Seq2SeqDataset uses linecache to save memory by @Pradhy729 (#5792)

Co-authored-by: Pradhy729 <49659913+Pradhy729@users.noreply.github.com>

examples/seq2seq/README.md

@@ -7,27 +7,24 @@ For `bertabs` instructions, see `bertabs/README.md`.
 
 
 ### Data
-
-CNN/DailyMail data
+XSUM Data:
 ```bash
 cd examples/seq2seq
-wget https://s3.amazonaws.com/datasets.huggingface.co/summarization/cnn_dm.tgz
-tar -xzvf cnn_dm.tgz
-
-export CNN_DIR=${PWD}/cnn_dm
+wget https://s3.amazonaws.com/datasets.huggingface.co/summarization/xsum.tar.gz
+tar -xzvf xsum.tar.gz
+export XSUM_DIR=${PWD}/xsum
 ```
-
 this should make a directory called cnn_dm/ with files like `test.source`.
 To use your own data, copy that files format. Each article to be summarized is on its own line.
 
-XSUM Data:
+CNN/DailyMail data
 ```bash
 cd examples/seq2seq
-wget https://s3.amazonaws.com/datasets.huggingface.co/summarization/xsum.tar.gz
-tar -xzvf xsum.tar.gz
-export XSUM_DIR=${PWD}/xsum
-```
+wget https://s3.amazonaws.com/datasets.huggingface.co/summarization/cnn_dm.tgz
+tar -xzvf cnn_dm.tgz
 
+export CNN_DIR=${PWD}/cnn_dm
+```
 
 WMT16 English-Romanian Translation Data:
 ```bash
@@ -64,6 +61,10 @@ Summarization Tips:
 - If you are finetuning on your own dataset, start from `distilbart-cnn-12-6` if you want long summaries and `distilbart-xsum-12-6` if you want short summaries.
 (It rarely makes sense to start from `bart-large` unless you are a researching finetuning methods). 
 
+**Update 2018-07-18**
+Datasets: Seq2SeqDataset will be used for all models besides MBart, for which MBartDataset will be used.**
+A new dataset is needed to support multilingual tasks.
+
 ### Summarization Finetuning
 Run/modify `finetune.sh`
 
@@ -78,8 +79,6 @@ The following command should work on a 16GB GPU:
     --model_name_or_path facebook/bart-large
 ```
 
-
-
 ### Translation Finetuning
 
 First, follow the wmt_en_ro download instructions.
@@ -124,23 +123,6 @@ from transformers import AutoModelForSeq2SeqLM
 model = AutoModelForSeq2SeqLM.from_pretrained(f'{output_dir}/best_tfmr')
 ```
 
-#### XSUM Shared Task
-Compare XSUM results with others by using `--logger_name wandb_shared`. This requires `wandb` registration.
-
-Here is an example command, but you can do whatever you want. Hopefully this will make debugging and collaboration easier!
-```bash
-WANDB_PROJECT='hf_xsum' ./finetune.sh \
-    --data_dir $XSUM_DIR \
-    --output_dir xsum_frozen_embs \
-    --model_name_or_path facebook/bart-large \
-    --train_batch_size 16 --eval_batch_size 16 --freeze_embeds --freeze_encoder \
-    --num_train_epochs 6 \
-    --max_target_length=60 --val_max_target_length=60 --test_max_target_length=100 \
-    --logger_name wandb
-```
-
-You can see your wandb logs [here](https://app.wandb.ai/sshleifer/hf_xsum?workspace=user-)
-
 ### Evaluation Commands
 
 To create summaries for each article in dataset, we use `run_eval.py`, here are a few commands that run eval for different tasks and models.

examples/seq2seq/distillation.py

@@ -15,28 +15,15 @@ from transformers import AdamW, BartConfig, BartForConditionalGeneration, T5Conf
 
 try:
     from .finetune import SummarizationModule
-    from .initialization_utils import init_student, copy_layers
-    from .utils import (
-        use_task_specific_params,
-        SummarizationDataset,
-        pickle_load,
-        freeze_params,
-        assert_all_frozen,
-        any_requires_grad,
-    )
     from .finetune import main as ft_main
+    from .initialization_utils import init_student, copy_layers
+    from .utils import use_task_specific_params, pickle_load, freeze_params, assert_all_frozen, any_requires_grad
+
 except ImportError:
     from finetune import SummarizationModule
     from finetune import main as ft_main
     from initialization_utils import init_student, copy_layers
-    from utils import (
-        use_task_specific_params,
-        SummarizationDataset,
-        pickle_load,
-        freeze_params,
-        assert_all_frozen,
-        any_requires_grad,
-    )
+    from utils import use_task_specific_params, pickle_load, freeze_params, assert_all_frozen, any_requires_grad
 
 
 class BartSummarizationDistiller(SummarizationModule):
@@ -115,11 +102,6 @@ class BartSummarizationDistiller(SummarizationModule):
         if self.different_encoder:
             copy_layers(teacher.encoder.block, student.encoder.block, e_layers_to_copy)
 
