exaone.py

# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project

# Adapted from
# https://huggingface.co/LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct/blob/main/modeling_exaone.py
# Copyright 2024 The LG U+ CTO AI Tech Lab.
# Copyright 2021 The LG AI Research EXAONE Lab
# Copyright (c) 2018, NVIDIA CORPORATION.  All rights reserved.
#
# This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
# and OPT implementations in this library. It has been modified from its
# original forms to accommodate minor architectural differences compared
# to GPT-NeoX and OPT used by the Meta AI team that trained the model.
#
# 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.
"""Inference-only Exaone model compatible with HuggingFace weights."""

from collections.abc import Iterable
from itertools import islice
from typing import Any, Optional, Union

import torch
from torch import nn
from transformers import PretrainedConfig

from vllm.attention import Attention
from vllm.compilation.decorators import support_torch_compile
from vllm.config import CacheConfig, VllmConfig
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
from vllm.model_executor.layers.activation import SiluAndMul
from vllm.model_executor.layers.layernorm import RMSNorm
from vllm.model_executor.layers.linear import (
    MergedColumnParallelLinear,
    QKVParallelLinear,
    RowParallelLinear,
)
from vllm.model_executor.layers.logits_processor import LogitsProcessor
from vllm.model_executor.layers.quantization import QuantizationConfig
from vllm.model_executor.layers.rotary_embedding import get_rope
from vllm.model_executor.layers.vocab_parallel_embedding import (
    DEFAULT_VOCAB_PADDING_SIZE,
    ParallelLMHead,
    VocabParallelEmbedding,
)
from vllm.model_executor.model_loader.weight_utils import (
    default_weight_loader,
    maybe_remap_kv_scale_name,
)
from vllm.sequence import IntermediateTensors

from .interfaces import SupportsLoRA, SupportsPP
from .utils import (
    AutoWeightsLoader,
    PPMissingLayer,
    is_pp_missing_parameter,
    make_empty_intermediate_tensors_factory,
    make_layers,
    maybe_prefix,
)


class ExaoneGatedMLP(nn.Module):
    def __init__(
        self,
        hidden_size: int,
        intermediate_size: int,
        hidden_act: str,
        quant_config: Optional[QuantizationConfig] = None,
        bias: bool = False,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.gate_up_proj = MergedColumnParallelLinear(
            input_size=hidden_size,
            output_sizes=[intermediate_size] * 2,
            bias=bias,
            quant_config=quant_config,
            prefix=f"{prefix}.gate_up_proj",
        )
        self.c_proj = RowParallelLinear(
            input_size=intermediate_size,
            output_size=hidden_size,
            bias=bias,
            quant_config=quant_config,
            prefix=f"{prefix}.c_proj",
        )
        if hidden_act != "silu":
            raise ValueError(
                f"Unsupported activation: {hidden_act}. Only silu is supported for now."
            )
        self.act_fn = SiluAndMul()

    def forward(self, x):
        gate_up, _ = self.gate_up_proj(x)
        x = self.act_fn(gate_up)
        x, _ = self.c_proj(x)
        return x


class ExaoneAttention(nn.Module):
    def __init__(
        self,
        config: PretrainedConfig,
        hidden_size: int,
        num_heads: int,
        num_kv_heads: int,
        rope_theta: float = 10000,
        rope_scaling: Optional[dict[str, Any]] = None,
        max_position_embeddings: int = 8192,
        quant_config: Optional[QuantizationConfig] = None,
        bias: bool = False,
        cache_config: Optional[CacheConfig] = None,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.hidden_size = hidden_size
        tp_size = get_tensor_model_parallel_world_size()
        self.total_num_heads = num_heads
        assert self.total_num_heads % tp_size == 0
        self.num_heads = self.total_num_heads // tp_size
        self.total_num_kv_heads = num_kv_heads
        if self.total_num_kv_heads >= tp_size:
            # Number of KV heads is greater than TP size, so we partition
            # the KV heads across multiple tensor parallel GPUs.
            assert self.total_num_kv_heads % tp_size == 0
        else:
            # Number of KV heads is less than TP size, so we replicate
            # the KV heads across multiple tensor parallel GPUs.
            assert tp_size % self.total_num_kv_heads == 0
        self.num_kv_heads = max(1, self.total_num_kv_heads // tp_size)
        # MistralConfig has an optional head_dim introduced by Mistral-Nemo
        self.head_dim = getattr(config, "head_dim", None)
        if self.head_dim is None:
            self.head_dim = self.hidden_size // self.total_num_heads
        self.q_size = self.num_heads * self.head_dim
        self.kv_size = self.num_kv_heads * self.head_dim
        self.scaling = self.head_dim**-0.5
        self.rope_theta = rope_theta
        self.max_position_embeddings = max_position_embeddings

