Instructions to use radce/Llama-3.2-1B-ru-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use radce/Llama-3.2-1B-ru-v2 with Transformers:

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

pipe = pipeline("text-generation", model="radce/Llama-3.2-1B-ru-v2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("radce/Llama-3.2-1B-ru-v2")
model = AutoModelForCausalLM.from_pretrained("radce/Llama-3.2-1B-ru-v2")
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 radce/Llama-3.2-1B-ru-v2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "radce/Llama-3.2-1B-ru-v2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "radce/Llama-3.2-1B-ru-v2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/radce/Llama-3.2-1B-ru-v2

SGLang

How to use radce/Llama-3.2-1B-ru-v2 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 "radce/Llama-3.2-1B-ru-v2" \
    --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": "radce/Llama-3.2-1B-ru-v2",
		"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 "radce/Llama-3.2-1B-ru-v2" \
        --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": "radce/Llama-3.2-1B-ru-v2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use radce/Llama-3.2-1B-ru-v2 with Docker Model Runner:
```
docker model run hf.co/radce/Llama-3.2-1B-ru-v2
```

Дообучались слои 14, 15, 16.

Датасет состоял из 13 862 816 токенов.

Видеокарта для дообучения: Tesla A100.

Обучение

{'loss': 0.8521, 'grad_norm': 0.5644629001617432, 'learning_rate': 2.9148375768217733e-05, 'epoch': 1.29}
{'loss': 0.6742, 'grad_norm': 0.5370610952377319, 'learning_rate': 7.199297629499562e-06, 'epoch': 2.58}
{'train_runtime': 5708.2175, 'train_samples_per_second': 22.869, 'train_steps_per_second': 0.204, 'train_loss': 0.7442483934749853, 'epoch': 3.0}

Downloads last month: 2

Safetensors

Model size

1B params

Tensor type

F32

Model tree for radce/Llama-3.2-1B-ru-v2

Base model

radce/llama3.2-1B-Instruct-ru

Finetuned

(1)

this model

radce
/

Llama-3.2-1B-ru-v2

Model tree for radce/Llama-3.2-1B-ru-v2

Datasets used to train radce/Llama-3.2-1B-ru-v2