File size: 4,184 Bytes
b653f91 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 |
# src/train.py
import os
import argparse
from datasets import Dataset
from transformers import (
AutoModelForSeq2SeqLM,
AutoTokenizer,
DataCollatorForSeq2Seq,
Seq2SeqTrainingArguments,
Seq2SeqTrainer,
)
def train_model():
"""
Fine-tunes a pre-trained NLLB model on a parallel dataset.
"""
parser = argparse.ArgumentParser(description="Fine-tune a translation model.")
parser.add_argument("--model_checkpoint", type=str, default="facebook/nllb-200-distilled-600M")
parser.add_argument("--source_lang", type=str, required=True, help="Source language code (e.g., 'ne')")
parser.add_argument("--target_lang", type=str, default="en")
parser.add_argument("--source_lang_tokenizer", type=str, required=True, help="Source language code for tokenizer (e.g., 'nep_Npan')")
parser.add_argument("--train_file_source", type=str, required=True, help="Path to the source language training file")
parser.add_argument("--train_file_target", type=str, required=True, help="Path to the target language training file")
parser.add_argument("--output_dir", type=str, required=True, help="Directory to save the fine-tuned model")
parser.add_argument("--epochs", type=int, default=3)
parser.add_argument("--batch_size", type=int, default=8)
args = parser.parse_args()
# --- 1. Configuration ---
MODEL_CHECKPOINT = args.model_checkpoint
SOURCE_LANG = args.source_lang
TARGET_LANG = args.target_lang
MODEL_OUTPUT_DIR = args.output_dir
# --- 2. Load Tokenizer and Model ---
print("Loading tokenizer and model...")
tokenizer = AutoTokenizer.from_pretrained(
MODEL_CHECKPOINT, src_lang=args.source_lang_tokenizer, tgt_lang="eng_Latn"
)
model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_CHECKPOINT)
# --- 3. Load and Preprocess Data (Memory-Efficiently) ---
print("Loading and preprocessing data...")
def generate_examples():
with open(args.train_file_source, "r", encoding="utf-8") as f_src, \
open(args.train_file_target, "r", encoding="utf-8") as f_tgt:
for src_line, tgt_line in zip(f_src, f_tgt):
yield {"translation": {SOURCE_LANG: src_line.strip(), TARGET_LANG: tgt_line.strip()}}
dataset = Dataset.from_generator(generate_examples)
split_datasets = dataset.train_test_split(train_size=0.95, seed=42)
split_datasets["validation"] = split_datasets.pop("test")
def preprocess_function(examples):
inputs = [ex[SOURCE_LANG] for ex in examples["translation"]]
targets = [ex[TARGET_LANG] for ex in examples["translation"]]
model_inputs = tokenizer(inputs, text_target=targets, max_length=128, truncation=True)
return model_inputs
tokenized_datasets = split_datasets.map(
preprocess_function,
batched=True,
remove_columns=split_datasets["train"].column_names,
)
# --- 4. Set Up Training Arguments ---
print("Setting up training arguments...")
training_args = Seq2SeqTrainingArguments(
output_dir=MODEL_OUTPUT_DIR,
eval_strategy="epoch",
learning_rate=2e-5,
per_device_train_batch_size=args.batch_size,
per_device_eval_batch_size=args.batch_size,
weight_decay=0.01,
save_total_limit=3,
num_train_epochs=args.epochs,
predict_with_generate=True,
fp16=False, # Set to True if you have a compatible GPU
)
# --- 5. Create the Trainer ---
data_collator = DataCollatorForSeq2Seq(tokenizer=tokenizer, model=model)
trainer = Seq2SeqTrainer(
model=model,
args=training_args,
train_dataset=tokenized_datasets["train"],
eval_dataset=tokenized_datasets["validation"],
tokenizer=tokenizer,
data_collator=data_collator,
)
# --- 6. Start Training ---
print("\n--- Starting model fine-tuning ---")
trainer.train()
print("--- Training complete ---")
# --- 7. Save the Final Model ---
print(f"Saving final model to {MODEL_OUTPUT_DIR}")
trainer.save_model()
print("Model saved successfully!")
if __name__ == "__main__":
train_model()
|