| import os |
| import subprocess |
| import sys |
|
|
| def run_cmd(cmd, cwd=None): |
| print(f"$ {cmd}") |
| result = subprocess.run(cmd, shell=True, cwd=cwd, capture_output=True, text=True) |
| if result.stdout: |
| print(result.stdout) |
| if result.stderr: |
| print(result.stderr, file=sys.stderr) |
| if result.returncode != 0: |
| raise RuntimeError(f"Command failed: {cmd}") |
| return result |
|
|
| def main(): |
| os.makedirs("/content/project_bob", exist_ok=True) |
| os.chdir("/content/project_bob") |
|
|
| run_cmd("pip install -q -U autotrain-advanced huggingface_hub huggingface_hub[cli] optimum transformers accelerate peft bitsandbytes datasets") |
|
|
| with open("dataset.jsonl", "w") as f: |
| f.write("""{"text": "<|begin_of_text|><|start_header_id|>user<|end_header_id|>What is the current price of Bitcoin?<|eot_id|><|start_header_id|>assistant<|end_header_id|><call_tool>get_crypto_price({\\"symbol\\": \\"BTC\\"})</call_tool><|eot_id|>"} |
| {"text": "<|begin_of_text|><|start_header_id|>user<|end_header_id|>Multiply 42 by 58 for me.<|eot_id|><|start_header_id|>assistant<|end_header_id|><call_tool>multiply_numbers({\\"a\\": 42, \\"b\\": 58})</call_tool><|eot_id|>"} |
| {"text": "<|begin_of_text|><|start_header_id|>user<|end_header_id|>What is the capital of France?<|eot_id|><|start_header_id|>assistant<|end_header_id|>The capital of France is Paris.<|eot_id|>"} |
| {"text": "<|begin_of_text|><|start_header_id|>user<|end_header_id|>Calculate 144 multiplied by 37.<|eot_id|><|start_header_id|>assistant<|end_header_id|><call_tool>multiply_numbers({\\"a\\": 144, \\"b\\": 37})</call_tool><|eot_id|>"} |
| {"text": "<|begin_of_text|><|start_header_id|>user<|end_header_id|>What's the current Ethereum price?<|eot_id|><|start_header_id|>assistant<|end_header_id|><call_tool>get_crypto_price({\\"symbol\\": \\"ETH\\"})</call_tool><|eot_id|>"} |
| """) |
|
|
| print("Uploading dataset.jsonl to Hugging Face Hub...") |
| from huggingface_hub import HfApi, HfFolder |
| api = HfApi() |
| api.upload_file( |
| path_or_fileobj="dataset.jsonl", |
| path_in_repo="dataset.jsonl", |
| repo_id="yummyfiles/Bob", |
| repo_type="dataset", |
| token=os.getenv("HF_TOKEN"), |
| ) |
| print("Dataset uploaded to yummyfiles/Bob") |
|
|
| print("Starting AutoTrain QLoRA fine-tuning...") |
| autotrain_cmd = ( |
| "autotrain llm --train " |
| "--project-name Bob " |
| "--model meta-llama/Llama-3.2-1B-Instruct " |
| "--train-data yummyfiles/Bob " |
| "--data-path dataset.jsonl " |
| "--text-column text " |
| "--trainer peft " |
| "--peft-r 16 " |
| "--peft-alpha 32 " |
| "--peft-dropout 0.05 " |
| "--learning-rate 2e-4 " |
| "--batch-size 1 " |
| "--epochs 3 " |
| "--block-size 2048 " |
| "--warmup-ratio 0.03 " |
| "--optimizer paged_adamw_8bit " |
| "--scheduler cosine " |
| "--gradient-accumulation 4 " |
| "--mixed-precision bf16 " |
| "--quantization 4bit " |
| "--push-to-hub " |
| "--repo-id yummyfiles/Bob " |
| f"--token {os.getenv('HF_TOKEN')}" |
| ) |
| run_cmd(autotrain_cmd) |
|
|
| print("Converting model to ONNX (4-bit)...") |
| run_cmd("pip install -q optimum[onnxruntime]") |
| run_cmd( |
| "optimum-cli export onnx " |
| "--model yummyfiles/Bob " |
| "--task text-generation " |
| "--quantize int4 " |
| "--output /content/project_bob/web_model" |
| ) |
|
|
| print("Uploading ONNX model to yummyfiles/Bob...") |
| api.upload_folder( |
| folder_path="/content/project_bob/web_model", |
| path_in_repo="web_model", |
| repo_id="yummyfiles/Bob", |
| repo_type="model", |
| token=os.getenv("HF_TOKEN"), |
| ) |
|
|
| print("Done! Model and ONNX artifacts uploaded to yummyfiles/Bob") |
|
|
| if __name__ == "__main__": |
| main() |
