size()
+
+        query_states = self.q_proj(hidden_states)
+        key_states = self.k_proj(hidden_states)
+        value_states = self.v_proj(hidden_states)
+
+        query_states = query_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        key_states = key_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+        value_states = value_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+
+        cos, sin = self.rotary_emb(value_states, position_ids, seq_len=None)
+        query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, None)
+
+        past_key_value = getattr(self, "past_key_value", past_key_value)
+
+        if past_key_value is not None:
+            # sin and cos are specific to RoPE models; position_ids needed for the static cache
+            cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
+            key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)
+
+        key_states = repeat_kv(key_states, self.num_key_value_groups)
+        value_states = repeat_kv(value_states, self.num_key_value_groups)
+
+        causal_mask = attention_mask
+        if attention_mask is not None and cache_position is not None:
+            causal_mask = causal_mask[:, :, cache_position, : key_states.shape[-2]]
+
+        # SDPA with memory-efficient backend is currently (torch==2.1.2) bugged with non-contiguous inputs with custom attn_mask,
+        # Reference: https://github.com/pytorch/pytorch/issues/112577.
+        if query_states.device.type == "cuda" and causal_mask is not None:
+            query_states = query_states.contiguous()
+            key_states = key_states.contiguous()
+            value_states = value_states.contiguous()
+
+        attn_output = torch.nn.functional.scaled_dot_product_attention(
+            query_states,
+            key_states,
+            value_states,
+            attn_mask=causal_mask,
+            dropout_p=self.attention_dropout if self.training else 0.0,
+        )
+
+        attn_output = attn_output.transpose(1, 2).contiguous()
+        attn_output = attn_output.view(bsz, q_len, -1)
+
+        attn_output = self.o_proj(attn_output)
+
+        return attn_output, None, past_key_value
+
+
+GEMMA_ATTENTION_CLASSES = {
+    "eager": GemmaAttention,
+    "flash_attention_2": GemmaFlashAttention2,
+    "sdpa": GemmaSdpaAttention,
+}
+
+
+# Copied from transformers.models.llama.modeling_llama.LlamaDecoderLayer with LLAMA->GEMMA,Llama->Gemma
+class GemmaDecoderLayer(nn.Module):
+    def __init__(self, config: GemmaConfig, layer_idx: int):
+        super().__init__()
+        self.hidden_size = config.hidden_size
+
+        self.self_attn = GEMMA_ATTENTION_CLASSES[config._attn_implementation](config=config, layer_idx=layer_idx)
+
+        self.mlp = GemmaMLP(config)
+        self.input_layernorm = GemmaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.post_attention_layernorm = GemmaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_value: Optional[Tuple[torch.Tensor]] = None,
+        output_attentions: Optional[bool] = False,
+        use_cache: Optional[bool] = False,
+        cache_position: Optional[torch.LongTensor] = None,
+        **kwargs,
+    ) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
+        """
+        Args:
+            hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)`
+            attention_mask (`torch.FloatTensor`, *optional*):
+                attention mask of size `(batch_size, sequence_length)` if flash attention is used or `(batch_size, 1,
+                query_sequence_length, key_sequence_length)` if default attention is used.
+            output_attentions (`bool`, *optional*):
+                Whether or not to return the attentions tensors of all attention layers. See `attentions` under
+                returned tensors for more detail.
+            use_cache (`bool`, *optional*):
+                If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding
+                (see `past_key_values`).
+            past_key_value (`Tuple(torch.FloatTensor)`, *optional*): cached past key and value projection states
+        """
+        if "padding_mask" in kwargs:
+            warnings.warn(
+                "Passing `padding_mask` is deprecated and will be removed in v4.37. Please make sure use `attention_mask` instead.`"
+            )
+
+        residual = hidden_states
+
+        hidden_states = self.input_layernorm(hidden_states)
+
+        # Self Attention
+        hidden_states, self_attn_weights, present_key_value = self.self_attn(
+            hidden_states=hidden_states,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_value=past_key_value,
+            output_attentions=output_attentions,
+            use_cache=use_cache,
+            cache_position=cache_position,
+            **kwargs,
+        )
+        hidden_states = residual + hidden_states
+
+        # Fully Connected
+        residual = hidden_states
+        hidden_states = self.post_attention_layernorm(hidden_states)
+        hidden_states = self.mlp(hidden_states)
+        hidden_states = residual + hidden_states
+
+        outputs = (hidden_states,)
+
+        if output_attentions:
+            outputs += (self_attn_weights,)
+
+        if use_cache:
+            outputs += (present_key_value,)
+
+        return outputs
+
+
+GEMMA_START_DOCSTRING = r"""
+    This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
+    library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
+    etc.)
+
+    This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
+    Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
+    and behavior.
+
+    Parameters:
+        config ([`GemmaConfig`]):
+            Model configuration class with all the parameters of the model. Initializing with a config file does not
+            load the weights associated with the model, only the configuration. Check out the
+            [`~PreTrainedModel.from_pretrained`] method to load the model weights.
+"""
+
+
+@add_start_docstrings(
+    "The bare Gemma Model outputting raw hidden-states without any specific head on top.",
+    GEMMA_START_DOCSTRING,
+)
+class GemmaPreTrainedModel(PreTrainedModel):
+    config_class = GemmaConfig
+    base_model_prefix = "model"
+    supports_gradient_checkpointing = True
+    _keep_in_fp32_modules = ["inv_freq", "rotary_emb", "cos_cached", "sin_cached"]
+    _no_split_modules = ["GemmaDecoderLayer"]
+    _skip_keys_device_placement = ["past_key_values", "causal_mask"]
+    _supports_flash_attn_2 = True
+    _supports_sdpa = True
+    _supports_cache_class = True
+
+    def _init_weights(self, module):
+        std = self.config.initializer_range
+        if isinstance(module, nn.Linear):
+            module.weight.data.normal_(mean=0.0, std=std)
+            if module.bias is not None:
+                module.bias.data.zero_()
+        elif isinstance(module, nn.Embedding):
+            module.weight.data.normal_(mean=0.0, std=std)
+            if module.padding_idx is not None:
+                module.weight.data[module.padding_idx].zero_()
+
+    def _setup_cache(self, cache_cls, max_batch_size, max_cache_len: Optional[int] = None):
+        if self.config._attn_implementation == "flash_attention_2" and cache_cls == StaticCache:
+            raise ValueError(
+                "`static` cache implementation is not compatible with `attn_implementation==flash_attention_2` "
+                "make sure to use `sdpa` in the mean time, and open an issue at https://github.com/huggingface/transformers"
+            )
+
+        if max_cache_len > self.model.causal_mask.shape[-1] or self.device != self.model.causal_mask.device:
+            causal_mask = torch.full((max_cache_len, max_cache_len), fill_value=1, device=self.device)
+            self.register_buffer("causal_mask", torch.triu(causal_mask, diagonal=1), persistent=False)
+
+        for layer in self.model.layers:
+            weights = layer.self_attn.o_proj.weight
+            layer.self_attn.past_key_value = cache_cls(
+                self.config, max_batch_size, max_cache_len, device=weights.device, dtype=weights.dtype
+            )
+
+    def _reset_cache(self):
+        for layer in self.model.layers:
+            layer.self_attn.past_key_value = None
+
+
+GEMMA_INPUTS_DOCSTRING = r"""
+    Args:
+        input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
+            Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
+            it.
+
+            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            [What are input IDs?](../glossary#input-ids)
+        attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
+
+            - 1 for tokens that are **not masked**,
+            - 0 for tokens that are **masked**.
+
+            [What are attention masks?](../glossary#attention-mask)
+
+            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            If `past_key_values` is used, optionally only the last `input_ids` have to be input (see
+            `past_key_values`).
+
+            If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`]
+            and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more
+            information on the default strategy.
+
+            - 1 indicates the head is **not masked**,
+            - 0 indicates the head is **masked**.
+        position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
+            config.n_positions - 1]`.
+
+            [What are position IDs?](../glossary#position-ids)
+        past_key_values (`Cache` or `tuple(tuple(torch.FloatTensor))`, *optional*):
+            Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
+            blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values`
+            returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`.
+
+            Two formats are allowed:
+            - a [`~cache_utils.Cache`] instance;
+            - Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of
+            shape `(batch_size, num_heads, sequence_length, embed_size_per_head)`). This is also known as the legacy
+            cache format.
+
+            The model will output the same cache format that is fed as input. If no `past_key_values` are passed, the
+            legacy cache format will be returned.
+
+            If `past_key_values` are used, the user can optionally input only the last `input_ids` (those that don't
+            have their past key value states given to this model) of shape `(batch_size, 1)` instead of all `input_ids`
+            of shape `(batch_size, sequence_length)`.
+        inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
+            Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
+            is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
+            model's internal embedding lookup matrix.
+        use_cache (`bool`, *optional*):
+            If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see
+            `past_key_values`).
+        output_attentions (`bool`, *optional*):
+            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
+            tensors for more detail.
+        output_hidden_states (`bool`, *optional*):
+            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
+            more detail.
+        return_dict (`bool`, *optional*):
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+"""
+
+
+@add_start_docstrings(
+    "The bare Gemma Model outputting raw hidden-states without any specific head on top.",
+    GEMMA_START_DOCSTRING,
+)
+# Copied from transformers.models.llama.modeling_llama.LlamaModel with LLAMA->GEMMA,Llama->Gemma
+class GemmaModel(GemmaPreTrainedModel):
+    """
+    Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`GemmaDecoderLayer`]
+
+    Args:
+        config: GemmaConfig
+    """
+
+    def __init__(self, config: GemmaConfig):
+        super().__init__(config)
+        self.padding_idx = config.pad_token_id
+        self.vocab_size = config.vocab_size
+
+        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
+        self.layers = nn.ModuleList(
+            [GemmaDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
+        )
+        self.norm = GemmaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.gradient_checkpointing = False
+
+        # register a causal mask to separate causal and padding mask creation. Merging happends in the attention class
+        causal_mask = torch.full((config.max_position_embeddings, config.max_position_embeddings), fill_value=1)
+        self.register_buffer("causal_mask", torch.triu(causal_mask, diagonal=1), persistent=False)
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.embed_tokens = value
+
+    @add_start_docstrings_to_model_forward(GEMMA_INPUTS_DOCSTRING)
+    # Ignore copy
+    def forward(
+        self,
+        input_ids: torch.LongTensor = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[List[torch.FloatTensor]] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        cache_position: Optional[torch.LongTensor] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPast]:
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        use_cache = use_cache if use_cache is not None else self.config.use_cache
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        if (input_ids is None) ^ (inputs_embeds is not None):
+            raise ValueError(
+                "You cannot specify both input_ids and inputs_embeds at the same time, and must specify either one"
+            )
+
+        if self.gradient_checkpointing and self.training and use_cache:
+            logger.warning_once(
+                "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`."
+            )
+            use_cache = False
+
+        if inputs_embeds is None:
+            inputs_embeds = self.embed_tokens(input_ids)
+
+        past_seen_tokens = 0
+        if use_cache:  # kept for BC (cache positions)
+            if not isinstance(past_key_values, StaticCache):
+                past_key_values = DynamicCache.from_legacy_cache(past_key_values)
+            past_seen_tokens = past_key_values.get_seq_length()
+
+        if cache_position is None:
+            cache_position = torch.arange(
+                past_seen_tokens, past_seen_tokens + inputs_embeds.shape[1], device=inputs_embeds.device
+            )
+
+        if position_ids is None:
+            position_ids = cache_position.unsqueeze(0)
+
+        causal_mask = self._update_causal_mask(attention_mask, inputs_embeds)
+
+        # embed positions
+        hidden_states = inputs_embeds
+
+        # normalized
+        hidden_states = hidden_states * (self.config.hidden_size**0.5)
+
+        # decoder layers
+        all_hidden_states = () if output_hidden_states else None
+        all_self_attns = () if output_attentions else None
+        next_decoder_cache = None
+
+        for decoder_layer in self.layers:
+            if output_hidden_states:
+                all_hidden_states += (hidden_states,)
+
+            if self.gradient_checkpointing and self.training:
+                layer_outputs = self._gradient_checkpointing_func(
+                    decoder_layer.__call__,
+                    hidden_states,
+                    causal_mask,
+                    position_ids,
+                    past_key_values,
+                    output_attentions,
+                    use_cache,
+                    cache_position,
+                )
+            else:
+                layer_outputs = decoder_layer(
+                    hidden_states,