        self.qkv_proj = QKVParallelLinear(
            hidden_size=hidden_size,
            head_size=self.head_dim,
            total_num_heads=self.total_num_heads,
            total_num_kv_heads=self.total_num_kv_heads,
            bias=bias,
            quant_config=quant_config,
            prefix=f"{prefix}.qkv_proj",
        )

        self.out_proj = RowParallelLinear(
            input_size=self.total_num_heads * self.head_dim,
            output_size=hidden_size,
            bias=bias,
            quant_config=quant_config,
            prefix=f"{prefix}.out_proj",
        )

        is_neox_style = True
        if quant_config is not None and quant_config.get_name() == "gguf":
            is_neox_style = False

        self.rotary_emb = get_rope(
            self.head_dim,
            rotary_dim=self.head_dim,
            max_position=max_position_embeddings,
            base=rope_theta,
            rope_scaling=rope_scaling,
            is_neox_style=is_neox_style,
        )
        self.attn = Attention(
            self.num_heads,
            self.head_dim,
            self.scaling,
            num_kv_heads=self.num_kv_heads,
            cache_config=cache_config,
            quant_config=quant_config,
            prefix=f"{prefix}.attn",
        )

    def forward(
        self,
        positions: torch.Tensor,
        hidden_states: torch.Tensor,
    ) -> torch.Tensor:
        qkv, _ = self.qkv_proj(hidden_states)
        q, k, v = qkv.split([self.q_size, self.kv_size, self.kv_size], dim=-1)
        q, k = self.rotary_emb(positions, q, k)
        attn_output = self.attn(q, k, v)
        output, _ = self.out_proj(attn_output)
        return output


class ExaoneBlockAttention(nn.Module):
    def __init__(
        self,
        config: PretrainedConfig,
        hidden_size: int,
        num_heads: int,
        num_kv_heads: int,
        rope_theta: float = 10000,
        rope_scaling: Optional[dict[str, Any]] = None,
        max_position_embeddings: int = 8192,
        quant_config: Optional[QuantizationConfig] = None,
        bias: bool = False,
        cache_config: Optional[CacheConfig] = None,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.attention = ExaoneAttention(
            config=config,
            hidden_size=hidden_size,
            num_heads=num_heads,
            num_kv_heads=num_kv_heads,
            rope_theta=rope_theta,
            rope_scaling=rope_scaling,
            max_position_embeddings=max_position_embeddings,
            quant_config=quant_config,
            bias=bias,
            cache_config=cache_config,
            prefix=f"{prefix}.attention",
        )

    def forward(
        self,
        positions: torch.Tensor,
        hidden_states: torch.Tensor,
    ) -> torch.Tensor:
        return self.attention(
            positions=positions,
            hidden_states=hidden_states,
        )


class ExaoneDecoderLayer(nn.Module):
    def __init__(
        self,
        config: PretrainedConfig,
        cache_config: Optional[CacheConfig] = None,
        quant_config: Optional[QuantizationConfig] = None,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.hidden_size = config.hidden_size
        rope_theta = getattr(config, "rope_theta", 10000)
        rope_scaling = getattr(config, "rope_scaling", None)
        if rope_scaling is not None and getattr(
            config, "original_max_position_embeddings", None
        ):
            rope_scaling["original_max_position_embeddings"] = (
                config.original_max_position_embeddings
            )
        max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
        # Support abacusai/Smaug-72B-v0.1 with attention_bias
        # Support internlm/internlm-7b with bias
        attention_bias = getattr(config, "attention_bias", False) or getattr(
            config, "bias", False
        )
        self.attn = ExaoneBlockAttention(
            config=config,
            hidden_size=self.hidden_size,
            num_heads=config.num_attention_heads,
            num_kv_heads=getattr(
                config, "num_key_value_heads", config.num_attention_heads
            ),
            rope_theta=rope_theta,
            rope_scaling=rope_scaling,
            max_position_embeddings=max_position_embeddings,
            quant_config=quant_config,
            bias=attention_bias,
            cache_config=cache_config,
            prefix=f"{prefix}.attn",
        )
        self.mlp = ExaoneGatedMLP(
            hidden_size=self.hidden_size,
            intermediate_size=config.intermediate_size,
            hidden_act=config.activation_function,
            quant_config=quant_config,
            bias=getattr(config, "mlp_bias", False),
            prefix=f"{prefix}.mlp",
        )
        self.ln_1 = RMSNorm(config.hidden_size, eps=config.layer_norm_epsilon)
        self.ln_2 = RMSNorm(config.hidden_size, eps=config.layer_norm_epsilon)