-    def get_dataset(self, type_path) -> SummarizationDataset:
-        n_obs = self.n_obs[type_path]
-        dataset = SummarizationDataset(self.tokenizer, type_path=type_path, n_obs=n_obs, **self.dataset_kwargs)
-        return dataset
-
     def calc_mse_loss(self, teacher_outputs: torch.Tensor, student_outputs: torch.Tensor, mask) -> torch.FloatTensor:
         if mask is not None:
             # mask has False at padding_idx

examples/seq2seq/finetune.py

@@ -21,7 +21,6 @@ try:
     from .utils import (
         assert_all_frozen,
         use_task_specific_params,
-        SummarizationDataset,
         lmap,
         flatten_list,
         pickle_save,
@@ -32,12 +31,17 @@ try:
         get_git_info,
         ROUGE_KEYS,
         calculate_bleu_score,
+        Seq2SeqDataset,
+        MBartDataset,
     )
+
     from .callbacks import Seq2SeqLoggingCallback, get_checkpoint_callback
 except ImportError:
     from utils import (
+        Seq2SeqDataset,
+        MBartDataset,
+        assert_all_frozen,
         use_task_specific_params,
-        SummarizationDataset,
         lmap,
         flatten_list,
         pickle_save,
@@ -48,7 +52,6 @@ except ImportError:
         get_git_info,
         ROUGE_KEYS,
         calculate_bleu_score,
-        assert_all_frozen,
     )
     from callbacks import Seq2SeqLoggingCallback, get_checkpoint_callback
 
@@ -100,6 +103,7 @@ class SummarizationModule(BaseTransformer):
         self.hparams.git_sha = get_git_info()["repo_sha"]
         self.num_workers = hparams.num_workers
         self.decoder_start_token_id = None
+        self.dataset_class = Seq2SeqDataset
 
     def freeze_embeds(self):
         """Freeze token embeddings and positional embeddings for bart, just token embeddings for t5."""
@@ -163,7 +167,7 @@ class SummarizationModule(BaseTransformer):
 
     def _generative_step(self, batch: dict) -> dict:
         pad_token_id = self.tokenizer.pad_token_id
-        source_ids, source_mask, y = SummarizationDataset.trim_seq2seq_batch(batch, pad_token_id)
+        source_ids, source_mask, y = Seq2SeqDataset.trim_seq2seq_batch(batch, pad_token_id)
         t0 = time.time()
         generated_ids = self.model.generate(
             input_ids=source_ids,
@@ -187,10 +191,10 @@ class SummarizationModule(BaseTransformer):
     def test_epoch_end(self, outputs):
         return self.validation_epoch_end(outputs, prefix="test")
 
-    def get_dataset(self, type_path) -> SummarizationDataset:
+    def get_dataset(self, type_path) -> Seq2SeqDataset:
         n_obs = self.n_obs[type_path]
         max_target_length = self.target_lens[type_path]
-        dataset = SummarizationDataset(
+        dataset = self.dataset_class(
             self.tokenizer,
             type_path=type_path,
             n_obs=n_obs,
@@ -303,6 +307,8 @@ class TranslationModule(SummarizationModule):
         self.dataset_kwargs["tgt_lang"] = hparams.tgt_lang
         if self.model.config.decoder_start_token_id is None and isinstance(self.tokenizer, MBartTokenizer):
             self.decoder_start_token_id = self.tokenizer.lang_code_to_id[hparams.tgt_lang]
+        if isinstance(self.tokenizer, MBartTokenizer):
+            self.dataset_class = MBartDataset
 
     def calc_generative_metrics(self, preds, target) -> dict:
         return calculate_bleu_score(preds, target)

examples/seq2seq/test_seq2seq_examples.py

@@ -9,16 +9,17 @@ from unittest.mock import patch
 
 import pytest
 import torch
+from pytest import param
 from torch.utils.data import DataLoader
 