+                    attention_mask=causal_mask,
+                    position_ids=position_ids,
+                    past_key_value=past_key_values,
+                    output_attentions=output_attentions,
+                    use_cache=use_cache,
+                    cache_position=cache_position,
+                )
+
+            hidden_states = layer_outputs[0]
+
+            if use_cache:
+                next_decoder_cache = layer_outputs[2 if output_attentions else 1]
+
+            if output_attentions:
+                all_self_attns += (layer_outputs[1],)
+
+        hidden_states = self.norm(hidden_states)
+
+        # add hidden states from the last decoder layer
+        if output_hidden_states:
+            all_hidden_states += (hidden_states,)
+
+        next_cache = None
+        if use_cache:
+            next_cache = (
+                next_decoder_cache.to_legacy_cache() if isinstance(next_decoder_cache, Cache) else next_decoder_cache
+            )
+        if not return_dict:
+            return tuple(v for v in [hidden_states, next_cache, all_hidden_states, all_self_attns] if v is not None)
+        return BaseModelOutputWithPast(
+            last_hidden_state=hidden_states,
+            past_key_values=next_cache,
+            hidden_states=all_hidden_states,
+            attentions=all_self_attns,
+        )
+
+    def _update_causal_mask(self, attention_mask, input_tensor):
+        if self.config._attn_implementation == "flash_attention_2":
+            if attention_mask is not None and 0.0 in attention_mask:
+                return attention_mask
+            return None
+
+        batch_size, seq_length = input_tensor.shape[:2]
+        dtype = input_tensor.dtype
+        device = input_tensor.device
+
+        # support going beyond cached `max_position_embedding`
+        if seq_length > self.causal_mask.shape[-1]:
+            causal_mask = torch.full((2 * self.causal_mask.shape[-1], 2 * self.causal_mask.shape[-1]), fill_value=1)
+            self.register_buffer("causal_mask", torch.triu(causal_mask, diagonal=1), persistent=False)
+
+        if hasattr(self, "causal_mask"):  # we use the current dtype to avoid any overflows
+            causal_mask = (
+                self.causal_mask[None, None, :, :].repeat(batch_size, 1, 1, 1).to(dtype) * torch.finfo(dtype).min
+            )
+        else:
+            mask = torch.full(
+                (self.config.max_position_embeddings, self.config.max_position_embeddings),
+                fill_value=torch.finfo(dtype).min,
+            )
+            causal_mask = torch.triu(mask, diagonal=1)
+
+        causal_mask = causal_mask.to(dtype=dtype, device=device)
+        if attention_mask is not None and attention_mask.dim() == 2:
+            mask_length = attention_mask.shape[-1]
+            padding_mask = causal_mask[..., :mask_length].eq(0.0) * attention_mask[:, None, None, :].eq(0.0)
+            causal_mask[..., :mask_length] = causal_mask[..., :mask_length].masked_fill(
+                padding_mask, torch.finfo(dtype).min
+            )
+
+        if self.config._attn_implementation == "sdpa":
+            is_tracing = torch.jit.is_tracing() or isinstance(input_tensor, torch.fx.Proxy)
+            if not is_tracing and attention_mask is not None and torch.any(attention_mask != 1):
+                causal_mask = causal_mask.mul(~torch.all(causal_mask == causal_mask.min(), dim=-1)[..., None]).to(
+                    dtype
+                )
+
+        return causal_mask
+
+
+# Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM with LLAMA->GEMMA,Llama->Gemma,llama->gemma
+class GemmaForCausalLM(GemmaPreTrainedModel):
+    _tied_weights_keys = ["lm_head.weight"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.model = GemmaModel(config)
+        self.vocab_size = config.vocab_size
+        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.model.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.model.embed_tokens = value
+
+    def get_output_embeddings(self):
+        return self.lm_head
+
+    def set_output_embeddings(self, new_embeddings):
+        self.lm_head = new_embeddings
+
+    def set_decoder(self, decoder):
+        self.model = decoder
+
+    def get_decoder(self):
+        return self.model
+
+    # Ignore copy
+    @add_start_docstrings_to_model_forward(GEMMA_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=CausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC)
+    def forward(
+        self,
+        input_ids: torch.LongTensor = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[List[torch.FloatTensor]] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        cache_position: Optional[torch.LongTensor] = None,
+    ) -> Union[Tuple, CausalLMOutputWithPast]:
+        r"""
+        Args:
+            labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+                Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
+                config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
+                (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.
+
+        Returns:
+
+        Example:
+
+        ```python
+        >>> from transformers import AutoTokenizer, GemmaForCausalLM
+
+        >>> model = GemmaForCausalLM.from_pretrained("google/gemma-7b")
+        >>> tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b")
+
+        >>> prompt = "What is your favorite condiment?"
+        >>> inputs = tokenizer(prompt, return_tensors="pt")
+
+        >>> # Generate
+        >>> generate_ids = model.generate(inputs.input_ids, max_length=30)
+        >>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
+        "What is your favorite condiment?"
+        ```"""
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
+        outputs = self.model(
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+            cache_position=cache_position,
+        )
+
+        hidden_states = outputs[0]
+        logits = self.lm_head(hidden_states)
+
+        loss = None
+        if labels is not None:
+            # Shift so that tokens < n predict n
+            shift_logits = logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            # Flatten the tokens
+            loss_fct = CrossEntropyLoss()
+            shift_logits = shift_logits.view(-1, self.config.vocab_size)
+            shift_labels = shift_labels.view(-1)
+            # Enable model parallelism
+            shift_labels = shift_labels.to(shift_logits.device)
+            loss = loss_fct(shift_logits, shift_labels)
+
+        if not return_dict:
+            output = (logits,) + outputs[1:]
+            return (loss,) + output if loss is not None else output
+
+        return CausalLMOutputWithPast(
+            loss=loss,
+            logits=logits,
+            past_key_values=outputs.past_key_values,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+    def prepare_inputs_for_generation(
+        self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs
+    ):
+        past_length = 0
+        if past_key_values is not None:
+            if isinstance(past_key_values, Cache):
+                cache_length = past_key_values.get_seq_length()
+                past_length = past_key_values.seen_tokens
+                max_cache_length = past_key_values.get_max_length()
+            else:
+                cache_length = past_length = past_key_values[0][0].shape[2]
+                max_cache_length = None
+
+            # Keep only the unprocessed tokens:
+            # 1 - If the length of the attention_mask exceeds the length of input_ids, then we are in a setting where
+            # some of the inputs are exclusively passed as part of the cache (e.g. when passing input_embeds as
+            # input)
+            if attention_mask is not None and attention_mask.shape[1] > input_ids.shape[1]:
+                input_ids = input_ids[:, -(attention_mask.shape[1] - past_length) :]
+            # 2 - If the past_length is smaller than input_ids', then input_ids holds all input tokens. We can discard
+            # input_ids based on the past_length.
+            elif past_length < input_ids.shape[1]:
+                input_ids = input_ids[:, past_length:]
+            # 3 - Otherwise (past_length >= input_ids.shape[1]), let's assume input_ids only has unprocessed tokens.
+
+            # If we are about to go beyond the maximum cache length, we need to crop the input attention mask.
+            if (
+                max_cache_length is not None
+                and attention_mask is not None
+                and cache_length + input_ids.shape[1] > max_cache_length
+            ):
+                attention_mask = attention_mask[:, -max_cache_length:]
+
+        position_ids = kwargs.get("position_ids", None)
+        if attention_mask is not None and position_ids is None:
+            # create position_ids on the fly for batch generation
+            position_ids = attention_mask.long().cumsum(-1) - 1
+            position_ids.masked_fill_(attention_mask == 0, 1)
+            if past_key_values:
+                position_ids = position_ids[:, -input_ids.shape[1] :]
+
+        if past_key_value := getattr(self.model.layers[0].self_attn, "past_key_value", None):
+            # generation with static cache
+            past_length = past_key_value.get_seq_length()
+            input_ids = input_ids[:, past_length:]
+            position_ids = position_ids[:, past_length:]
+
+        # TODO @gante we should only keep a `cache_position` in generate, and do +=1.
+        # same goes for position ids. Could also help with continued generation.
+        cache_position = kwargs.get("cache_position", None)
+        if cache_position is None:
+            cache_position = torch.arange(
+                past_length, past_length + position_ids.shape[-1], device=position_ids.device
+            )
+
+        # if `inputs_embeds` are passed, we only want to use them in the 1st generation step
+        if inputs_embeds is not None and past_key_values is None:
+            model_inputs = {"inputs_embeds": inputs_embeds}
+        else:
+            model_inputs = {"input_ids": input_ids}
+
+        model_inputs.update(
+            {
+                "position_ids": position_ids,
+                "cache_position": cache_position,
+                "past_key_values": past_key_values,
+                "use_cache": kwargs.get("use_cache"),
+                "attention_mask": attention_mask,
+            }
+        )
+        return model_inputs
+
+    @staticmethod
+    def _reorder_cache(past_key_values, beam_idx):
+        reordered_past = ()
+        for layer_past in past_key_values:
+            reordered_past += (
+                tuple(past_state.index_select(0, beam_idx.to(past_state.device)) for past_state in layer_past),
+            )
+        return reordered_past
+
+
+@add_start_docstrings(
+    """
+    The Gemma Model transformer with a sequence classification head on top (linear layer).
+
+    [`GemmaForSequenceClassification`] uses the last token in order to do the classification, as other causal models
+    (e.g. GPT-2) do.
+
+    Since it does classification on the last token, it requires to know the position of the last token. If a
+    `pad_token_id` is defined in the configuration, it finds the last token that is not a padding token in each row. If
+    no `pad_token_id` is defined, it simply takes the last value in each row of the batch. Since it cannot guess the
+    padding tokens when `inputs_embeds` are passed instead of `input_ids`, it does the same (take the last value in
+    each row of the batch).
+    """,
+    GEMMA_START_DOCSTRING,
+)
+# Copied from transformers.models.llama.modeling_llama.LlamaForSequenceClassification with LLAMA->GEMMA,Llama->Gemma
+class GemmaForSequenceClassification(GemmaPreTrainedModel):
+    def __init__(self, config):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+        self.model = GemmaModel(config)
+        self.score = nn.Linear(config.hidden_size, self.num_labels, bias=False)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.model.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.model.embed_tokens = value
+
+    @add_start_docstrings_to_model_forward(GEMMA_INPUTS_DOCSTRING)
+    def forward(
+        self,
+        input_ids: torch.LongTensor = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[List[torch.FloatTensor]] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, SequenceClassifierOutputWithPast]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
+            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
+            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        transformer_outputs = self.model(
+            input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        hidden_states = transformer_outputs[0]
+        logits = self.score(hidden_states)
+
+        if input_ids is not None:
+            batch_size = input_ids.shape[0]
+        else:
+            batch_size = inputs_embeds.shape[0]
+
+        if self.config.pad_token_id is None and batch_size != 1:
+            raise ValueError("Cannot handle batch sizes > 1 if no padding token is defined.")
+        if self.config.pad_token_id is None:
+            sequence_lengths = -1
+        else:
+            if input_ids is not None:
+                # if no pad token found, use modulo instead of reverse indexing for ONNX compatibility
+                sequence_lengths = torch.eq(input_ids, self.config.pad_token_id).int().argmax(-1) - 1
+                sequence_lengths = sequence_lengths % input_ids.shape[-1]
+                sequence_lengths = sequence_lengths.to(logits.device)
+            else:
+                sequence_lengths = -1
+
+        pooled_logits = logits[torch.arange(batch_size, device=logits.device), sequence_lengths]
+
+        loss = None
+        if labels is not None:
+            labels = labels.to(logits.device)
+            if self.config.problem_type is None:
+                if self.num_labels == 1:
+                    self.config.problem_type = "regression"
+                elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int):
+                    self.config.problem_type = "single_label_classification"
+                else:
+                    self.config.problem_type = "multi_label_classification"
+
+            if self.config.problem_type == "regression":
+                loss_fct = MSELoss()
+                if self.num_labels == 1:
+                    loss = loss_fct(pooled_logits.squeeze(), labels.squeeze())
+                else:
+                    loss = loss_fct(pooled_logits, labels)
+            elif self.config.problem_type == "single_label_classification":
+                loss_fct = CrossEntropyLoss()
+                loss = loss_fct(pooled_logits.view(-1, self.num_labels), labels.view(-1))
+            elif self.config.problem_type == "multi_label_classification":
+                loss_fct = BCEWithLogitsLoss()
+                loss = loss_fct(pooled_logits, labels)
+        if not return_dict:
+            output = (pooled_logits,) + transformer_outputs[1:]
+            return ((loss,) + output) if loss is not None else output
+
+        return SequenceClassifierOutputWithPast(
+            loss=loss,
+            logits=pooled_logits,
+            past_key_values=transformer_outputs.past_key_values,
+            hidden_states=transformer_outputs.hidden_states,
+            attentions=transformer_outputs.attentions,
+        )

