Instructions to use FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep")
model = AutoModelForCausalLM.from_pretrained("FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep

SGLang

How to use FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep with Docker Model Runner:
```
docker model run hf.co/FlyPig23/Llama3.2-3B_Paper_Impact_code_SFT_1ep
```

Llama3.2-3B_Paper_Impact_code_SFT_1ep

File size: 1,883 Bytes

a3b1b88

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 29,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1724137931034483,
      "grad_norm": 0.8689085245132446,
      "learning_rate": 1.992708874098054e-05,
      "loss": 0.143,
      "step": 5
    },
    {
      "epoch": 0.3448275862068966,
      "grad_norm": 0.11412197351455688,
      "learning_rate": 1.7485107481711014e-05,
      "loss": 0.194,
      "step": 10
    },
    {
      "epoch": 0.5172413793103449,
      "grad_norm": 0.07396583259105682,
      "learning_rate": 1.2393156642875579e-05,
      "loss": 0.0887,
      "step": 15
    },
    {
      "epoch": 0.6896551724137931,
      "grad_norm": 0.029150845482945442,
      "learning_rate": 6.453951129574644e-06,
      "loss": 0.0875,
      "step": 20
    },
    {
      "epoch": 0.8620689655172413,
      "grad_norm": 0.13400128483772278,
      "learning_rate": 1.7701613410634367e-06,
      "loss": 0.0887,
      "step": 25
    },
    {
      "epoch": 1.0,
      "step": 29,
      "total_flos": 1.5530368955108557e+17,
      "train_loss": 0.11576244132272129,
      "train_runtime": 811.0761,
      "train_samples_per_second": 4.543,
      "train_steps_per_second": 0.036
    }
  ],
  "logging_steps": 5,
  "max_steps": 29,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.5530368955108557e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}