try to resolve the issue with sftt trainer or trackio

trainer.py

@@ -98,40 +98,44 @@ class SmolLM3Trainer:
         callbacks.append(SimpleConsoleCallback())
         logger.info("Added simple console monitoring callback")
         
-        # Try to add Trackio callback if available
-        if self.monitor and self.monitor.enable_tracking:
-            try:
-                trackio_callback = self.monitor.create_monitoring_callback()
-                if trackio_callback:
-                    callbacks.append(trackio_callback)
-                    logger.info("Added Trackio monitoring callback")
-                else:
-                    logger.warning("Failed to create Trackio callback")
-            except Exception as e:
-                logger.error(f"Error creating Trackio callback: {e}")
-                logger.info("Continuing with console monitoring only")
+        # Try to add Trackio callback if available (temporarily disabled for debugging)
+        logger.info("Skipping Trackio callback to debug training issue")
+        # if self.monitor and self.monitor.enable_tracking:
+        #     try:
+        #         trackio_callback = self.monitor.create_monitoring_callback()
+        #         if trackio_callback:
+        #         callbacks.append(trackio_callback)
+        #         logger.info("Added Trackio monitoring callback")
+        #     else:
+        #         logger.warning("Failed to create Trackio callback")
+        # except Exception as e:
+        #     logger.error(f"Error creating Trackio callback: {e}")
+        #     logger.info("Continuing with console monitoring only")
         
-        if self.use_sft_trainer:
-            # Use SFTTrainer for supervised fine-tuning
-            trainer = SFTTrainer(
+        # Try standard Trainer first (more stable with callbacks)
+        try:
+            trainer = Trainer(
                 model=self.model.model,
+                tokenizer=self.model.tokenizer,
+                args=training_args,
                 train_dataset=train_dataset,
                 eval_dataset=eval_dataset,
-                args=training_args,
                 data_collator=data_collator,
                 callbacks=callbacks,
             )
-        else:
-            # Use standard Trainer
-            trainer = Trainer(
+            logger.info("Using standard Hugging Face Trainer")
+        except Exception as e:
+            logger.warning(f"Standard Trainer failed: {e}")
+            # Fallback to SFTTrainer
+            trainer = SFTTrainer(
                 model=self.model.model,
-                tokenizer=self.model.tokenizer,
-                args=training_args,
                 train_dataset=train_dataset,
                 eval_dataset=eval_dataset,
+                args=training_args,
                 data_collator=data_collator,
                 callbacks=callbacks,
             )
+            logger.info("Using SFTTrainer")
         
         return trainer