src/transformers/models/gemma/tokenization_gemma.py

+# coding=utf-8
+# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Tokenization classes for Gemma."""
+import os
+from shutil import copyfile
+from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple
+
+import sentencepiece as spm
+
+from ...tokenization_utils import AddedToken, PreTrainedTokenizer
+from ...utils import logging
+
+
+if TYPE_CHECKING:
+    pass
+
+logger = logging.get_logger(__name__)
+
+VOCAB_FILES_NAMES = {"vocab_file": "tokenizer.model"}
+
+SPIECE_UNDERLINE = "▁"
+
+
+class GemmaTokenizer(PreTrainedTokenizer):
+    """
+    Construct a Gemma tokenizer. Based on byte-level Byte-Pair-Encoding. The default padding token is unset as there is
+    no padding token in the original model.
+
+    Args:
+        vocab_file (`str`):
+            Path to the vocabulary file.
+        unk_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<unk>"`):
+            The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this
+            token instead.
+        bos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<bos>"`):
+            The beginning of sequence token that was used during pretraining. Can be used a sequence classifier token.
+        eos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<eos>"`):
+            The end of sequence token.
+        pad_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<pad>"`):
+            A special token used to make arrays of tokens the same size for batching purpose. Will then be ignored by
+            attention mechanisms or loss computation.
+        sp_model_kwargs (`Dict[str, Any]`, `Optional`, *optional*):
+            Will be passed to the `SentencePieceProcessor.__init__()` method. The [Python wrapper for
+            SentencePiece](https://github.com/google/sentencepiece/tree/master/python) can be used, among other things,
+            to set:
+
+            - `enable_sampling`: Enable subword regularization.
+            - `nbest_size`: Sampling parameters for unigram. Invalid for BPE-Dropout.
+
+              - `nbest_size = {0,1}`: No sampling is performed.
+              - `nbest_size > 1`: samples from the nbest_size results.
+              - `nbest_size < 0`: assuming that nbest_size is infinite and samples from the all hypothesis (lattice)
+                using forward-filtering-and-backward-sampling algorithm.
+
+            - `alpha`: Smoothing parameter for unigram sampling, and dropout probability of merge operations for
+              BPE-dropout.
+
+        add_bos_token (`bool`, *optional*, defaults to `True`):
+            Whether or not to add an `bos_token` at the start of sequences.
+        add_eos_token (`bool`, *optional*, defaults to `False`):
+            Whether or not to add an `eos_token` at the end of sequences.
+        clean_up_tokenization_spaces (`bool`, *optional*, defaults to `False`):
+            Whether or not to cleanup spaces after decoding, cleanup consists in removing potential artifacts like
+            extra spaces.
+        use_default_system_prompt (`bool`, *optional*, defaults to `False`):
+            Whether or not the default system prompt for Gemma should be used.
+        spaces_between_special_tokens (`bool`, *optional*, defaults to `False`):
+            Whether or not to add spaces between special tokens.
+    """
+
+    vocab_files_names = VOCAB_FILES_NAMES
+    model_input_names = ["input_ids", "attention_mask"]
+
+    def __init__(
+        self,
+        vocab_file,
+        unk_token="<unk>",
+        bos_token="<bos>",
+        eos_token="<eos>",
+        pad_token="<pad>",
+        sp_model_kwargs: Optional[Dict[str, Any]] = None,
+        add_bos_token=True,
+        add_eos_token=False,
+        clean_up_tokenization_spaces=False,
+        use_default_system_prompt=False,
+        spaces_between_special_tokens=False,
+        **kwargs,
+    ):
+        self.sp_model_kwargs = {} if sp_model_kwargs is None else sp_model_kwargs
+        bos_token = AddedToken(bos_token, normalized=False, special=True) if isinstance(bos_token, str) else bos_token
+        eos_token = AddedToken(eos_token, normalized=False, special=True) if isinstance(eos_token, str) else eos_token
+        unk_token = AddedToken(unk_token, normalized=False, special=True) if isinstance(unk_token, str) else unk_token
+        pad_token = AddedToken(pad_token, normalized=False, special=True) if isinstance(pad_token, str) else pad_token
+
+        self.vocab_file = vocab_file
+        self.add_bos_token = add_bos_token
+        self.add_eos_token = add_eos_token
+        self.use_default_system_prompt = use_default_system_prompt
+
+        self.sp_model = spm.SentencePieceProcessor(**self.sp_model_kwargs)
+        self.sp_model.Load(vocab_file)
+
+        super().__init__(
+            bos_token=bos_token,
+            eos_token=eos_token,
+            unk_token=unk_token,
+            pad_token=pad_token,
+            add_bos_token=add_bos_token,
+            add_eos_token=add_eos_token,
+            sp_model_kwargs=self.sp_model_kwargs,
+            clean_up_tokenization_spaces=clean_up_tokenization_spaces,
+            use_default_system_prompt=use_default_system_prompt,
+            spaces_between_special_tokens=spaces_between_special_tokens,
+            **kwargs,
+        )
+
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer.__getstate__
+    def __getstate__(self):
+        state = self.__dict__.copy()
+        state["sp_model"] = None
+        state["sp_model_proto"] = self.sp_model.serialized_model_proto()
+        return state
+
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer.__setstate__
+    def __setstate__(self, d):
+        self.__dict__ = d
+        self.sp_model = spm.SentencePieceProcessor(**self.sp_model_kwargs)
+        self.sp_model.LoadFromSerializedProto(self.sp_model_proto)
+
+    @property
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer.vocab_size
+    def vocab_size(self):
+        """Returns vocab size"""
+        return self.sp_model.get_piece_size()
+
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer.get_vocab
+    def get_vocab(self):
+        """Returns vocab as a dict"""
+        vocab = {self.convert_ids_to_tokens(i): i for i in range(self.vocab_size)}
+        vocab.update(self.added_tokens_encoder)
+        return vocab
+
+    def _tokenize(self, text, **kwargs):
+        """
+        Returns a tokenized string. The Gemma tokenizer never adds a prefix space.
+        """
+        return self.sp_model.encode(text, out_type=str)
+
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer._convert_token_to_id
+    def _convert_token_to_id(self, token):
+        """Converts a token (str) in an id using the vocab."""
+        return self.sp_model.piece_to_id(token)
+
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer._convert_id_to_token
+    def _convert_id_to_token(self, index):
+        """Converts an index (integer) in a token (str) using the vocab."""
+        token = self.sp_model.IdToPiece(index)
+        return token
+
+    def _decode(
+        self,
+        token_ids: List[int],
+        skip_special_tokens: bool = False,
+        spaces_between_special_tokens: bool = False,
+        **kwargs,
+    ) -> str:
+        sub_texts = []
+        current_sub_text = []
+        for ids in token_ids:
+            if skip_special_tokens and ids in self.all_special_ids:
+                continue
+            if ids in self._added_tokens_decoder:
+                if current_sub_text:
+                    sub_texts.append(self.sp_model.decode(current_sub_text))
+                sub_texts.append(self._added_tokens_decoder[ids].content)
+                current_sub_text = []
+            else:
+                current_sub_text.append(ids)
+        if current_sub_text:
+            sub_texts.append(self.sp_model.decode(current_sub_text))
+
+        if spaces_between_special_tokens:
+            sub_texts = " ".join(sub_texts)
+        else:
+            sub_texts = "".join(sub_texts)
+
+        return sub_texts
+
+    def convert_tokens_to_string(self, tokens):
+        """Converts a sequence of tokens (string) in a single string."""
+        current_sub_tokens = []
+        out_string = ""
+        for token in tokens:
+            # make sure that special tokens are not decoded using sentencepiece model
+            if token in self._added_tokens_encoder:
+                out_string += self.sp_model.decode(current_sub_tokens) + token
+                current_sub_tokens = []
+            else:
+                current_sub_tokens.append(token)
+        out_string += self.sp_model.decode(current_sub_tokens)
+        return out_string
+
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer.save_vocabulary
+    def save_vocabulary(self, save_directory, filename_prefix: Optional[str] = None) -> Tuple[str]:
+        """
+        Save the vocabulary and special tokens file to a directory.
+
+        Args:
+            save_directory (`str`):
+                The directory in which to save the vocabulary.
+
+        Returns:
+            `Tuple(str)`: Paths to the files saved.
+        """
+        if not os.path.isdir(save_directory):
+            logger.error(f"Vocabulary path ({save_directory}) should be a directory")
+            return
+        out_vocab_file = os.path.join(
+            save_directory, (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"]
+        )
+
+        if os.path.abspath(self.vocab_file) != os.path.abspath(out_vocab_file) and os.path.isfile(self.vocab_file):
+            copyfile(self.vocab_file, out_vocab_file)
+        elif not os.path.isfile(self.vocab_file):
+            with open(out_vocab_file, "wb") as fi:
+                content_spiece_model = self.sp_model.serialized_model_proto()
+                fi.write(content_spiece_model)
+
+        return (out_vocab_file,)
+
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer.build_inputs_with_special_tokens
+    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
+        bos_token_id = [self.bos_token_id] if self.add_bos_token else []
+        eos_token_id = [self.eos_token_id] if self.add_eos_token else []
+
+        output = bos_token_id + token_ids_0 + eos_token_id
+
+        if token_ids_1 is not None:
+            output = output + bos_token_id + token_ids_1 + eos_token_id
+
+        return output
+
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer.get_special_tokens_mask
+    def get_special_tokens_mask(
+        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False
+    ) -> List[int]:
+        """
+        Retrieve sequence ids from a token list that has no special tokens added. This method is called when adding
+        special tokens using the tokenizer `prepare_for_model` method.
+
+        Args:
+            token_ids_0 (`List[int]`):
+                List of IDs.
+            token_ids_1 (`List[int]`, *optional*):
+                Optional second list of IDs for sequence pairs.
+            already_has_special_tokens (`bool`, *optional*, defaults to `False`):
+                Whether or not the token list is already formatted with special tokens for the model.
+
+        Returns:
+            `List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.
+        """
+        if already_has_special_tokens:
+            return super().get_special_tokens_mask(
+                token_ids_0=token_ids_0, token_ids_1=token_ids_1, already_has_special_tokens=True
+            )
+
+        bos_token_id = [1] if self.add_bos_token else []
+        eos_token_id = [1] if self.add_eos_token else []
+
+        if token_ids_1 is None:
+            return bos_token_id + ([0] * len(token_ids_0)) + eos_token_id
+        return (
+            bos_token_id
+            + ([0] * len(token_ids_0))
+            + eos_token_id
+            + bos_token_id
+            + ([0] * len(token_ids_1))
+            + eos_token_id
+        )
+
+    # Copied from transformers.models.llama.tokenization_llama.LlamaTokenizer.create_token_type_ids_from_sequences
+    def create_token_type_ids_from_sequences(
+        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
+    ) -> List[int]:
+        """
+        Creates a mask from the two sequences passed to be used in a sequence-pair classification task. An ALBERT
+        sequence pair mask has the following format:
+
+        ```
+        0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1
+        | first sequence    | second sequence |
+        ```
+
+        if token_ids_1 is None, only returns the first portion of the mask (0s).
+
+        Args:
+            token_ids_0 (`List[int]`):
+                List of ids.
+            token_ids_1 (`List[int]`, *optional*):
+                Optional second list of IDs for sequence pairs.
+
+        Returns:
+            `List[int]`: List of [token type IDs](../glossary#token-type-ids) according to the given sequence(s).
+        """
+        bos_token_id = [self.bos_token_id] if self.add_bos_token else []
+        eos_token_id = [self.eos_token_id] if self.add_eos_token else []
+
+        output = [0] * len(bos_token_id + token_ids_0 + eos_token_id)
+
+        if token_ids_1 is not None:
+            output += [1] * len(bos_token_id + token_ids_1 + eos_token_id)
+
+        return output

