Upload app.py with huggingface_hub

app.py

@@ -37,34 +37,46 @@ model_config = config.get("model_config", {})
 # Model details from config
 MODEL_NAME = model_config.get("model_name_or_path", "unsloth/DeepSeek-R1-Distill-Qwen-14B-bnb-4bit")
 SPACE_NAME = os.getenv("HF_SPACE_NAME", "phi4training")
-TRAINING_ACTIVE = os.path.exists("TRAINING_ACTIVE")
 
 # Function to start the training process
 def start_training():
     try:
-        # Create TRAINING_ACTIVE file
-        with open("TRAINING_ACTIVE", "w") as f:
-            f.write("Training in progress")
-            
-        # Run the training script in the background
-        subprocess.Popen(["python", "run_cloud_training.py"], 
-                        stdout=subprocess.PIPE, 
-                        stderr=subprocess.PIPE)
+        # Run the training script using HF's native logging
+        process = subprocess.Popen(
+            ["python", "run_cloud_training.py"], 
+            stdout=subprocess.PIPE, 
+            stderr=subprocess.PIPE,
+            universal_newlines=True
+        )
         
-        return "✅ Training started! Check status below for updates."
+        # Log the start of training
+        logger.info("Training started - Check Hugging Face logs for details")
+        
+        return """
+        ✅ Training process initiated! 
+        
+        The model is now being fine-tuned in the background.
+        
+        To monitor progress:
+        1. Check the Hugging Face space logs in the "Logs" tab
+        2. Training metrics will be available in the Hugging Face UI
+        3. The process will continue running in the background
+        
+        NOTE: This is a research training phase only, no model outputs will be available.
+        """
     except Exception as e:
         logger.error(f"Error starting training: {str(e)}")
         return f"❌ Error starting training: {str(e)}"
 
 # Create Gradio interface - training status only, no model outputs
 with gr.Blocks(css="footer {visibility: hidden}") as demo:
-    gr.Markdown(f"# {SPACE_NAME}: Training Status Dashboard")
+    gr.Markdown(f"# {SPACE_NAME}: Research Training Dashboard")
     
     with gr.Row():
         with gr.Column():
             status = gr.Markdown(
                 f"""
-                ## Research Training Phase Active
+                ## DeepSeek-R1-Distill-Qwen-14B Research Training
                 
                 **Model**: {MODEL_NAME}
                 **Dataset**: phi4-cognitive-dataset
@@ -78,62 +90,39 @@ with gr.Blocks(css="footer {visibility: hidden}") as demo:
                 - **Learning Rate**: {config.get("training_config", {}).get("learning_rate", 2e-5)}
                 - **Max Sequence Length**: {config.get("training_config", {}).get("max_seq_length", 2048)}
                 
-                ### Training Status:
-                {"🟢 Training in progress" if TRAINING_ACTIVE else "⚪ Training not currently active"}
-                
                 ⚠️ **NOTE**: This space does not provide model outputs during the research training phase.
+                All logs are available in the Hugging Face "Logs" tab.
                 """
             )
     
     with gr.Row():
-        # Add buttons for starting training and refreshing status
+        # Add button for starting training
         start_btn = gr.Button("Start Training", variant="primary")
-        refresh_btn = gr.Button("Refresh Status")
         
     # Output area for training start messages
     training_output = gr.Markdown("")
     
-    def refresh_status():
-        # Re-check if training is active
-        training_active = os.path.exists("TRAINING_ACTIVE")
-        return f"""
-        ## Research Training Phase Active
-        
-        **Model**: {MODEL_NAME}
-        **Dataset**: phi4-cognitive-dataset
-        
-        This is a multidisciplinary research training phase. The model is not available for interactive use.
-        
-        ### Training Configuration:
-        - **Epochs**: {config.get("training_config", {}).get("num_train_epochs", 3)}
-        - **Batch Size**: {config.get("training_config", {}).get("per_device_train_batch_size", 2)}
-        - **Gradient Accumulation Steps**: {config.get("training_config", {}).get("gradient_accumulation_steps", 4)}
-        - **Learning Rate**: {config.get("training_config", {}).get("learning_rate", 2e-5)}
-        - **Max Sequence Length**: {config.get("training_config", {}).get("max_seq_length", 2048)}
-        
-        ### Training Status:
-        {"🟢 Training in progress" if training_active else "⚪ Training not currently active"}
-        
-        ⚠️ **NOTE**: This space does not provide model outputs during the research training phase.
-        """
-    
-    # Connect button clicks to functions
+    # Connect start button to function
     start_btn.click(start_training, outputs=training_output)
-    refresh_btn.click(refresh_status, outputs=status)
     
     gr.Markdown("""
     ### Research Training Information
+    
     This model is being fine-tuned on research-focused datasets and is not available for interactive querying.
-    Training logs are available to authorized researchers only.
+    The training process will run in the background and logs will be available in the Hugging Face UI.
     
-    ### Instructions
+    #### Instructions
     1. Click "Start Training" to begin the fine-tuning process
-    2. Use "Refresh Status" to check training progress
-    3. Training logs are saved to the output directory
+    2. Monitor progress in the Hugging Face "Logs" tab
+    3. Training metrics and results will be saved to the output directory
+    
+    #### About This Project
+    The model is being fine-tuned on the phi4-cognitive-dataset with a focus on research capabilities.
+    This training phase does not include any interactive features or output generation.
     """)
 
 # Launch the interface
 if __name__ == "__main__":
     # Start Gradio with minimal features
-    logger.info("Starting training status dashboard")
-    demo.launch(share=False)  # Removed enable_queue parameter which is no longer supported in Gradio 5.x 
+    logger.info("Starting research training dashboard")
+    demo.launch(share=False)