-from transformers import AutoTokenizer
+from transformers import AutoTokenizer, MBartTokenizer
 from transformers.testing_utils import require_multigpu
 
 from .distillation import distill_main, evaluate_checkpoint
 from .finetune import main
 from .pack_dataset import pack_data_dir
 from .run_eval import generate_summaries_or_translations, run_generate
-from .utils import SummarizationDataset, lmap, load_json
+from .utils import MBartDataset, Seq2SeqDataset, lmap, load_json
 
 
 logging.basicConfig(level=logging.DEBUG)
@@ -26,6 +27,7 @@ logging.basicConfig(level=logging.DEBUG)
 logger = logging.getLogger()
 CUDA_AVAILABLE = torch.cuda.is_available()
 CHEAP_ARGS = {
+    "label_smoothing_eps": 0.2,
     "logger_name": "default",
     "length_penalty": 0.5,
     "cache_dir": "",
@@ -80,11 +82,11 @@ CHEAP_ARGS = {
 
 
 def _dump_articles(path: Path, articles: list):
-    with path.open("w") as f:
-        f.write("\n".join(articles))
+    content = "\n".join(articles)
+    Path(path).open("w").writelines(content)
 
 
-ARTICLES = [" Sam ate lunch today", "Sams lunch ingredients"]
+ARTICLES = [" Sam ate lunch today.", "Sams lunch ingredients."]
 SUMMARIES = ["A very interesting story about what I ate for lunch.", "Avocado, celery, turkey, coffee"]
 T5_TINY = "patrickvonplaten/t5-tiny-random"
 BART_TINY = "sshleifer/bart-tiny-random"
@@ -208,7 +210,7 @@ def test_run_eval_bart(model):
 
 
 @pytest.mark.parametrize(
-    ["model"], [pytest.param(T5_TINY), pytest.param(BART_TINY), pytest.param(MBART_TINY), pytest.param(MARIAN_TINY)]
+    ["model"], [pytest.param(T5_TINY), pytest.param(BART_TINY), pytest.param(MBART_TINY), pytest.param(MARIAN_TINY)],
 )
 def test_finetune(model):
     args_d: dict = CHEAP_ARGS.copy()
@@ -260,22 +262,50 @@ def test_pack_dataset():
     assert orig_paths == new_paths
 
 
-@pytest.mark.parametrize(
-    ["tok"], [pytest.param(T5_TINY), pytest.param(BART_TINY), pytest.param(MBART_TINY), pytest.param(MARIAN_TINY)]
-)
-def test_dataset(tok):
-    tokenizer = AutoTokenizer.from_pretrained(tok)
+def test_mbart_dataset_truncation():
+    tokenizer = MBartTokenizer.from_pretrained(MBART_TINY)
     tmp_dir = make_test_data_dir()
     max_len_source = max(len(tokenizer.encode(a)) for a in ARTICLES)
     max_len_target = max(len(tokenizer.encode(a)) for a in SUMMARIES)
-    trunc_target = 4
-    train_dataset = SummarizationDataset(
+    trunc = 4
+    src_lang, tgt_lang = "ro_RO", "de_DE"  # NOT WHAT IT WAS TRAINED ON
+    train_dataset = MBartDataset(
         tokenizer,
         data_dir=tmp_dir,
         type_path="train",
-        max_source_length=20,
-        max_target_length=trunc_target,
-        tgt_lang="ro_RO",
+        max_source_length=trunc,
+        max_target_length=1000,  # ignored
+        src_lang=src_lang,
+        tgt_lang=tgt_lang,
+    )
+    dataloader = DataLoader(train_dataset, batch_size=2, collate_fn=train_dataset.collate_fn)
+    for batch in dataloader:
+        assert isinstance(batch, dict)
+        assert batch["attention_mask"].shape == batch["input_ids"].shape
+        # show that articles were trimmed.
+        assert batch["input_ids"].shape[1] == trunc
+        # show that targets are the same len
+        assert batch["decoder_input_ids"].shape[1] == trunc
+        # check language codes in correct place
+        assert batch["decoder_input_ids"][0, 0].item() == tokenizer.lang_code_to_id[tgt_lang]
+        assert batch["decoder_input_ids"][0, -1].item() == tokenizer.eos_token_id
+        assert batch["input_ids"][0, -2].item() == tokenizer.eos_token_id
+        assert batch["input_ids"][0, -1].item() == tokenizer.lang_code_to_id[src_lang]
+
+        assert max_len_target > trunc  # Truncated
+        assert max_len_source > trunc
+        break  # No need to test every batch
+
+
+@pytest.mark.parametrize(["tok"], [pytest.param(T5_TINY), pytest.param(BART_TINY), param(MARIAN_TINY)])
+def test_summarization_dataset_truncation(tok):
+    tokenizer = AutoTokenizer.from_pretrained(tok)
+    tmp_dir = make_test_data_dir()
+    max_len_source = max(len(tokenizer.encode(a)) for a in ARTICLES)
+    max_len_target = max(len(tokenizer.encode(a)) for a in SUMMARIES)
+    trunc_target = 4
+    train_dataset = Seq2SeqDataset(
+        tokenizer, data_dir=tmp_dir, type_path="train", max_source_length=20, max_target_length=trunc_target,
     )
     dataloader = DataLoader(train_dataset, batch_size=2, collate_fn=train_dataset.collate_fn)
     for batch in dataloader:
@@ -286,3 +316,4 @@ def test_dataset(tok):
         # show that targets were truncated
         assert batch["decoder_input_ids"].shape[1] == trunc_target  # Truncated
         assert max_len_target > trunc_target  # Truncated
+        break  # No need to test every batch