src/transformers/models/gemma/tokenization_gemma_fast.py

+# coding=utf-8
+# Copyright 2024 The HuggingFace Inc. team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import os
+from shutil import copyfile
+from typing import Optional, Tuple
+
+from tokenizers import processors
+
+from ...tokenization_utils_fast import PreTrainedTokenizerFast
+from ...utils import is_sentencepiece_available, logging
+from ...utils.versions import require_version
+
+
+require_version("tokenizers>=0.13.3")
+
+if is_sentencepiece_available():
+    from .tokenization_gemma import GemmaTokenizer
+else:
+    GemmaTokenizer = None
+
+logger = logging.get_logger(__name__)
+VOCAB_FILES_NAMES = {"vocab_file": "tokenizer.model", "tokenizer_file": "tokenizer.json"}
+
+
+class GemmaTokenizerFast(PreTrainedTokenizerFast):
+    """
+    Construct a Gemma tokenizer fast. Based on byte-level Byte-Pair-Encoding.
+
+    This uses notably ByteFallback and no prefix space. Normalization is applied to replace  `" "` with `"▁"`
+
+    ```python
+    >>> from transformers import GemmaTokenizerFast
+
+    >>> tokenizer = GemmaTokenizerFast.from_pretrained("hf-internal-testing/dummy-gemma")
+    >>> tokenizer.encode("Hello this is a test")
+    [2, 4521, 736, 603, 476, 2121]
+    ```
+
+    If you want to change the `bos_token` or the `eos_token`, make sure to specify them when initializing the model, or
+    call `tokenizer.update_post_processor()` to make sure that the post-processing is correctly done (otherwise the
+    values of the first token and final token of an encoded sequence will not be correct). For more details, checkout
+    [post-processors] (https://huggingface.co/docs/tokenizers/api/post-processors) documentation.
+
+
+    This tokenizer inherits from [`PreTrainedTokenizerFast`] which contains most of the main methods. Users should
+    refer to this superclass for more information regarding those methods.
+
+    Args:
+        vocab_file (`str`, *optional*):
+            [SentencePiece](https://github.com/google/sentencepiece) file (generally has a .model extension) that
+            contains the vocabulary necessary to instantiate a tokenizer.
+        tokenizer_file (`str`, *optional*):
+            [tokenizers](https://github.com/huggingface/tokenizers) file (generally has a .json extension) that
+            contains everything needed to load the tokenizer.
+        clean_up_tokenization_spaces (`bool`, *optional*, defaults to `False`):
+            Whether or not to cleanup spaces after decoding, cleanup consists in removing potential artifacts like
+            extra spaces.
+        unk_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<unk>"`):
+            The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this
+            token instead.
+        bos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<bos>"`):
+            The beginning of sequence token that was used during pretraining. Can be used a sequence classifier token.
+        eos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<eos>"`):
+            The end of sequence token.
+        pad_token (`str`, *optional*, defaults to `"<pad>"`):
+            The padding token
+        add_bos_token (`bool`, *optional*, defaults to `True`):
+            Whether or not to add an `bos_token` at the start of sequences.
+        add_eos_token (`bool`, *optional*, defaults to `False`):
+            Whether or not to add an `eos_token` at the end of sequences.
+    """
+
+    vocab_files_names = VOCAB_FILES_NAMES
+    slow_tokenizer_class = GemmaTokenizer
+    padding_side = "left"
+    model_input_names = ["input_ids", "attention_mask"]
+
+    def __init__(
+        self,
+        vocab_file=None,
+        tokenizer_file=None,
+        clean_up_tokenization_spaces=False,
+        unk_token="<unk>",
+        bos_token="<bos>",
+        eos_token="<eos>",
+        pad_token="<pad>",
+        add_bos_token=True,
+        add_eos_token=False,
+        **kwargs,
+    ):
+        super().__init__(
+            vocab_file=vocab_file,
+            tokenizer_file=tokenizer_file,
+            clean_up_tokenization_spaces=clean_up_tokenization_spaces,
+            unk_token=unk_token,
+            bos_token=bos_token,
+            eos_token=eos_token,
+            pad_token=pad_token,
+            add_bos_token=add_bos_token,
+            add_eos_token=add_eos_token,
+            **kwargs,
+        )
+        self._add_bos_token = add_bos_token
+        self._add_eos_token = add_eos_token
+        self.update_post_processor()
+        self.vocab_file = vocab_file
+
+    @property
+    def can_save_slow_tokenizer(self) -> bool:
+        return os.path.isfile(self.vocab_file) if self.vocab_file else False
+
+    # Copied from transformers.models.llama.tokenization_llama_fast.LlamaTokenizerFast.update_post_processor
+    def update_post_processor(self):
+        """
+        Updates the underlying post processor with the current `bos_token` and `eos_token`.
+        """
+        bos = self.bos_token
+        bos_token_id = self.bos_token_id
+        if bos is None and self.add_bos_token:
+            raise ValueError("add_bos_token = True but bos_token = None")
+
+        eos = self.eos_token
+        eos_token_id = self.eos_token_id
+        if eos is None and self.add_eos_token:
+            raise ValueError("add_eos_token = True but eos_token = None")
+
+        single = f"{(bos+':0 ') if self.add_bos_token else ''}$A:0{(' '+eos+':0') if self.add_eos_token else ''}"
+        pair = f"{single}{(' '+bos+':1') if self.add_bos_token else ''} $B:1{(' '+eos+':1') if self.add_eos_token else ''}"
+
+        special_tokens = []
+        if self.add_bos_token:
+            special_tokens.append((bos, bos_token_id))
+        if self.add_eos_token:
+            special_tokens.append((eos, eos_token_id))
+        self._tokenizer.post_processor = processors.TemplateProcessing(
+            single=single, pair=pair, special_tokens=special_tokens
+        )
+
+    @property
+    def add_eos_token(self):
+        return self._add_eos_token
+
+    @property
+    def add_bos_token(self):
+        return self._add_bos_token
+
+    @add_eos_token.setter
+    def add_eos_token(self, value):
+        self._add_eos_token = value
+        self.update_post_processor()
+
+    @add_bos_token.setter
+    def add_bos_token(self, value):
+        self._add_bos_token = value
+        self.update_post_processor()
+
+    # Copied from transformers.models.llama.tokenization_llama_fast.LlamaTokenizerFast.save_vocabulary
+    def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str] = None) -> Tuple[str]:
+        if not self.can_save_slow_tokenizer:
+            raise ValueError(
+                "Your fast tokenizer does not have the necessary information to save the vocabulary for a slow "
+                "tokenizer."
+            )
+
+        if not os.path.isdir(save_directory):
+            logger.error(f"Vocabulary path ({save_directory}) should be a directory")
+            return
+        out_vocab_file = os.path.join(
+            save_directory, (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"]
+        )
+
+        if os.path.abspath(self.vocab_file) != os.path.abspath(out_vocab_file):
+            copyfile(self.vocab_file, out_vocab_file)
+
+        return (out_vocab_file,)
+
+    # Copied from transformers.models.llama.tokenization_llama_fast.LlamaTokenizerFast.build_inputs_with_special_tokens
+    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
+        bos_token_id = [self.bos_token_id] if self.add_bos_token else []
+        eos_token_id = [self.eos_token_id] if self.add_eos_token else []
+
+        output = bos_token_id + token_ids_0 + eos_token_id
+
+        if token_ids_1 is not None:
+            output = output + bos_token_id + token_ids_1 + eos_token_id
+
+        return output

src/transformers/utils/dummy_flax_objects.py

@@ -737,6 +737,27 @@ class FlaxEncoderDecoderModel(metaclass=DummyObject):
         requires_backends(self, ["flax"])
 
 
+class FlaxGemmaForCausalLM(metaclass=DummyObject):
+    _backends = ["flax"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
+
+class FlaxGemmaModel(metaclass=DummyObject):
+    _backends = ["flax"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
+
+class FlaxGemmaPreTrainedModel(metaclass=DummyObject):
+    _backends = ["flax"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
+
 class FlaxGPT2LMHeadModel(metaclass=DummyObject):
     _backends = ["flax"]
 

src/transformers/utils/dummy_pt_objects.py

@@ -3811,6 +3811,34 @@ class FuyuPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
+class GemmaForCausalLM(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class GemmaForSequenceClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class GemmaModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class GemmaPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
 GIT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
 

src/transformers/utils/dummy_sentencepiece_objects.py

@@ -79,6 +79,13 @@ class FNetTokenizer(metaclass=DummyObject):
         requires_backends(self, ["sentencepiece"])
 
 
+class GemmaTokenizer(metaclass=DummyObject):
+    _backends = ["sentencepiece"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["sentencepiece"])
+
+
 class GPTSw3Tokenizer(metaclass=DummyObject):
     _backends = ["sentencepiece"]
 

src/transformers/utils/dummy_tokenizers_objects.py

@@ -170,6 +170,13 @@ class FunnelTokenizerFast(metaclass=DummyObject):
         requires_backends(self, ["tokenizers"])
 
 
+class GemmaTokenizerFast(metaclass=DummyObject):
+    _backends = ["tokenizers"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["tokenizers"])
+
+
 class GPT2TokenizerFast(metaclass=DummyObject):
     _backends = ["tokenizers"]
 

tests/models/gemma/test_modeling_flax_gemma.py

+# Copyright 2024 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+
+import unittest
+
+import numpy as np
+
+from transformers import AutoTokenizer, GemmaConfig, is_flax_available
+from transformers.testing_utils import require_flax, slow
+
+from ...generation.test_flax_utils import FlaxGenerationTesterMixin
+from ...test_modeling_flax_common import FlaxModelTesterMixin, ids_tensor
+
+
+if is_flax_available():
+    import jax
+    import jax.numpy as jnp
+
+    from transformers.models.gemma.modeling_flax_gemma import (
+        FlaxGemmaForCausalLM,
+        FlaxGemmaModel,
+    )
+
+
+class FlaxGemmaModelTester:
+    def __init__(
+        self,
+        parent,
+        batch_size=2,
+        seq_length=7,
+        is_training=True,
+        use_input_mask=True,
+        use_token_type_ids=False,
+        use_labels=True,
+        vocab_size=99,
+        hidden_size=32,
+        num_hidden_layers=2,
+        num_attention_heads=4,
+        num_key_value_heads=2,
+        intermediate_size=37,
+        hidden_act="gelu",
+        hidden_dropout_prob=0.1,
+        attention_probs_dropout_prob=0.1,
+        max_position_embeddings=512,
+        initializer_range=0.02,
+    ):
+        self.parent = parent
+        self.batch_size = batch_size
+        self.seq_length = seq_length
+        self.is_training = is_training
+        self.use_input_mask = use_input_mask
+        self.use_token_type_ids = use_token_type_ids
+        self.use_labels = use_labels
+        self.vocab_size = vocab_size
+        self.hidden_size = hidden_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.num_key_value_heads = num_key_value_heads
+        self.intermediate_size = intermediate_size
+        self.hidden_act = hidden_act
+        self.hidden_dropout_prob = hidden_dropout_prob
+        self.attention_probs_dropout_prob = attention_probs_dropout_prob
+        self.max_position_embeddings = max_position_embeddings
+        self.initializer_range = initializer_range
+        self.scope = None
+        self.bos_token_id = vocab_size - 1
+        self.eos_token_id = vocab_size - 1
+        self.pad_token_id = vocab_size - 1
+
+    def prepare_config_and_inputs(self):
+        input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size)
+
+        input_mask = None
+        if self.use_input_mask:
+            input_mask = np.tril(np.ones((self.batch_size, self.seq_length)))
+
+        config = GemmaConfig(
+            vocab_size=self.vocab_size,
+            hidden_size=self.hidden_size,
+            num_hidden_layers=self.num_hidden_layers,
+            num_attention_heads=self.num_attention_heads,
+            num_key_value_heads=self.num_key_value_heads,
+            head_dim=self.hidden_size // self.num_attention_heads,
+            intermediate_size=self.intermediate_size,
+            hidden_act=self.hidden_act,
+            hidden_dropout_prob=self.hidden_dropout_prob,
+            attention_probs_dropout_prob=self.attention_probs_dropout_prob,
+            max_position_embeddings=self.max_position_embeddings,
+            use_cache=True,
+            is_decoder=False,
+            initializer_range=self.initializer_range,
+        )
+
+        return config, input_ids, input_mask
+
+    def prepare_config_and_inputs_for_common(self):
+        config_and_inputs = self.prepare_config_and_inputs()
+        config, input_ids, attention_mask = config_and_inputs
+        inputs_dict = {"input_ids": input_ids, "attention_mask": attention_mask}
+        return config, inputs_dict
+
+    def check_use_cache_forward(self, model_class_name, config, input_ids, attention_mask):
+        max_decoder_length = 20
+        model = model_class_name(config)
+
+        past_key_values = model.init_cache(input_ids.shape[0], max_decoder_length)
+        attention_mask = jnp.ones((input_ids.shape[0], max_decoder_length), dtype="i4")
+
+        position_ids = jnp.broadcast_to(
+            jnp.arange(input_ids.shape[-1] - 1)[None, :], (input_ids.shape[0], input_ids.shape[-1] - 1)
+        )
+        outputs_cache = model(
+            input_ids[:, :-1],
+            attention_mask=attention_mask,
+            past_key_values=past_key_values,
+            position_ids=position_ids,
+        )
+
+        position_ids = jnp.array(input_ids.shape[0] * [[input_ids.shape[-1] - 1]], dtype="i4")
+        outputs_cache_next = model(
+            input_ids[:, -1:],
+            attention_mask=attention_mask,
+            past_key_values=outputs_cache.past_key_values,
+            position_ids=position_ids,
+        )
+
+        outputs = model(input_ids)
+
+        diff = np.max(np.abs((outputs_cache_next[0][:, -1, :5] - outputs[0][:, -1, :5])))
+        self.parent.assertTrue(diff < 1e-3, msg=f"Max diff is {diff}")
+
+    def check_use_cache_forward_with_attn_mask(self, model_class_name, config, input_ids, attention_mask):
+        max_decoder_length = 20
+        model = model_class_name(config)
+
+        attention_mask_cache = jnp.concatenate(
+            [attention_mask, jnp.zeros((attention_mask.shape[0], max_decoder_length - attention_mask.shape[1]))],
+            axis=-1,
+        )
+
+        past_key_values = model.init_cache(input_ids.shape[0], max_decoder_length)
+        position_ids = jnp.broadcast_to(
+            jnp.arange(input_ids.shape[-1] - 1)[None, :], (input_ids.shape[0], input_ids.shape[-1] - 1)
+        )
+
+        outputs_cache = model(
+            input_ids[:, :-1],
+            attention_mask=attention_mask_cache,
+            past_key_values=past_key_values,
+            position_ids=position_ids,
+        )
+        position_ids = jnp.array(input_ids.shape[0] * [[input_ids.shape[-1] - 1]], dtype="i4")
+        outputs_cache_next = model(
+            input_ids[:, -1:],
+            past_key_values=outputs_cache.past_key_values,
+            attention_mask=attention_mask_cache,
+            position_ids=position_ids,
+        )
+
+        outputs = model(input_ids, attention_mask=attention_mask)
+
+        diff = np.max(np.abs((outputs_cache_next[0][:, -1, :5] - outputs[0][:, -1, :5])))
+        self.parent.assertTrue(diff < 1e-3, msg=f"Max diff is {diff}")
+
+
+@require_flax
+class FlaxGemmaModelTest(FlaxModelTesterMixin, FlaxGenerationTesterMixin, unittest.TestCase):
+    all_model_classes = (FlaxGemmaModel, FlaxGemmaForCausalLM) if is_flax_available() else ()
+    all_generative_model_classes = (FlaxGemmaForCausalLM,) if is_flax_available() else ()
+
+    def setUp(self):
+        self.model_tester = FlaxGemmaModelTester(self)
+
+    def test_use_cache_forward(self):
+        for model_class_name in self.all_model_classes:
+            config, input_ids, attention_mask = self.model_tester.prepare_config_and_inputs()
+            self.model_tester.check_use_cache_forward(model_class_name, config, input_ids, attention_mask)
+
+    def test_use_cache_forward_with_attn_mask(self):
+        for model_class_name in self.all_model_classes:
+            config, input_ids, attention_mask = self.model_tester.prepare_config_and_inputs()
+            self.model_tester.check_use_cache_forward_with_attn_mask(
+                model_class_name, config, input_ids, attention_mask
+            )
+
+    @slow
+    def test_model_from_pretrained(self):
+        for model_class_name in self.all_model_classes:
+            model = model_class_name.from_pretrained("google/gemma-2b", from_pt=True)
+            outputs = model(np.ones((1, 1)))
+            self.assertIsNotNone(outputs)
+
+
+@slow
+@require_flax
+class FlaxGemmaIntegrationTest(unittest.TestCase):
+    input_text = ["The capital of France is", "To play the perfect cover drive"]
+    model_id = "google/gemma-2b"
+    revision = "flax"
+
+    def setUp(self):
+        self.model, self.params = FlaxGemmaForCausalLM.from_pretrained(
+            self.model_id, revision=self.revision, _do_init=False
+        )
+        self.tokenizer = AutoTokenizer.from_pretrained(self.model_id)
+        self.tokenizer.padding_side = "left"
+
+    def test_logits(self):
+        inputs = self.tokenizer(self.input_text, return_tensors="np", padding=True)
+        # fmt: off
+        EXPECTED_MEAN = [
+            [-16.427, -21.386, -35.491, -36.258, -31.401, -36.370, -37.598],
+            [-21.386, -32.150, -33.155, -34.344, -34.706, -34.678, -38.495],
+        ]
+        EXPECTED_SLICE = [-33.462, -16.481, -30.837, -32.195, -33.113]
+        # fmt: on
+
+        logits = self.model(**inputs, params=self.params).logits
+
+        diff_mean = jnp.abs(logits.mean(-1) - np.array(EXPECTED_MEAN)).max()
+        diff_slice = jnp.abs(logits[0, -1, 475:480] - np.array(EXPECTED_SLICE)).max()
+
+        self.assertAlmostEqual(diff_mean, 0, places=3)
+        self.assertAlmostEqual(diff_slice, 0, places=3)
+
+    def test_generation(self):
+        EXPECTED_TEXTS = [
+            "The capital of France is a city of contrasts. It is a city of history, of art, of culture, of fashion",
+            "To play the perfect cover drive, you need to have a good technique and a good mindset.\n\nThe cover drive is a shot",
+        ]
+        inputs = self.tokenizer(self.input_text, return_tensors="np", padding=True)
+
+        output = self.model.generate(**inputs, params=self.params, max_new_tokens=20, do_sample=False)
+        output_text = self.tokenizer.batch_decode(output.sequences, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    def test_jit_generation(self):
+        EXPECTED_TEXTS = [
+            "The capital of France is a city of contrasts. It is a city of history, culture, and art, but it is",
+            "To play the perfect cover drive, you need to have a good technique and a good mindset.\n\nThe cover drive is a shot",
+        ]
+        inputs = self.tokenizer(self.input_text, return_tensors="np", padding=True)
+
+        def generate(input_ids, attention_mask):
+            outputs = self.model.generate(
+                input_ids, attention_mask=attention_mask, params=self.params, max_new_tokens=20, do_sample=False
+            )
+            return outputs
+
+        jit_generate = jax.jit(generate)
+        output_sequences = jit_generate(**inputs).sequences
+        output_text = self.tokenizer.batch_decode(output_sequences, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)

