init

training_scripts/finetune_t5.py

@@ -140,7 +140,11 @@ def train(
         batch = [64] if batch is None else batch
         for n, (lr_tmp, batch_tmp) in enumerate(product(lr, batch)):
             logging.info(f"[TRAIN {n}/{len(lr) * len(batch)}] lr: {lr_tmp}, batch: {batch_tmp}")
+            model = load_model(
+                model_name=model_name, use_auth_token=use_auth_token, low_cpu_mem_usage=model_low_cpu_mem_usage
+            )
             trainer = Seq2SeqTrainer(
+                model=model,
                 args=Seq2SeqTrainingArguments(
                     num_train_epochs=epoch,
                     learning_rate=lr_tmp,
@@ -151,10 +155,7 @@ def train(
                     seed=random_seed,
                     per_device_train_batch_size=batch_tmp,
                 ),
-                data_collator=transformers.DataCollatorForSeq2Seq(tokenizer, model=load_model(
-                    model_name=model_name,
-                    use_auth_token=use_auth_token,
-                    low_cpu_mem_usage=model_low_cpu_mem_usage)),
+                data_collator=transformers.DataCollatorForSeq2Seq(tokenizer, model=model),
                 train_dataset=tokenized_dataset['train_ds'],
                 eval_dataset=tokenized_dataset['validation_ds'],
                 compute_metrics=compute_metric,