Upload run_cloud_training.py with huggingface_hub

run_cloud_training.py

@@ -24,12 +24,54 @@ from unsloth import FastLanguageModel
 # Disable flash attention globally
 os.environ["TRANSFORMERS_NO_FLASH_ATTENTION"] = "1"
 
-# Try to install flash-attention (for systems that support it)
+# Configure logging first
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+    handlers=[
+        logging.StreamHandler(),
+        logging.FileHandler("training.log")
+    ]
+)
+logger = logging.getLogger(__name__)
+
+# Make sure torch is installed and available before proceeding
+try:
+    logger.info("Importing torch...")
+    import torch
+    logger.info(f"PyTorch version: {torch.__version__}")
+    logger.info(f"CUDA available: {torch.cuda.is_available()}")
+    if torch.cuda.is_available():
+        logger.info(f"CUDA version: {torch.version.cuda}")
+        logger.info(f"GPU: {torch.cuda.get_device_name(0)}")
+except ImportError:
+    logger.error("PyTorch not found. Installing torch first...")
+    try:
+        import subprocess
+        import sys
+        subprocess.check_call([sys.executable, "-m", "pip", "install", "torch"])
+        logger.info("PyTorch installed successfully. Importing...")
+        import torch
+        logger.info(f"PyTorch version: {torch.__version__}")
+    except Exception as e:
+        logger.error(f"Failed to install PyTorch: {e}")
+        logger.error("Cannot proceed without PyTorch. Exiting.")
+        raise
+
+# Now try to install flash-attention (for systems that support it)
 try:
     import subprocess
     import sys
     
-    logger = logging.getLogger(__name__)
+    # Make sure torch is installed before attempting flash-attn
+    try:
+        logger.info("Ensuring PyTorch is installed before flash-attention...")
+        subprocess.check_call([sys.executable, "-m", "pip", "install", "torch", "--quiet"])
+        logger.info("PyTorch installation verified")
+    except Exception as torch_error:
+        logger.warning(f"PyTorch installation check failed: {torch_error}")
+        logger.info("Will continue with flash-attention installation anyway")
+    
     logger.info("Attempting to install flash-attention...")
     
     # Try multiple installation approaches for flash-attention
@@ -74,17 +116,6 @@ except ImportError:
     TENSORBOARD_AVAILABLE = False
     print("Tensorboard not available. Will skip tensorboard logging.")
 
-# Configure logging
-logging.basicConfig(
-    level=logging.INFO,
-    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
-    handlers=[
-        logging.StreamHandler(),
-        logging.FileHandler("training.log")
-    ]
-)
-logger = logging.getLogger(__name__)
-
 # Default dataset path - use the correct path with username
 DEFAULT_DATASET = "George-API/phi4-cognitive-dataset"