tests/models/gemma/test_modeling_gemma.py

+# coding=utf-8
+# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+""" Testing suite for the PyTorch Gemma model. """
+
+import tempfile
+import unittest
+
+import pytest
+from parameterized import parameterized
+
+from transformers import AutoModelForCausalLM, AutoTokenizer, GemmaConfig, is_torch_available
+from transformers.testing_utils import (
+    require_bitsandbytes,
+    require_flash_attn,
+    require_torch,
+    require_torch_gpu,
+    slow,
+    torch_device,
+)
+
+from ...generation.test_utils import GenerationTesterMixin
+from ...test_configuration_common import ConfigTester
+from ...test_modeling_common import ModelTesterMixin, ids_tensor
+from ...test_pipeline_mixin import PipelineTesterMixin
+
+
+if is_torch_available():
+    import torch
+
+    from transformers import GemmaForCausalLM, GemmaForSequenceClassification, GemmaModel
+
+
+class GemmaModelTester:
+    def __init__(
+        self,
+        parent,
+        batch_size=13,
+        seq_length=7,
+        is_training=True,
+        use_input_mask=True,
+        use_token_type_ids=False,
+        use_labels=True,
+        vocab_size=99,
+        hidden_size=32,
+        num_hidden_layers=2,
+        num_attention_heads=4,
+        num_key_value_heads=2,
+        intermediate_size=37,
+        hidden_act="gelu",
+        hidden_dropout_prob=0.1,
+        attention_probs_dropout_prob=0.1,
+        max_position_embeddings=512,
+        type_vocab_size=16,
+        type_sequence_label_size=2,
+        initializer_range=0.02,
+        num_labels=3,
+        num_choices=4,
+        pad_token_id=0,
+        scope=None,
+    ):
+        self.parent = parent
+        self.batch_size = batch_size
+        self.seq_length = seq_length
+        self.is_training = is_training
+        self.use_input_mask = use_input_mask
+        self.use_token_type_ids = use_token_type_ids
+        self.use_labels = use_labels
+        self.vocab_size = vocab_size
+        self.hidden_size = hidden_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.num_key_value_heads = num_key_value_heads
+        self.intermediate_size = intermediate_size
+        self.hidden_act = hidden_act
+        self.hidden_dropout_prob = hidden_dropout_prob
+        self.attention_probs_dropout_prob = attention_probs_dropout_prob
+        self.max_position_embeddings = max_position_embeddings
+        self.type_vocab_size = type_vocab_size
+        self.type_sequence_label_size = type_sequence_label_size
+        self.initializer_range = initializer_range
+        self.num_labels = num_labels
+        self.num_choices = num_choices
+        self.pad_token_id = pad_token_id
+        self.scope = scope
+        self.head_dim = self.hidden_size // self.num_attention_heads
+
+    # Copied from tests.models.mistral.test_modeling_mistral.MistralModelTester.prepare_config_and_inputs
+    def prepare_config_and_inputs(self):
+        input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size)
+
+        input_mask = None
+        if self.use_input_mask:
+            input_mask = torch.tril(torch.ones(self.batch_size, self.seq_length)).to(torch_device)
+
+        token_type_ids = None
+        if self.use_token_type_ids:
+            token_type_ids = ids_tensor([self.batch_size, self.seq_length], self.type_vocab_size)
+
+        sequence_labels = None
+        token_labels = None
+        choice_labels = None
+        if self.use_labels:
+            sequence_labels = ids_tensor([self.batch_size], self.type_sequence_label_size)
+            token_labels = ids_tensor([self.batch_size, self.seq_length], self.num_labels)
+            choice_labels = ids_tensor([self.batch_size], self.num_choices)
+
+        config = self.get_config()
+
+        return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
+
+    # Ignore copy
+    def get_config(self):
+        return GemmaConfig(
+            vocab_size=self.vocab_size,
+            hidden_size=self.hidden_size,
+            num_hidden_layers=self.num_hidden_layers,
+            num_attention_heads=self.num_attention_heads,
+            num_key_value_heads=self.num_key_value_heads,
+            intermediate_size=self.intermediate_size,
+            hidden_act=self.hidden_act,
+            hidden_dropout_prob=self.hidden_dropout_prob,
+            attention_probs_dropout_prob=self.attention_probs_dropout_prob,
+            max_position_embeddings=self.max_position_embeddings,
+            type_vocab_size=self.type_vocab_size,
+            is_decoder=False,
+            initializer_range=self.initializer_range,
+            pad_token_id=self.pad_token_id,
+            head_dim=self.head_dim,
+        )
+
+    # Copied from tests.models.llama.test_modeling_llama.LlamaModelTester.create_and_check_model with Llama->Gemma
+    def create_and_check_model(
+        self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
+    ):
+        model = GemmaModel(config=config)
+        model.to(torch_device)
+        model.eval()
+        result = model(input_ids, attention_mask=input_mask)
+        result = model(input_ids)
+        self.parent.assertEqual(result.last_hidden_state.shape, (self.batch_size, self.seq_length, self.hidden_size))
+
+    # Copied from tests.models.llama.test_modeling_llama.LlamaModelTester.create_and_check_model_as_decoder with Llama->Gemma
+    def create_and_check_model_as_decoder(
+        self,
+        config,
+        input_ids,
+        token_type_ids,
+        input_mask,
+        sequence_labels,
+        token_labels,
+        choice_labels,
+        encoder_hidden_states,
+        encoder_attention_mask,
+    ):
+        config.add_cross_attention = True
+        model = GemmaModel(config)
+        model.to(torch_device)
+        model.eval()
+        result = model(
+            input_ids,
+            attention_mask=input_mask,
+            encoder_hidden_states=encoder_hidden_states,
+            encoder_attention_mask=encoder_attention_mask,
+        )
+        result = model(
+            input_ids,
+            attention_mask=input_mask,
+            encoder_hidden_states=encoder_hidden_states,
+        )
+        result = model(input_ids, attention_mask=input_mask)
+        self.parent.assertEqual(result.last_hidden_state.shape, (self.batch_size, self.seq_length, self.hidden_size))
+
+    # Copied from tests.models.llama.test_modeling_llama.LlamaModelTester.create_and_check_for_causal_lm with Llama->Gemma
+    def create_and_check_for_causal_lm(
+        self,
+        config,
+        input_ids,
+        token_type_ids,
+        input_mask,
+        sequence_labels,
+        token_labels,
+        choice_labels,
+        encoder_hidden_states,
+        encoder_attention_mask,
+    ):
+        model = GemmaForCausalLM(config=config)
+        model.to(torch_device)
+        model.eval()
+        result = model(input_ids, attention_mask=input_mask, labels=token_labels)
+        self.parent.assertEqual(result.logits.shape, (self.batch_size, self.seq_length, self.vocab_size))
+
+    # Copied from tests.models.llama.test_modeling_llama.LlamaModelTester.create_and_check_decoder_model_past_large_inputs with Llama->Gemma
+    def create_and_check_decoder_model_past_large_inputs(
+        self,
+        config,
+        input_ids,
+        token_type_ids,
+        input_mask,
+        sequence_labels,
+        token_labels,
+        choice_labels,
+        encoder_hidden_states,
+        encoder_attention_mask,
+    ):
+        config.is_decoder = True
+        config.add_cross_attention = True
+        model = GemmaForCausalLM(config=config)
+        model.to(torch_device)
+        model.eval()
+
+        # first forward pass
+        outputs = model(
+            input_ids,
+            attention_mask=input_mask,
+            encoder_hidden_states=encoder_hidden_states,
+            encoder_attention_mask=encoder_attention_mask,
+            use_cache=True,
+        )
+        past_key_values = outputs.past_key_values
+
+        # create hypothetical multiple next token and extent to next_input_ids
+        next_tokens = ids_tensor((self.batch_size, 3), config.vocab_size)
+        next_mask = ids_tensor((self.batch_size, 3), vocab_size=2)
+
+        # append to next input_ids and
+        next_input_ids = torch.cat([input_ids, next_tokens], dim=-1)
+        next_attention_mask = torch.cat([input_mask, next_mask], dim=-1)
+
+        output_from_no_past = model(
+            next_input_ids,
+            attention_mask=next_attention_mask,
+            encoder_hidden_states=encoder_hidden_states,
+            encoder_attention_mask=encoder_attention_mask,
+            output_hidden_states=True,
+        )["hidden_states"][0]
+        output_from_past = model(
+            next_tokens,
+            attention_mask=next_attention_mask,
+            encoder_hidden_states=encoder_hidden_states,
+            encoder_attention_mask=encoder_attention_mask,
+            past_key_values=past_key_values,
+            output_hidden_states=True,
+        )["hidden_states"][0]
+
+        # select random slice
+        random_slice_idx = ids_tensor((1,), output_from_past.shape[-1]).item()
+        output_from_no_past_slice = output_from_no_past[:, -3:, random_slice_idx].detach()
+        output_from_past_slice = output_from_past[:, :, random_slice_idx].detach()
+
+        self.parent.assertTrue(output_from_past_slice.shape[1] == next_tokens.shape[1])
+
+        # test that outputs are equal for slice
+        self.parent.assertTrue(torch.allclose(output_from_past_slice, output_from_no_past_slice, atol=1e-3))
+
+    # Copied from tests.models.llama.test_modeling_llama.LlamaModelTester.prepare_config_and_inputs_for_common with Llama->Gemma
+    def prepare_config_and_inputs_for_common(self):
+        config_and_inputs = self.prepare_config_and_inputs()
+        (
+            config,
+            input_ids,
+            token_type_ids,
+            input_mask,
+            sequence_labels,
+            token_labels,
+            choice_labels,
+        ) = config_and_inputs
+        inputs_dict = {"input_ids": input_ids, "attention_mask": input_mask}
+        return config, inputs_dict
+
+
+@require_torch
+class GemmaModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin, unittest.TestCase):
+    all_model_classes = (GemmaModel, GemmaForCausalLM, GemmaForSequenceClassification) if is_torch_available() else ()
+    all_generative_model_classes = (GemmaForCausalLM,) if is_torch_available() else ()
+    pipeline_model_mapping = (
+        {
+            "feature-extraction": GemmaModel,
+            "text-classification": GemmaForSequenceClassification,
+            "text-generation": GemmaForCausalLM,
+            "zero-shot": GemmaForSequenceClassification,
+        }
+        if is_torch_available()
+        else {}
+    )
+    test_headmasking = False
+    test_pruning = False
+
+    # TODO (ydshieh): Check this. See https://app.circleci.com/pipelines/github/huggingface/transformers/79245/workflows/9490ef58-79c2-410d-8f51-e3495156cf9c/jobs/1012146
+    def is_pipeline_test_to_skip(
+        self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
+    ):
+        return True
+
+    def setUp(self):
+        self.model_tester = GemmaModelTester(self)
+        self.config_tester = ConfigTester(self, config_class=GemmaConfig, hidden_size=37)
+
+    def test_config(self):
+        self.config_tester.run_common_tests()
+
+    def test_model(self):
+        config_and_inputs = self.model_tester.prepare_config_and_inputs()
+        self.model_tester.create_and_check_model(*config_and_inputs)
+
+    def test_model_various_embeddings(self):
+        config_and_inputs = self.model_tester.prepare_config_and_inputs()
+        for type in ["absolute", "relative_key", "relative_key_query"]:
