Upload train_survival.py with huggingface_hub

train_survival.py

@@ -18,7 +18,6 @@ OUTPUT_MODEL_ID = "sunkencity/survival-expert-3b"
 dataset = load_dataset(DATASET_ID, split="train")
 
 # SANITIZE DATASET
-# Filter out any rows that have None or empty strings
 def filter_empty(example):
     return (
         example["instruction"] is not None 
@@ -27,9 +26,7 @@ def filter_empty(example):
         and len(example["response"]) > 0
     )
 
-print(f"Original dataset size: {len(dataset)}")
 dataset = dataset.filter(filter_empty)
-print(f"Filtered dataset size: {len(dataset)}")
 
 # Load Model
 bnb_config = BitsAndBytesConfig(
@@ -68,26 +65,21 @@ training_args = SFTConfig(
     push_to_hub=True,
     hub_model_id=OUTPUT_MODEL_ID,
     fp16=True,
-    dataset_text_field="text",
     packing=False,
     max_length=1024
+    # Removed dataset_text_field="text" as it conflicted with formatting_func
 )
 
 def formatting_prompts_func(example):
     output_texts = []
-    # Ensure we handle list input (batched)
     instructions = example['instruction']
     responses = example['response']
     
     for i in range(len(instructions)):
-        if i >= len(responses):
-            break # Should not happen after filtering, but safety first
-            
+        if i >= len(responses): break
         instruction = instructions[i]
         response = responses[i]
-        
-        if not instruction or not response:
-            continue
+        if not instruction or not response: continue
 
         text = f"<|im_start|>user\n{instruction}<|im_end|>\n<|im_start|>assistant\n{response}<|im_end|>"
         output_texts.append(text)
@@ -108,4 +100,4 @@ trainer.train()
 
 print("Pushing to hub...")
 trainer.push_to_hub()
-print("Done!")
+print("Done!")