Upload finetune_soc.py with huggingface_hub

finetune_soc.py

@@ -29,12 +29,12 @@ MODEL_ID = "swiss-ai/Apertus-8B-2509"
 DATASET_ID = "marcodsn/SOC-2508"
 OUTPUT_REPO = "Colby/apertus-8b-soc"
 
-print("📦 Loading tokenizer...")
+print("Loading tokenizer...")
 tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
 
-print("📦 Loading dataset...")
+print("Loading dataset...")
 dataset = load_dataset(DATASET_ID, split="train")
-print(f"✅ Loaded {len(dataset)} conversations")
+print(f"Loaded {len(dataset)} conversations")
 
 
 def format_conversation(example):
@@ -77,12 +77,12 @@ def format_conversation(example):
     return {"text": text}
 
 
-print("🔀 Formatting conversations to ChatML...")
+print("Formatting conversations to ChatML...")
 dataset = dataset.map(format_conversation, remove_columns=dataset.column_names)
 split = dataset.train_test_split(test_size=0.05, seed=42)
 train_dataset = split["train"]
 eval_dataset = split["test"]
-print(f"   Train: {len(train_dataset)}  Eval: {len(eval_dataset)}")
+print(f"  Train: {len(train_dataset)}  Eval: {len(eval_dataset)}")
 
 peft_config = LoraConfig(
     r=16,
@@ -128,7 +128,7 @@ config = SFTConfig(
     run_name="apertus-8b-soc-v1",
 )
 
-print("🎯 Initializing trainer...")
+print("Initializing trainer...")
 trainer = SFTTrainer(
     model=MODEL_ID,
     train_dataset=train_dataset,
@@ -137,12 +137,12 @@ trainer = SFTTrainer(
     args=config,
 )
 
-print("🚀 Starting training...")
+print("Starting training...")
 trainer.train()
 
-print("💾 Pushing to Hub...")
+print("Pushing to Hub...")
 trainer.push_to_hub()
 trackio.finish()
 
-print(f"✅ Done! Model at: https://huggingface.co/{OUTPUT_REPO}")
-print(f"📊 Metrics at: https://huggingface.co/spaces/Colby/trackio")
+print(f"Done! Model at: https://huggingface.co/{OUTPUT_REPO}")
+print(f"Metrics at: https://huggingface.co/spaces/Colby/trackio")