+            config_and_inputs[0].position_embedding_type = type
+            self.model_tester.create_and_check_model(*config_and_inputs)
+
+    def test_Gemma_sequence_classification_model(self):
+        config, input_dict = self.model_tester.prepare_config_and_inputs_for_common()
+        print(config)
+        config.num_labels = 3
+        input_ids = input_dict["input_ids"]
+        attention_mask = input_ids.ne(1).to(torch_device)
+        sequence_labels = ids_tensor([self.model_tester.batch_size], self.model_tester.type_sequence_label_size)
+        model = GemmaForSequenceClassification(config)
+        model.to(torch_device)
+        model.eval()
+        result = model(input_ids, attention_mask=attention_mask, labels=sequence_labels)
+        self.assertEqual(result.logits.shape, (self.model_tester.batch_size, self.model_tester.num_labels))
+
+    def test_Gemma_sequence_classification_model_for_single_label(self):
+        config, input_dict = self.model_tester.prepare_config_and_inputs_for_common()
+        config.num_labels = 3
+        config.problem_type = "single_label_classification"
+        input_ids = input_dict["input_ids"]
+        attention_mask = input_ids.ne(1).to(torch_device)
+        sequence_labels = ids_tensor([self.model_tester.batch_size], self.model_tester.type_sequence_label_size)
+        model = GemmaForSequenceClassification(config)
+        model.to(torch_device)
+        model.eval()
+        result = model(input_ids, attention_mask=attention_mask, labels=sequence_labels)
+        self.assertEqual(result.logits.shape, (self.model_tester.batch_size, self.model_tester.num_labels))
+
+    def test_Gemma_sequence_classification_model_for_multi_label(self):
+        config, input_dict = self.model_tester.prepare_config_and_inputs_for_common()
+        config.num_labels = 3
+        config.problem_type = "multi_label_classification"
+        input_ids = input_dict["input_ids"]
+        attention_mask = input_ids.ne(1).to(torch_device)
+        sequence_labels = ids_tensor(
+            [self.model_tester.batch_size, config.num_labels], self.model_tester.type_sequence_label_size
+        ).to(torch.float)
+        model = GemmaForSequenceClassification(config)
+        model.to(torch_device)
+        model.eval()
+        result = model(input_ids, attention_mask=attention_mask, labels=sequence_labels)
+        self.assertEqual(result.logits.shape, (self.model_tester.batch_size, self.model_tester.num_labels))
+
+    @unittest.skip("TODO @gante fix this for Llama")
+    @parameterized.expand([(1, False), (1, True), (4, False)])
+    def test_new_cache_format(self, num_beams, do_sample):
+        pass
+
+    @unittest.skip("Gemma buffers include complex numbers, which breaks this test")
+    def test_save_load_fast_init_from_base(self):
+        pass
+
+    @unittest.skip("Gemma uses GQA on all models so the KV cache is a non standard format")
+    def test_past_key_values_format(self):
+        pass
+
+    @require_flash_attn
+    @require_torch_gpu
+    @pytest.mark.flash_attn_test
+    @slow
+    def test_flash_attn_2_generate_padding_right(self):
+        import torch
+
+        for model_class in self.all_generative_model_classes:
+            config, _ = self.model_tester.prepare_config_and_inputs_for_common()
+            model = model_class(config)
+
+            with tempfile.TemporaryDirectory() as tmpdirname:
+                model.save_pretrained(tmpdirname)
+                model = model_class.from_pretrained(tmpdirname, torch_dtype=torch.float16, low_cpu_mem_usage=True).to(
+                    torch_device
+                )
+
+                dummy_input = torch.LongTensor([[0, 2, 3, 4], [0, 2, 3, 4]]).to(torch_device)
+                dummy_attention_mask = torch.LongTensor([[1, 1, 1, 1], [1, 1, 1, 0]]).to(torch_device)
+
+                model.generate(dummy_input, attention_mask=dummy_attention_mask, max_new_tokens=1, do_sample=False)
+
+                model = model_class.from_pretrained(
+                    tmpdirname,
+                    torch_dtype=torch.float16,
+                    attn_implementation="flash_attention_2",
+                    low_cpu_mem_usage=True,
+                ).to(torch_device)
+
+                with self.assertRaises(ValueError):
+                    _ = model.generate(
+                        dummy_input, attention_mask=dummy_attention_mask, max_new_tokens=1, do_sample=False
+                    )
+
+    @require_flash_attn
+    @require_torch_gpu
+    @pytest.mark.flash_attn_test
+    @slow
+    def test_flash_attn_2_generate_use_cache(self):
+        import torch
+
+        max_new_tokens = 30
+
+        for model_class in self.all_generative_model_classes:
+            config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
+
+            dummy_input = inputs_dict[model_class.main_input_name]
+            if dummy_input.dtype in [torch.float32, torch.bfloat16]:
+                dummy_input = dummy_input.to(torch.float16)
+
+            # make sure that all models have enough positions for generation
+            if hasattr(config, "max_position_embeddings"):
+                config.max_position_embeddings = max_new_tokens + dummy_input.shape[1] + 1
+
+            model = model_class(config)
+
+            with tempfile.TemporaryDirectory() as tmpdirname:
+                model.save_pretrained(tmpdirname)
+
+                dummy_attention_mask = inputs_dict.get("attention_mask", torch.ones_like(dummy_input))
+                # NOTE: Gemma apparently does not support right padding + use_cache with FA2.
+                dummy_attention_mask[:, -1] = 1
+
+                model = model_class.from_pretrained(
+                    tmpdirname,
+                    torch_dtype=torch.float16,
+                    attn_implementation="flash_attention_2",
+                    low_cpu_mem_usage=True,
+                ).to(torch_device)
+
+                # Just test that a large cache works as expected
+                _ = model.generate(
+                    dummy_input,
+                    attention_mask=dummy_attention_mask,
+                    max_new_tokens=max_new_tokens,
+                    do_sample=False,
+                    use_cache=True,
+                )
+
+    @require_flash_attn
+    @require_torch_gpu
+    @pytest.mark.flash_attn_test
+    @slow
+    def test_flash_attn_2_inference_padding_right(self):
+        self.skipTest("Gemma flash attention does not support right padding")
+
+
+@require_torch_gpu
+@slow
+class GemmaIntegrationTest(unittest.TestCase):
+    input_text = ["Hello I am doing", "Hi today"]
+
+    def test_model_2b_fp32(self):
+        model_id = "google/gemma-2b"
+        EXPECTED_TEXTS = [
+            "Hello I am doing a project on the 1990s and I need to know what the most popular music",
+            "Hi today I am going to share with you a very easy and simple recipe of <strong><em>Kaju Kat",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True).to(torch_device)
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    def test_model_2b_fp16(self):
+        model_id = "google/gemma-2b"
+        EXPECTED_TEXTS = [
+            "Hello I am doing a project on the 1990s and I need to know what the most popular music",
+            "Hi today I am going to share with you a very easy and simple recipe of <strong><em>Kaju Kat",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True, torch_dtype=torch.float16).to(
+            torch_device
+        )
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    def test_model_2b_fp16_static_cache(self):
+        model_id = "google/gemma-2b"
+        EXPECTED_TEXTS = [
+            "Hello I am doing a project on the 1990s and I need to know what the most popular music",
+            "Hi today I am going to share with you a very easy and simple recipe of <strong><em>Kaju Kat",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True, torch_dtype=torch.float16).to(
+            torch_device
+        )
+
+        model.generation_config.cache_implementation = "static"
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    def test_model_2b_bf16(self):
+        model_id = "google/gemma-2b"
+        EXPECTED_TEXTS = [
+            "Hello I am doing a project on the 1990s and I need to know what the most popular music",
+            "Hi today I am going to share with you a very easy and simple recipe of <strong><em>Khichdi",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True, torch_dtype=torch.bfloat16).to(
+            torch_device
+        )
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    @require_bitsandbytes
+    def test_model_2b_4bit(self):
+        model_id = "google/gemma-2b"
+        EXPECTED_TEXTS = [
+            "Hello I am doing a project and I need to make a 3d model of a house. I have been using",
+            "Hi today I'd like to share with you my experience with the new wattpad wattpad wattpad wattpad wattpad wattpad wattpad",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True, load_in_4bit=True)
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    @unittest.skip("The test will not fit our CI runners")
+    def test_model_7b_fp32(self):
+        model_id = "google/gemma-7b"
+        EXPECTED_TEXTS = [
+            "Hello my name is ***** ***** I will be assisting you today. I am sorry to hear about your issue. I will",
+            "Hi,\n\nI have a problem with my 2005 1.6 16",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True).to(torch_device)
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    def test_model_7b_fp16(self):
+        model_id = "google/gemma-7b"
+        EXPECTED_TEXTS = [
+            """Hello I am doing a project on a 1999 4.0L 4x4. I""",
+            "Hi today I am going to show you how to make a simple and easy to make a DIY 3D",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True, torch_dtype=torch.float16).to(
+            torch_device
+        )
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    def test_model_7b_bf16(self):
+        model_id = "google/gemma-7b"
+        EXPECTED_TEXTS = [
+            """Hello I am doing a project on a 1991 240sx and I am trying to find""",
+            "Hi today I am going to show you how to make a very simple and easy to make a very simple and",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True, torch_dtype=torch.bfloat16).to(
+            torch_device
+        )
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    def test_model_7b_fp16_static_cache(self):
+        model_id = "google/gemma-7b"
+        EXPECTED_TEXTS = [
+            """Hello I am doing a project on a 1999 4.0L 4x4. I""",
+            "Hi today I am going to show you how to make a simple and easy to make a DIY 3D",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True, torch_dtype=torch.float16).to(
+            torch_device
+        )
+
+        model.generation_config.cache_implementation = "static"
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)
+
+    @require_bitsandbytes
+    def test_model_7b_4bit(self):
+        model_id = "google/gemma-7b"
+        EXPECTED_TEXTS = [
+            "Hello I am doing a project for my school and I am trying to make a program that will take a number and then",
+            """Hi today I am going to talk about the new update for the game called "The new update" and I""",
+        ]
+
+        model = AutoModelForCausalLM.from_pretrained(model_id, low_cpu_mem_usage=True, load_in_4bit=True)
+
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)
+
+        output = model.generate(**inputs, max_new_tokens=20, do_sample=False)
+        output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
+
+        self.assertEqual(output_text, EXPECTED_TEXTS)