examples/seq2seq/utils.py

 import itertools
 import json
+import linecache
 import os
 import pickle
+import warnings
 from logging import getLogger
 from pathlib import Path
 from typing import Callable, Dict, Iterable, List
@@ -13,50 +15,20 @@ from rouge_score import rouge_scorer, scoring
 from sacrebleu import corpus_bleu
 from torch import nn
 from torch.utils.data import Dataset, Sampler
-from tqdm import tqdm
 
 from transformers import BartTokenizer
 
 
-def encode_file(
-    tokenizer,
-    data_path,
-    max_length,
-    pad_to_max_length=True,
-    return_tensors="pt",
-    overwrite_cache=False,
-    prefix="",
-    tok_name="",
-):
+def encode_line(tokenizer, line, max_length, pad_to_max_length=True, return_tensors="pt"):
     extra_kw = {"add_prefix_space": True} if isinstance(tokenizer, BartTokenizer) else {}
-    cache_path = Path(f"{data_path}_{tok_name}{max_length}.pt")
-    if not overwrite_cache and cache_path.exists():
-        try:
-            examples = torch.load(cache_path)
-            assert isinstance(examples, list)
-            return examples
-
-        except Exception:
-            print(f"failed to load from {cache_path}, retokenizing {data_path}")
-    data_path = Path(data_path)
-
-    lns = lmap(str.strip, data_path.open().readlines())
-    lns = [prefix + text for text in lns]
-    assert lns, f"found empty file at {data_path}"
-    examples = []
-    for text in tqdm(lns, desc=f"Tokenizing {data_path.name}"):
-        tokenized = tokenizer(
-            [text],
+    return tokenizer(
+        [line],
         max_length=max_length,
         padding="max_length" if pad_to_max_length else None,
         truncation=True,
         return_tensors=return_tensors,
         **extra_kw,
     )
-        assert tokenized.input_ids.shape[1] == max_length
-        examples.append(tokenized)
-    torch.save(lmap(dict, examples), cache_path.open("wb"))
-    return examples
 
 
 def lmap(f: Callable, x: Iterable) -> List:
@@ -80,73 +52,111 @@ def trim_batch(
         return (input_ids[:, keep_column_mask], attention_mask[:, keep_column_mask])
 
