Update app.py

app.py

@@ -333,43 +333,32 @@ def train_model_inline(uploaded_file, text_column, label_column, num_epochs, bat
         TRAINING_LOGS.append(f"- Warmup steps: {warmup_steps}")
         yield "\n".join(TRAINING_LOGS)
         
-        # Create a dictionary for training arguments to handle version compatibility
-        training_args_dict = {
-            "output_dir": str(output_dir),
-            "num_train_epochs": num_epochs,
-            "per_device_train_batch_size": batch_size,
-            "per_device_eval_batch_size": batch_size,
-            "warmup_steps": warmup_steps,
-            "weight_decay": 0.01,
-            "learning_rate": learning_rate,
-            "logging_dir": str(output_dir / "logs"),
-            "logging_steps": logging_steps,
-            "eval_steps": eval_steps,
-            "save_steps": save_steps,
-            "save_total_limit": 2,
-            "load_best_model_at_end": True,
-            "metric_for_best_model": "eval_accuracy",
-            "greater_is_better": True,
-            "push_to_hub": push_to_hub,
-            "hub_model_id": hub_model_id if push_to_hub else None,
-            "report_to": None,
-            "dataloader_num_workers": 0,
-            "fp16": torch.cuda.is_available(),
-            "seed": 42,
-            "remove_unused_columns": False,
-        }
+        training_args = TrainingArguments(
+            output_dir=str(output_dir),
+            num_train_epochs=num_epochs,
+            per_device_train_batch_size=batch_size,
+            per_device_eval_batch_size=batch_size,
+            warmup_steps=warmup_steps,
+            weight_decay=0.01,
+            learning_rate=learning_rate,
+            logging_dir=str(output_dir / "logs"),
+            logging_steps=logging_steps,
+            evaluation_strategy="steps", # Corrected parameter name
+            eval_steps=eval_steps,
+            save_steps=save_steps,
+            save_total_limit=2,
+            load_best_model_at_end=True,
+            metric_for_best_model="eval_accuracy",
+            greater_is_better=True,
+            push_to_hub=push_to_hub,
+            hub_model_id=hub_model_id if push_to_hub else None,
+            report_to=None,
+            dataloader_num_workers=0,
+            fp16=torch.cuda.is_available(),
+            seed=42,
+            remove_unused_columns=False,
+        )
 
-        # Try to use 'eval_strategy' and fall back to 'evaluation_strategy' if a TypeError occurs
-        try:
-            training_args_dict["eval_strategy"] = "steps"
-            training_args = TrainingArguments(**training_args_dict)
-        except TypeError as e:
-            if "unexpected keyword argument 'eval_strategy'" in str(e):
-                training_args_dict["evaluation_strategy"] = "steps"
-                training_args = TrainingArguments(**training_args_dict)
-            else:
-                raise e
-        
         # Data collator
         data_collator = DataCollatorWithPadding(tokenizer=tokenizer)
         
@@ -635,8 +624,8 @@ def push_to_hub_after_training(model_path, username, model_name, token):
 
 def count_tokens(text):
     """Count tokens in input text"""
-    global CURRENT_TOKENIZER # Ensure we can modify the global tokenizer
-    if not text:
+    global CURRENT_TOKENIZER
+    if text is None:
         return "Enter text to see token count"
     
     # Attempt to load a default tokenizer if it's not set