tests/models/gemma/test_tokenization_gemma.py

+# coding=utf-8
+# Copyright 2024 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import os
+import tempfile
+import unittest
+
+from datasets import load_dataset
+
+from transformers import (
+    AddedToken,
+    GemmaTokenizer,
+    GemmaTokenizerFast,
+    is_torch_available,
+)
+from transformers.convert_slow_tokenizer import convert_slow_tokenizer
+from transformers.testing_utils import (
+    get_tests_dir,
+    nested_simplify,
+    require_jinja,
+    require_sentencepiece,
+    require_tokenizers,
+    require_torch,
+    slow,
+)
+
+from ...test_tokenization_common import TokenizerTesterMixin
+
+
+SAMPLE_VOCAB = get_tests_dir("fixtures/test_sentencepiece.model")
+
+
+if is_torch_available():
+    pass
+
+
+@require_sentencepiece
+@require_tokenizers
+class GemmaTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
+    tokenizer_class = GemmaTokenizer
+    rust_tokenizer_class = GemmaTokenizerFast
+
+    test_rust_tokenizer = False
+    test_sentencepiece = True
+    from_pretrained_kwargs = {}
+
+    def setUp(self):
+        super().setUp()
+        # We have a SentencePiece fixture for testing
+        tokenizer = GemmaTokenizer(SAMPLE_VOCAB, keep_accents=True)
+        tokenizer.pad_token = tokenizer.eos_token
+        tokenizer.save_pretrained(self.tmpdirname)
+
+    @require_torch
+    def test_batch_tokenization(self):
+        if not self.test_seq2seq:
+            return
+
+        tokenizers = self.get_tokenizers()
+        for tokenizer in tokenizers:
+            with self.subTest(f"{tokenizer.__class__.__name__}"):
+                # Longer text that will definitely require truncation.
+                text = [
+                    " UN Chief Says There Is No Military Solution in Syria",
+                    " Secretary-General Ban Ki-moon says his response to Russia's stepped up military support for"
+                    " Syria is that 'there is no military solution' to the nearly five-year conflict and more weapons"
+                    " will only worsen the violence and misery for millions of people.",
+                ]
+                try:
+                    batch = tokenizer(
+                        text=text,
+                        max_length=3,
+                        max_target_length=10,
+                        return_tensors="pt",
+                    )
+                except NotImplementedError:
+                    return
+                self.assertEqual(batch.input_ids.shape[1], 3)
+                # max_target_length will default to max_length if not specified
+                batch = tokenizer(text, max_length=3, return_tensors="pt")
+                self.assertEqual(batch.input_ids.shape[1], 3)
+
+                batch_encoder_only = tokenizer(text=text, max_length=3, max_target_length=10, return_tensors="pt")
+                self.assertEqual(batch_encoder_only.input_ids.shape[1], 3)
+                self.assertEqual(batch_encoder_only.attention_mask.shape[1], 3)
+                self.assertNotIn("decoder_input_ids", batch_encoder_only)
+
+    @unittest.skip("Unfortunately way too slow to build a BPE with SentencePiece.")
+    def test_save_slow_from_fast_and_reload_fast(self):
+        pass
+
+    def test_special_tokens_initialization(self):
+        for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
+            with self.subTest(f"{tokenizer.__class__.__name__} ({pretrained_name})"):
+                added_tokens = [AddedToken("<special>", lstrip=True)]
+
+                tokenizer_r = self.rust_tokenizer_class.from_pretrained(
+                    pretrained_name, additional_special_tokens=added_tokens, **kwargs
+                )
+                r_output = tokenizer_r.encode("Hey this is a <special> token")
+
+                special_token_id = tokenizer_r.encode("<special>", add_special_tokens=False)[0]
+
+                self.assertTrue(special_token_id in r_output)
+
+                if self.test_slow_tokenizer:
+                    tokenizer_cr = self.rust_tokenizer_class.from_pretrained(
+                        pretrained_name,
+                        additional_special_tokens=added_tokens,
+                        **kwargs,  # , from_slow=True <- unfortunately too slow to convert
+                    )
+                    tokenizer_p = self.tokenizer_class.from_pretrained(
+                        pretrained_name, additional_special_tokens=added_tokens, **kwargs
+                    )
+
+                    p_output = tokenizer_p.encode("Hey this is a <special> token")
+
+                    cr_output = tokenizer_cr.encode("Hey this is a <special> token")
+
+                    self.assertEqual(p_output, r_output)
+                    self.assertEqual(cr_output, r_output)
+                    self.assertTrue(special_token_id in p_output)
+                    self.assertTrue(special_token_id in cr_output)
+
+    @slow
+    def test_tokenizer_integration(self):
+        expected_encoding =  {'input_ids': [[2, 158434, 591, 84193, 3836, 685, 6599, 31223, 235290, 140247, 578, 6599, 31223, 235290, 145139, 235290, 3491, 235275, 6572, 3311, 235290, 38197, 109959, 591, 25894, 235269, 162174, 235290, 235284, 235269, 1791, 6362, 12481, 235269, 1576, 18622, 235269, 2900, 1136, 86684, 235269, 29092, 4632, 16994, 604, 13146, 14944, 40371, 591, 19700, 235327, 235275, 578, 13146, 14944, 25511, 591, 235300, 12474, 235275, 675, 1163, 235248, 235304, 235284, 235340, 229903, 5377, 575, 235248, 235274, 235276, 235276, 235340, 17044, 578, 5271, 1061, 118345, 1865, 125247, 235269, 8745, 111226, 578, 176888, 235265], [2, 25894, 603, 6869, 577, 953, 235290, 8297, 5271, 209099, 41642, 774, 748, 78253, 2793, 731, 51506, 34346, 611, 2145, 2731, 578, 1833, 4807, 575, 832, 16630, 235265], [2, 651, 4320, 8426, 25341, 36271, 1163, 573, 27894, 5929, 235265]], 'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]]}  # fmt: skip
+        self.tokenizer_integration_test_util(
+            expected_encoding=expected_encoding,
+            model_name="hf-internal-testing/dummy-gemma",
+            revision="",
+            padding=False,
+        )
+
+    @unittest.skip("worker 'gw4' crashed on CI, passing locally.")
+    def test_pickle_subword_regularization_tokenizer(self):
+        pass
+
+    @unittest.skip("worker 'gw4' crashed on CI, passing locally.")
+    def test_subword_regularization_tokenizer(self):
+        pass
+
+    @unittest.skip("This test will be removed from main @LysandreJik")
+    def test_pretrained_model_lists(self):
+        pass
+
+    @unittest.skip("Skipping")
+    def test_torch_encode_plus_sent_to_model(self):
+        pass
+
+
+@require_torch
+@require_sentencepiece
+@require_tokenizers
+class GemmaIntegrationTest(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        checkpoint_name = "hf-internal-testing/dummy-gemma"
+        cls.tokenizer: GemmaTokenizer = GemmaTokenizer.from_pretrained(
+            checkpoint_name, eos_token="<s>"
+        )  # add this token
+        cls.rust_tokenizer = GemmaTokenizerFast.from_pretrained(
+            checkpoint_name, eos_token="<s>", from_slow=True
+        )  # add this token
+        return cls
+
+    @require_torch
+    def integration_tests(self):
+        inputs = self.tokenizer(
+            ["The following string should be properly encoded: Hello.", "But ird and ปี   ird   ด"],
+            return_tensors="pt",
+        )
+
+        self.assertEqual(
+            nested_simplify(inputs),
+            {
+                "input_ids": [
+                    [2, 450, 1494, 1347, 881, 367, 6284, 18511, 29901, 15043, 29889],
+                    [2, 1205, 29871, 1823, 322, 29871, 31010, 30691, 1678, 1823, 1678, 30718],
+                ],
+                "attention_mask": [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]],
+            },
+        )
+
+    def test_fast_special_tokens(self):
+        slow_tokenizer = self.tokenizer
+        fast_tokenizer = self.rust_tokenizer
+        slow = slow_tokenizer.encode("A sample test", add_special_tokens=True)
+        assert slow == [2, 235280, 6453, 2121]
+
+        fast_tokenizer.add_eos_token = False
+        fast = fast_tokenizer.encode("A sample test", add_special_tokens=True)
+        assert fast == [2, 235280, 6453, 2121]
+
+        fast_tokenizer.add_eos_token = True
+        fast = fast_tokenizer.encode("A sample test", add_special_tokens=True)
+        assert fast == [2, 235280, 6453, 2121, 204]
+
+        slow_tokenizer.add_eos_token = True
+        slow = slow_tokenizer.encode("A sample test", add_special_tokens=True)
+        assert slow == [2, 235280, 6453, 2121, 204]
+
+        self.tokenizer.add_eos_token = False
+        self.rust_tokenizer.add_eos_token = False
+
+    @unittest.skip("Not super important and always failing. Let's skip it")
+    @slow
+    def test_conversion(self):
+        # This is excruciatingly slow since it has to recreate the entire merge
+        # list from the original vocabulary in spm
+        self.rust_tokenizer.save_pretrained("./out")
+        with tempfile.TemporaryDirectory() as dirname:
+            self.rust_tokenizer.save_pretrained(dirname)
+
+            with open(os.path.join(dirname, "tokenizer.json"), "r") as f:
+                old_serialized = f.read()
+
+        new_tokenizer = convert_slow_tokenizer(self.tokenizer)
+        with tempfile.NamedTemporaryFile() as f:
+            new_tokenizer.save(f.name)
+            # Re-opening since `f` is in bytes.
+            new_serialized = open(f.name, "r").read()
+            with open("out_tokenizer.json", "w") as g:
+                g.write(new_serialized)
+
+            self.assertEqual(old_serialized, new_serialized)
+
+    def test_simple_encode_decode(self):
+        pyth_tokenizer = self.tokenizer
+        rust_tokenizer = self.rust_tokenizer
+
+        self.tokenizer.add_eos_token = False
+        self.rust_tokenizer.add_eos_token = False
+
+        self.assertEqual(pyth_tokenizer.encode("This is a test"), [2, 1596, 603, 476, 2121])
+        self.assertEqual(rust_tokenizer.encode("This is a test"), [2, 1596, 603, 476, 2121])
+        self.assertEqual(pyth_tokenizer.decode([2, 1596, 603, 476, 2121], skip_special_tokens=True), "This is a test")
+        self.assertEqual(rust_tokenizer.decode([2, 1596, 603, 476, 2121], skip_special_tokens=True), "This is a test")
+
+        # bytefallback showcase
+        self.assertEqual(pyth_tokenizer.encode("生活的真谛是"), [2, 122182, 235710, 245467, 235427] )  # fmt: skip
+        self.assertEqual(rust_tokenizer.encode("生活的真谛是"), [2, 122182, 235710, 245467, 235427] )  # fmt: skip
+        self.assertEqual(
+            pyth_tokenizer.decode([2, 122182, 235710, 245467, 235427], skip_special_tokens=True),
+            "生活的真谛是",
+        )
+        self.assertEqual(
+            rust_tokenizer.decode([2, 122182, 235710, 245467, 235427], skip_special_tokens=True),
+            "生活的真谛是",
+        )
+
+        # Inner spaces showcase
+        self.assertEqual(pyth_tokenizer.encode("Hi  Hello"), [2, 2151, 139, 4521])
+        self.assertEqual(rust_tokenizer.encode("Hi  Hello"), [2, 2151, 139, 4521])
+        self.assertEqual(pyth_tokenizer.decode([2, 2151, 139, 4521], skip_special_tokens=True), "Hi  Hello")
+        self.assertEqual(rust_tokenizer.decode([2, 2151, 139, 4521], skip_special_tokens=True), "Hi  Hello")
+
+        self.assertEqual(pyth_tokenizer.encode("Hi   Hello"), [2, 2151, 140, 4521])
+        self.assertEqual(rust_tokenizer.encode("Hi   Hello"), [2, 2151, 140, 4521])
+        self.assertEqual(pyth_tokenizer.decode([2, 2151, 140, 4521], skip_special_tokens=True), "Hi   Hello")
+        self.assertEqual(rust_tokenizer.decode([2, 2151, 140, 4521], skip_special_tokens=True), "Hi   Hello")
+
+        self.assertEqual(pyth_tokenizer.encode(""), [2])