 
-class SummarizationDataset(Dataset):
+class Seq2SeqDataset(Dataset):
     def __init__(
         self,
         tokenizer,
         data_dir,
+        max_source_length,
+        max_target_length,
         type_path="train",
-        max_source_length=1024,
-        max_target_length=56,
         n_obs=None,
-        overwrite_cache=False,
-        prefix="",
         src_lang=None,
         tgt_lang=None,
+        prefix="",
     ):
         super().__init__()
-        # FIXME: the rstrip logic strips all the chars, it seems.
-        tok_name = tokenizer.__class__.__name__.lower().rstrip("tokenizer")
-        if hasattr(tokenizer, "set_lang") and src_lang is not None:
-            tokenizer.set_lang(src_lang)  # HACK: only applies to mbart
-        self.source = encode_file(
-            tokenizer,
-            os.path.join(data_dir, type_path + ".source"),
-            max_source_length,
-            overwrite_cache=overwrite_cache,
-            prefix=prefix,
-            tok_name=tok_name,
-        )
-        tgt_path = os.path.join(data_dir, type_path + ".target")
-        if hasattr(tokenizer, "set_lang"):
-            assert tgt_lang is not None, "--tgt_lang must be passed to build a translation"
-            tokenizer.set_lang(tgt_lang)  # HACK: only applies to mbart
-        self.target = encode_file(
-            tokenizer, tgt_path, max_target_length, overwrite_cache=overwrite_cache, tok_name=tok_name
-        )
+        self.src_file = Path(data_dir).joinpath(type_path + ".source")
+        self.tgt_file = Path(data_dir).joinpath(type_path + ".target")
+        self.src_lens = self.get_char_lens(self.src_file)
+        self.max_source_length = max_source_length
+        self.max_target_length = max_target_length
+        assert min(self.src_lens) > 0, f"found empty line in {self.src_file}"
+        self.tokenizer = tokenizer
+        self.prefix = prefix
         if n_obs is not None:
-            self.source = self.source[:n_obs]
-            self.target = self.target[:n_obs]
-        self.pad_token_id = tokenizer.pad_token_id
+            self.src_lens = self.src_lens[:n_obs]
+        self.pad_token_id = self.tokenizer.pad_token_id
+        self.src_lang = src_lang
+        self.tgt_lang = tgt_lang
 
     def __len__(self):
-        return len(self.source)
+        return len(self.src_lens)
+
+    def __getitem__(self, index) -> Dict[str, torch.Tensor]:
+        index = index + 1  # linecache starts at 1
+        source_line = self.prefix + linecache.getline(str(self.src_file), index).rstrip("\n")
+        tgt_line = linecache.getline(str(self.tgt_file), index).rstrip("\n")
+        assert source_line, f"empty source line for index {index}"
+        assert tgt_line, f"empty tgt line for index {index}"
+        source_inputs = encode_line(self.tokenizer, source_line, self.max_source_length)
+        target_inputs = encode_line(self.tokenizer, tgt_line, self.max_target_length)
+
+        source_ids = source_inputs["input_ids"].squeeze()
+        target_ids = target_inputs["input_ids"].squeeze()
+        src_mask = source_inputs["attention_mask"].squeeze()
+        return {
+            "input_ids": source_ids,
+            "attention_mask": src_mask,
+            "decoder_input_ids": target_ids,
+        }
 
-    def __getitem__(self, index):
-        source_ids = self.source[index]["input_ids"].squeeze()
-        target_ids = self.target[index]["input_ids"].squeeze()
-        src_mask = self.source[index]["attention_mask"].squeeze()
-        return {"input_ids": source_ids, "attention_mask": src_mask, "decoder_input_ids": target_ids}
+    @staticmethod
+    def get_char_lens(data_file):
+        return [len(x) for x in Path(data_file).open().readlines()]
 
     @staticmethod
-    def trim_seq2seq_batch(batch, pad_token_id):
+    def trim_seq2seq_batch(batch, pad_token_id) -> tuple:
         y = trim_batch(batch["decoder_input_ids"], pad_token_id)
         source_ids, source_mask = trim_batch(batch["input_ids"], pad_token_id, attention_mask=batch["attention_mask"])
         return source_ids, source_mask, y
 
-    def collate_fn(self, batch) -> dict:
+    def collate_fn(self, batch) -> Dict[str, torch.Tensor]:
         input_ids = torch.stack([x["input_ids"] for x in batch])
         masks = torch.stack([x["attention_mask"] for x in batch])
         target_ids = torch.stack([x["decoder_input_ids"] for x in batch])
         pad_token_id = self.pad_token_id
         y = trim_batch(target_ids, pad_token_id)
         source_ids, source_mask = trim_batch(input_ids, pad_token_id, attention_mask=masks)
-        batch = {"input_ids": source_ids, "attention_mask": source_mask, "decoder_input_ids": y}
+        batch = {
+            "input_ids": source_ids,
+            "attention_mask": source_mask,
+            "decoder_input_ids": y,
+        }
         return batch
 