    def forward(
        self,
        positions: torch.Tensor,
        hidden_states: torch.Tensor,
        residual: Optional[torch.Tensor],
    ) -> tuple[torch.Tensor, torch.Tensor]:
        # Self Attention
        if residual is None:
            residual = hidden_states
            hidden_states = self.ln_1(hidden_states)
        else:
            hidden_states, residual = self.ln_1(hidden_states, residual)
        hidden_states = self.attn(
            positions=positions,
            hidden_states=hidden_states,
        )

        # Fully Connected
        hidden_states, residual = self.ln_2(hidden_states, residual)
        hidden_states = self.mlp(hidden_states)
        return hidden_states, residual


@support_torch_compile
class ExaoneModel(nn.Module):
    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
        super().__init__()

        config = vllm_config.model_config.hf_config
        cache_config = vllm_config.cache_config
        quant_config = vllm_config.quant_config
        lora_config = vllm_config.lora_config

        self.config = config
        self.quant_config = quant_config
        lora_vocab = (
            (lora_config.lora_extra_vocab_size * (lora_config.max_loras or 1))
            if lora_config
            else 0
        )
        self.vocab_size = config.vocab_size + lora_vocab
        self.wte = config.vocab_size
        if get_pp_group().is_first_rank or (
            config.tie_word_embeddings and get_pp_group().is_last_rank
        ):
            self.wte = VocabParallelEmbedding(
                self.vocab_size,
                config.hidden_size,
                org_num_embeddings=config.vocab_size,
                quant_config=quant_config,
            )
        else:
            self.wte = PPMissingLayer()
        self.start_layer, self.end_layer, self.h = make_layers(
            config.num_hidden_layers,
            lambda prefix: ExaoneDecoderLayer(
                config=config,
                cache_config=cache_config,
                quant_config=quant_config,
                prefix=prefix,
            ),
            prefix=f"{prefix}.h",
        )
        if get_pp_group().is_last_rank:
            self.ln_f = RMSNorm(config.hidden_size, eps=config.layer_norm_epsilon)
        else:
            self.ln_f = PPMissingLayer()

        self.make_empty_intermediate_tensors = make_empty_intermediate_tensors_factory(
            ["hidden_states", "residual"], config.hidden_size
        )

    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
        return self.wte(input_ids)

    def forward(
        self,
        input_ids: Optional[torch.Tensor],
        positions: torch.Tensor,
        intermediate_tensors: Optional[IntermediateTensors],
        inputs_embeds: Optional[torch.Tensor] = None,
    ) -> Union[torch.Tensor, IntermediateTensors]:
        if get_pp_group().is_first_rank:
            if inputs_embeds is not None:
                hidden_states = inputs_embeds
            else:
                hidden_states = self.get_input_embeddings(input_ids)
            residual = None
        else:
            assert intermediate_tensors is not None
            hidden_states = intermediate_tensors["hidden_states"]
            residual = intermediate_tensors["residual"]

        for layer in islice(self.h, self.start_layer, self.end_layer):
            hidden_states, residual = layer(
                positions,
                hidden_states,
                residual,
            )

        if not get_pp_group().is_last_rank:
            return IntermediateTensors(
                {"hidden_states": hidden_states, "residual": residual}
            )

        hidden_states, _ = self.ln_f(hidden_states, residual)
        return hidden_states

