f

app.py

@@ -5,12 +5,12 @@ from transformers import AutoTokenizer, TrainingArguments, Trainer, AutoModelFor
 import torch
 import os
 
-# Force CPU
+# Force CPU-only execution
 os.environ["CUDA_VISIBLE_DEVICES"] = ""
 
 def train_model(file, hf_token):
     try:
-        # Basic data loading test
+        # Basic data loading
         df = pd.read_csv(file.name)
         print(f"Loaded CSV with {len(df)} rows")
         
@@ -20,19 +20,19 @@ def train_model(file, hf_token):
         model = AutoModelForCausalLM.from_pretrained(
             model_name,
             low_cpu_mem_usage=True,
-            torch_dtype=torch.float32
+            torch_dtype=torch.float32  # Ensure compatibility with CPU
         )
         
-        # Basic dataset creation
+        # Prepare dataset
         dataset = Dataset.from_pandas(df)
         
         args = TrainingArguments(
             output_dir="./results",
             per_device_train_batch_size=1,
             num_train_epochs=1,
-            no_cuda=True,
-            local_rank=-1,
-            use_cpu=True
+            no_cuda=True,  # Disable GPU
+            use_cpu=True,  # Ensure CPU usage
+            fp16=False     # Disable mixed precision
         )
         
         trainer = Trainer(
@@ -42,11 +42,12 @@ def train_model(file, hf_token):
             tokenizer=tokenizer
         )
         
-        return f"Setup successful! Loaded {len(df)} rows"
+        return f"Setup successful! Loaded {len(df)} rows for training."
         
     except Exception as e:
         return f"Error: {str(e)}\nType: {type(e)}"
 
+# Gradio interface
 demo = gr.Interface(
     fn=train_model,
     inputs=[
@@ -58,4 +59,4 @@ demo = gr.Interface(
 )
 
 if __name__ == "__main__":
-    demo.launch(debug=True)
+    demo.launch(debug=True, ssr=False)  # Disable SSR for compatibility

requirements.txt

@@ -1,4 +1,4 @@
-gradio==4.19.2
+gradio==5.12.0
 transformers==4.37.2
 torch==2.1.2
 datasets==2.16.1