     def make_sortish_sampler(self, batch_size):
-        lens = [x["input_ids"].ne(self.pad_token_id).sum() for x in self.source]
-        return SortishSampler(lens, batch_size)
+        return SortishSampler(self.src_lens, batch_size)
+
+
+class MBartDataset(Seq2SeqDataset):
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        if self.max_source_length != self.max_target_length:
+            warnings.warn(
+                f"Mbart will ignore max_target_length = {self.max_target_length} and use {self.max_source_length} for both sides."
+            )
+
+    def __getitem__(self, index) -> Dict[str, str]:
+        index = index + 1  # linecache starts at 1
+        source_line = self.prefix + linecache.getline(str(self.src_file), index).rstrip("\n")
+        tgt_line = linecache.getline(str(self.tgt_file), index).rstrip("\n")
+        assert source_line, f"empty source line for index {index}"
+        assert tgt_line, f"empty tgt line for index {index}"
+        return {
+            "tgt_texts": source_line,
+            "src_texts": tgt_line,
+        }
+
+    def collate_fn(self, batch) -> Dict[str, torch.Tensor]:
+        batch_encoding = self.tokenizer.prepare_translation_batch(
+            [x["src_texts"] for x in batch],
+            src_lang=self.src_lang,
+            tgt_texts=[x["tgt_texts"] for x in batch],
+            tgt_lang=self.tgt_lang,
+            max_length=self.max_source_length,
+        )
+        return batch_encoding.data
 
 
 class SortishSampler(Sampler):

src/transformers/tokenization_bart.py

@@ -118,12 +118,7 @@ class MBartTokenizer(XLMRobertaTokenizer):
         self.fairseq_tokens_to_ids.update(self.lang_code_to_id)
         self.fairseq_ids_to_tokens = {v: k for k, v in self.fairseq_tokens_to_ids.items()}
         self._additional_special_tokens = list(self.lang_code_to_id.keys())
-        self.reset_special_tokens()
-
-    def reset_special_tokens(self) -> None:
-        """Reset the special tokens to the source lang setting. No prefix and suffix=[eos, cur_lang_code]."""
-        self.prefix_tokens = []
-        self.suffix_tokens = [self.eos_token_id, self.cur_lang_code]
+        self.set_src_lang_special_tokens(kwargs.get("src_lang", "en_XX"))
 
     def build_inputs_with_special_tokens(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
@@ -183,12 +178,6 @@ class MBartTokenizer(XLMRobertaTokenizer):
             return prefix_ones + ([0] * len(token_ids_0)) + suffix_ones
         return prefix_ones + ([0] * len(token_ids_0)) + ([0] * len(token_ids_1)) + suffix_ones
 
-    def set_lang(self, lang: str) -> None:
-        """Set the current language code in order to call tokenizer properly."""
-        self.cur_lang_code = self.lang_code_to_id[lang]
-        self.prefix_tokens = [self.cur_lang_code]
-        self.suffix_tokens = [self.eos_token_id]
-
     def prepare_translation_batch(
         self,
         src_texts: List[str],
@@ -215,7 +204,7 @@ class MBartTokenizer(XLMRobertaTokenizer):
         """
         if max_length is None:
             max_length = self.max_len
-        self.cur_lang_code = self.lang_code_to_id[src_lang]
+        self.set_src_lang_special_tokens(src_lang)
         model_inputs: BatchEncoding = self(
             src_texts,
             add_special_tokens=True,
@@ -227,7 +216,7 @@ class MBartTokenizer(XLMRobertaTokenizer):
         )
         if tgt_texts is None:
             return model_inputs
-        self.set_lang(tgt_lang)
+        self.set_tgt_lang_special_tokens(tgt_lang)
         decoder_inputs: BatchEncoding = self(
             tgt_texts,
             add_special_tokens=True,
@@ -239,6 +228,18 @@ class MBartTokenizer(XLMRobertaTokenizer):
         )
         for k, v in decoder_inputs.items():
             model_inputs[f"decoder_{k}"] = v
-        self.cur_lang_code = self.lang_code_to_id[src_lang]
-        self.reset_special_tokens()  # sets to src_lang
+
+        self.set_src_lang_special_tokens(src_lang)  # sets to src_lang
         return model_inputs
+
+    def set_src_lang_special_tokens(self, src_lang) -> None:
+        """Reset the special tokens to the source lang setting. No prefix and suffix=[eos, cur_lang_code]."""
+        self.cur_lang_code = self.lang_code_to_id[src_lang]
+        self.prefix_tokens = []
+        self.suffix_tokens = [self.eos_token_id, self.cur_lang_code]
+
+    def set_tgt_lang_special_tokens(self, lang: str) -> None:
+        """Reset the special tokens to the target language setting. Prefix [tgt_lang_code], suffix =[eos]."""
+        self.cur_lang_code = self.lang_code_to_id[lang]
+        self.prefix_tokens = [self.cur_lang_code]
+        self.suffix_tokens = [self.eos_token_id]