    def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
        stacked_params_mapping = [
            # (param_name, shard_name, shard_id)
            (".qkv_proj", ".q_proj", "q"),
            (".qkv_proj", ".k_proj", "k"),
            (".qkv_proj", ".v_proj", "v"),
            (".gate_up_proj", ".c_fc_0", 0),
            (".gate_up_proj", ".c_fc_1", 1),
        ]
        params_dict = dict(self.named_parameters())
        loaded_params: set[str] = set()
        for name, loaded_weight in weights:
            if "rotary_emb.inv_freq" in name:
                continue
            if "rotary_emb.cos_cached" in name or "rotary_emb.sin_cached" in name:
                # Models trained using ColossalAI may include these tensors in
                # the checkpoint. Skip them.
                continue
            if self.quant_config is not None and (
                scale_name := self.quant_config.get_cache_scale(name)
            ):
                # Loading kv cache quantization scales
                param = params_dict[scale_name]
                weight_loader = getattr(param, "weight_loader", default_weight_loader)
                loaded_weight = (
                    loaded_weight if loaded_weight.dim() == 0 else loaded_weight[0]
                )
                weight_loader(param, loaded_weight)
                loaded_params.add(scale_name)
                continue
            for param_name, weight_name, shard_id in stacked_params_mapping:
                if weight_name not in name:
                    continue
                name = name.replace(weight_name, param_name)
                # Skip loading extra bias for GPTQ models.
                if name.endswith(".bias") and name not in params_dict:
                    continue

                if is_pp_missing_parameter(name, self):
                    continue

                param = params_dict[name]
                weight_loader = param.weight_loader
                weight_loader(param, loaded_weight, shard_id)

                break
            else:
                # Skip loading extra bias for GPTQ models.
                if name.endswith(".bias") and name not in params_dict:
                    continue
                # Remapping the name of FP8 kv-scale.
                name = maybe_remap_kv_scale_name(name, params_dict)
                if name is None:
                    continue

                if is_pp_missing_parameter(name, self):
                    continue

                param = params_dict[name]
                weight_loader = getattr(param, "weight_loader", default_weight_loader)
                weight_loader(param, loaded_weight)
            loaded_params.add(name)
        return loaded_params


class ExaoneForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
    packed_modules_mapping = {
        "qkv_proj": [
            "q_proj",
            "k_proj",
            "v_proj",
        ],
        "gate_up_proj": [
            "c_fc_0",
            "c_fc_1",
        ],
    }

    # LoRA specific attributes
    embedding_modules = {
        "wte": "input_embeddings",
        "lm_head": "output_embeddings",
    }
    embedding_padding_modules = ["lm_head"]

    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
        super().__init__()
        config = vllm_config.model_config.hf_config
        quant_config = vllm_config.quant_config
        lora_config = vllm_config.lora_config

        self.config = config
        self.lora_config = lora_config
        self.quant_config = quant_config

        self.transformer = ExaoneModel(
            vllm_config=vllm_config,
            prefix=maybe_prefix(prefix, "model"),
        )
        if get_pp_group().is_last_rank:
            self.unpadded_vocab_size = config.vocab_size
            if lora_config:
                self.unpadded_vocab_size += lora_config.lora_extra_vocab_size
            self.lm_head = ParallelLMHead(
                self.unpadded_vocab_size,
                config.hidden_size,
                org_num_embeddings=config.vocab_size,
                padding_size=DEFAULT_VOCAB_PADDING_SIZE
                # We need bigger padding if using lora for kernel
                # compatibility
                if not lora_config
                else lora_config.lora_vocab_padding_size,
                quant_config=quant_config,
                prefix=maybe_prefix(prefix, "lm_head"),
            )
            if config.tie_word_embeddings:
                self.lm_head.weight = self.transformer.wte.weight

            logit_scale = getattr(config, "logit_scale", 1.0)
            self.logits_processor = LogitsProcessor(
                self.unpadded_vocab_size, config.vocab_size, logit_scale
            )
        else:
            self.lm_head = PPMissingLayer()

        self.make_empty_intermediate_tensors = (
            self.transformer.make_empty_intermediate_tensors
        )

    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
        return self.model.get_input_embeddings(input_ids)

    def forward(
        self,
        input_ids: torch.Tensor,
        positions: torch.Tensor,
        intermediate_tensors: Optional[IntermediateTensors] = None,
        inputs_embeds: Optional[torch.Tensor] = None,
    ) -> Union[torch.Tensor, IntermediateTensors]:
        model_output = self.transformer(
            input_ids, positions, intermediate_tensors, inputs_embeds
        )
        return model_output

    def compute_logits(
        self,
        hidden_states: torch.Tensor,
    ) -> Optional[torch.Tensor]:
        logits = self.logits_processor(self.lm_head, hidden_states)
        return logits

    def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
        loader = AutoWeightsLoader(
            self,
            # With tie_word_embeddings, we can skip lm_head.weight
            # The weight might appear unnecessarily in the files if the model is
            # processed with quantization, LoRA, fine-tuning, etc.
            skip_prefixes=(["lm_head."] if self.config.tie_word_embeddings else None),
        )
        return loader.load_weights(weights)