+        self.assertEqual(rust_tokenizer.encode(""), [2])
+
+        self.assertEqual(pyth_tokenizer.encode(" "), [2, 235248])
+        self.assertEqual(rust_tokenizer.encode(" "), [2, 235248])
+
+        self.assertEqual(pyth_tokenizer.encode("  "), [2, 139])
+        self.assertEqual(rust_tokenizer.encode("  "), [2, 139])
+
+        self.assertEqual(pyth_tokenizer.encode(" Hello"), [2, 25957])
+        self.assertEqual(rust_tokenizer.encode(" Hello"), [2, 25957])
+
+    def test_no_differences_decode(self):
+        self.tokenizer.add_eos_token = False
+        self.rust_tokenizer.add_eos_token = False
+        pyth_tokenizer = self.tokenizer
+        rust_tokenizer = self.rust_tokenizer
+
+        self.assertEqual(pyth_tokenizer.decode([869]), "og")
+        self.assertEqual(rust_tokenizer.decode([869]), "og")
+
+        self.assertEqual(pyth_tokenizer.decode([30112, 869]), " expenditureog")
+        self.assertEqual(rust_tokenizer.decode([30112, 869]), " expenditureog")
+
+    def test_no_differences_special_tokens(self):
+        pyth_tokenizer = self.tokenizer
+        rust_tokenizer = self.rust_tokenizer
+        self.assertEqual(pyth_tokenizer.encode(""), [2])
+        self.assertEqual(rust_tokenizer.encode(""), [2])
+
+        self.assertEqual(pyth_tokenizer.encode("<s>"), [2, 204])
+        self.assertEqual(rust_tokenizer.encode("<s>"), [2, 204])
+
+    @unittest.skipIf(
+        os.getenv("RUN_TOKENIZER_INTEGRATION", "0") == "0",
+        "RUN_TOKENIZER_INTEGRATION=1 to run tokenizer integration tests",
+    )
+    def test_integration_test_xnli(self):
+        import tqdm
+
+        pyth_tokenizer = self.tokenizer
+        rust_tokenizer = self.rust_tokenizer
+
+        dataset = load_dataset("code_x_glue_ct_code_to_text", "go")
+        for item in tqdm.tqdm(dataset["validation"]):
+            string = item["code"]
+            encoded1 = pyth_tokenizer.encode(string)
+            encoded2 = rust_tokenizer.encode(string)
+
+            self.assertEqual(encoded1, encoded2)
+
+            decoded1 = pyth_tokenizer.decode(encoded1, skip_special_tokens=True)
+            decoded2 = rust_tokenizer.decode(encoded1, skip_special_tokens=True)
+
+            self.assertEqual(decoded1, decoded2)
+
+        dataset = load_dataset("xnli", "all_languages")
+
+        for item in tqdm.tqdm(dataset["train"]):
+            for string in item["premise"].values():
+                encoded1 = pyth_tokenizer.encode(string)
+                encoded2 = rust_tokenizer.encode(string)
+
+                self.assertEqual(encoded1, encoded2)
+
+                decoded1 = pyth_tokenizer.decode(encoded1, skip_special_tokens=True)
+                decoded2 = rust_tokenizer.decode(encoded2, skip_special_tokens=True)
+
+                self.assertEqual(decoded1, decoded2)
+
+    def test_special_token_special_word(self):
+        # the word inform should be split as ['in', 'form']
+        tokenizer = GemmaTokenizer.from_pretrained("hf-internal-testing/dummy-gemma")
+        tokenizer.add_tokens([AddedToken("<REPR_END>", rstrip=True, lstrip=True)], special_tokens=False)
+        out1 = tokenizer.decode(
+            tokenizer.encode("<REPR_END>inform", add_special_tokens=False), spaces_between_special_tokens=False
+        )
+        self.assertEqual(out1, "<REPR_END>inform")
+        out2 = tokenizer.decode(
+            tokenizer.encode("<REPR_END>inform", add_special_tokens=False), spaces_between_special_tokens=True
+        )
+        # decoding strips the added prefix space.
+        self.assertEqual(out2, "<REPR_END> inform")
+        input_ids = tokenizer.encode("<REPR_END>inform", add_special_tokens=False)
+        self.assertEqual(input_ids, [256000, 43910])
+
+        out2 = tokenizer.decode(
+            tokenizer.encode(" <REPR_END> inform", add_special_tokens=False), spaces_between_special_tokens=False
+        )
+        # TODO @ArthurZ currently we strip left and right, so this will not keep the spaces
+        self.assertEqual(out2, "<REPR_END>inform")
+
+        ### Let's make sure decoding does not add extra spaces here and there
+        # TODO @ArthurZ this should be affected by the lstrip/rstrip/single word /normalize refactoring
+        # Since currently we always strip left and right of the token, results are as such
+        input_ids = tokenizer.encode("<s> Hello<s>how", add_special_tokens=False)
+        self.assertEqual(input_ids, [204, 25957, 204, 1139])
+        tokens = tokenizer.tokenize("<s> Hello<s>how", add_special_tokens=False)
+        self.assertEqual(tokens, ["<s>", "▁Hello", "<s>", "how"])
+        decoded_tokens = tokenizer.decode(input_ids)
+        self.assertEqual(decoded_tokens, "<s> Hello<s>how")
+
+        # Let's make sure that if there are any spaces, we don't remove them!
+        input_ids = tokenizer.encode(" <s> Hello<s> how", add_special_tokens=False)
+        self.assertEqual(input_ids, [235248, 204, 25957, 204, 1368])
+        tokens = tokenizer.tokenize(" <s> Hello<s> how", add_special_tokens=False)
+        self.assertEqual(tokens, ["▁", "<s>", "▁Hello", "<s>", "▁how"])
+        decoded_tokens = tokenizer.decode(input_ids)
+        self.assertEqual(decoded_tokens, " <s> Hello<s> how")
+
+    def test_some_edge_cases(self):
+        tokenizer = GemmaTokenizer.from_pretrained("hf-internal-testing/dummy-gemma")
+
+        sp_tokens = tokenizer.sp_model.encode("<s>>", out_type=str)
+        self.assertEqual(sp_tokens, ["<s>", ">"])
+        tokens = tokenizer.tokenize("<s>>")
+        self.assertEqual(sp_tokens, tokens)
+        self.assertEqual(tokens, ["<s>", ">"])
+
+        tokens = tokenizer.tokenize("")
+        self.assertEqual(tokens, [])
+        self.assertEqual(tokens, tokenizer.sp_model.encode("", out_type=str))
+
+        tokens = tokenizer.tokenize(" ")
+        self.assertEqual(tokens, ["▁"])
+        # a dummy prefix space is not added by the sp_model as it was de-activated
+        self.assertEqual(tokens, tokenizer.sp_model.encode(" ", out_type=str))
+
+        tokens = tokenizer.tokenize("▁")
+        self.assertEqual(tokens, ["▁"])
+        # a dummy prefix space is not added by the sp_model as it was de-activated
+        self.assertEqual(tokens, tokenizer.sp_model.encode("▁", out_type=str))
+
+        tokens = tokenizer.tokenize(" ▁")
+        self.assertEqual(tokens, ["▁▁"])
+        # a dummy prefix space is not added by the sp_model as it was de-activated
+        self.assertEqual(tokens, tokenizer.sp_model.encode("▁▁", out_type=str))
+
+    @require_jinja
+    def test_tokenization_for_chat(self):
+        tokenizer = GemmaTokenizer.from_pretrained("hf-internal-testing/dummy-gemma")
+
+        test_chats = [
+            [{"role": "user", "content": "Hello!"}],
+            [
+                {"role": "user", "content": "Hello!"},
+                {"role": "assistant", "content": "Nice to meet you."},
+            ],
+            [{"role": "user", "content": "Hello!"}],
+        ]
+        # Matt: The third test case tests the default system message, but if this is ever changed in the
+        #       class/repo code then that test will fail, and the case will need to be updated.
+        tokenized_chats = [tokenizer.apply_chat_template(test_chat) for test_chat in test_chats]
+        expected_tokens = [[235322, 235371, 571, 235298, 2997, 73786, 1645, 108, 4521, 149907, 235371, 571, 235298, 615, 73786, 108], [235322, 235371, 571, 235298, 2997, 73786, 1645, 108, 4521, 149907, 235371, 571, 235298, 615, 73786, 108, 235322, 235371, 571, 235298, 2997, 73786, 105776, 108, 7731, 577, 4664, 692, 35606, 235371, 571, 235298, 615, 73786, 108], [235322, 235371, 571, 235298, 2997, 73786, 1645, 108, 4521, 149907, 235371, 571, 235298, 615, 73786, 108]]  # fmt: skip
+        for tokenized_chat, expected_tokens in zip(tokenized_chats, expected_tokens):
+            self.assertListEqual(tokenized_chat, expected_tokens)
+
+
+@require_sentencepiece
+@require_tokenizers
+class CommonSpmIntegrationTests(unittest.TestCase):
+    """
+    A class that regroups important test to make sure that we properly handle the special tokens.
+    """
+
+    def test_edge_case_tabulation(self):
+        fast_tokenizer = GemmaTokenizerFast.from_pretrained("hf-internal-testing/dummy-gemma")
+        slow_tokenizer = GemmaTokenizer.from_pretrained("hf-internal-testing/dummy-gemma")
+        input_text = "Hey<eos>. \t\t \n\nyou  é  @#😈  🤗!       , 1234 15 5,61"
+        EXPECTED_IDS = [ 2, 6750, 1, 235265, 235248, 255969, 235248, 109, 4747, 139, 235335, 139, 216311, 241316, 139, 239880, 235341, 144, 235269, 235248, 235274, 235284, 235304, 235310, 235248, 235274, 235308, 235248, 235308, 235269, 235318, 235274]  # fmt: skip
+        EXPECTED_TOKENS = [ "Hey", "<eos>", ".", "▁", "\t\t", "▁", "\n\n", "you", "▁▁", "é", "▁▁", "@#", "😈", "▁▁", "🤗", "!", "▁▁▁▁▁▁▁", ",", "▁", "1", "2", "3", "4", "▁", "1", "5", "▁", "5", ",", "6", "1"]  # fmt: skip
+
+        tokens = fast_tokenizer.tokenize(input_text)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(tokens, EXPECTED_TOKENS)
+
+        tokens = slow_tokenizer.tokenize(input_text)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(tokens, EXPECTED_TOKENS)
+
+        input_ids = fast_tokenizer.encode(input_text)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(input_ids, EXPECTED_IDS)
+
+        input_ids = slow_tokenizer.encode(input_text)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(input_ids, EXPECTED_IDS)
+
+        text = fast_tokenizer.decode(EXPECTED_IDS)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(text, "<bos>Hey<eos>. \t\t \n\nyou  é  @#😈  🤗!       , 1234 15 5,61")
+
+        text = slow_tokenizer.decode(EXPECTED_IDS)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(text, "<bos>Hey<eos>. \t\t \n\nyou  é  @#😈  🤗!       , 1234 15 5,61")
+
+        input_text = "\t\t\t\t \n\n61"
+        EXPECTED_IDS = [2, 255971, 235248, 109, 235318, 235274]
+        EXPECTED_TOKENS = ["\t\t\t\t", "▁", "\n\n", "6", "1"]
+
+        tokens = fast_tokenizer.tokenize(input_text)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(tokens, EXPECTED_TOKENS)
+
+        tokens = slow_tokenizer.tokenize(input_text)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(tokens, EXPECTED_TOKENS)
+
+        input_ids = fast_tokenizer.encode(input_text)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(input_ids, EXPECTED_IDS)
+
+        input_ids = slow_tokenizer.encode(input_text)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(input_ids, EXPECTED_IDS)
+
+        text = fast_tokenizer.decode(EXPECTED_IDS)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(text, "<bos>\t\t\t\t \n\n61")
+
+        text = slow_tokenizer.decode(EXPECTED_IDS)
+        with self.subTest("test fast edge case fast"):
+            self.assertEqual(text, "<bos>\t\t\t\t \n\n61")

utils/check_docstrings.py

@@ -233,6 +233,8 @@ OBJECTS_TO_IGNORE = [
     "FlaxGPTNeoModel",
     "FlaxLlamaForCausalLM",
     "FlaxLlamaModel",
+    "FlaxGemmaForCausalLM",
+    "FlaxGemmaModel",
     "FlaxMBartForConditionalGeneration",
     "FlaxMBartForQuestionAnswering",
     "FlaxMBartForSequenceClassification",

utils/not_doctested.txt

@@ -575,6 +575,10 @@ src/transformers/models/funnel/convert_funnel_original_tf_checkpoint_to_pytorch.
 src/transformers/models/funnel/modeling_funnel.py
 src/transformers/models/funnel/modeling_tf_funnel.py
 src/transformers/models/fuyu/convert_fuyu_model_weights_to_hf.py
+src/transformers/models/gemma/configuration_gemma.py
+src/transformers/models/gemma/convert_gemma_weights_to_hf.py
+src/transformers/models/gemma/modeling_flax_gemma.py
+src/transformers/models/gemma/modeling_gemma.py
 src/transformers/models/git/configuration_git.py
 src/transformers/models/git/convert_git_to_pytorch.py
 src/transformers/models/glpn/configuration_glpn.py