Upload train.py with huggingface_hub

train.py

@@ -76,12 +76,12 @@ def main():
         train_dataset=dataset,
         peft_config=peft_config,
         formatting_func=formatting_prompts_func,
-        max_seq_length=2048,
         tokenizer=tokenizer,
-        args=TrainingArguments(
+        args=SFTConfig(
             output_dir=OUTPUT_DIR,
+            max_seq_length=2048, # Moved here
             per_device_train_batch_size=2,
-            gradient_accumulation_steps=4, # Simulate larger batch size
+            gradient_accumulation_steps=4,
             learning_rate=2e-4,
             logging_steps=10,
             num_train_epochs=1,
@@ -89,9 +89,10 @@ def main():
             fp16=True,
             group_by_length=True,
             save_strategy="epoch",
-            report_to="none", # Change to "wandb" if desired
+            report_to="none",
             push_to_hub=True,
-            hub_model_id=f"ceperaltab/{OUTPUT_DIR}", # Pushes to your namespace
+            hub_model_id=f"ceperaltab/{OUTPUT_DIR}",
+            dataset_text_field="text", # SFTConfig requires this or packing, though we use formatting_func
         ),
     )