Instructions to use vovaRL/NanoColibri-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use vovaRL/NanoColibri-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="vovaRL/NanoColibri-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("vovaRL/NanoColibri-Instruct")
model = AutoModelForCausalLM.from_pretrained("vovaRL/NanoColibri-Instruct", device_map="auto")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use vovaRL/NanoColibri-Instruct with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/vovaRL/NanoColibri-Instruct

SGLang

How to use vovaRL/NanoColibri-Instruct 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 "vovaRL/NanoColibri-Instruct" \
    --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": "vovaRL/NanoColibri-Instruct",
		"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 "vovaRL/NanoColibri-Instruct" \
        --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": "vovaRL/NanoColibri-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use vovaRL/NanoColibri-Instruct with Docker Model Runner:
```
docker model run hf.co/vovaRL/NanoColibri-Instruct
```

NanoColibri-Instruct

File size: 1,891 Bytes

b3ac35d

{
  "step": 21500,
  "global_tokens": 5439488000,
  "contributor": "vovaRL",
  "cloud": "home",
  "base_model": "scratch:colibri-nano",
  "wall_clock_s": 11548.808067560196,
  "history_tail": [
    {
      "step": 21460,
      "t": 1785260063.755,
      "ntp": 1.8235763311386108,
      "zloss": 0.3778407573699951,
      "lr": 3.0002803191548873e-05,
      "grad_norm": 0.322265625,
      "ppl": 6.193970576015481,
      "balance_max_imbalance": 0.0843505859375,
      "balance_dead_frac": 0.015625,
      "tokens": 5434245120
    },
    {
      "step": 21470,
      "t": 1785260140.412,
      "ntp": 3.5947928428649902,
      "zloss": 0.40486156940460205,
      "lr": 3.000157679763363e-05,
      "grad_norm": 0.37109375,
      "ppl": 36.40815699711934,
      "balance_max_imbalance": 0.229736328125,
      "balance_dead_frac": 0.109375,
      "tokens": 5435555840
    },
    {
      "step": 21480,
      "t": 1785260217.057,
      "ntp": 1.5007089376449585,
      "zloss": 0.40441009402275085,
      "lr": 3.0000700799706186e-05,
      "grad_norm": 0.33984375,
      "ppl": 4.484867434931063,
      "balance_max_imbalance": 0.0440673828125,
      "balance_dead_frac": 0.0,
      "tokens": 5436866560
    },
    {
      "step": 21490,
      "t": 1785260293.595,
      "ntp": 1.7874301671981812,
      "zloss": 0.3486020863056183,
      "lr": 3.000017520004023e-05,
      "grad_norm": 0.328125,
      "ppl": 5.974080335333217,
      "balance_max_imbalance": 0.2159423828125,
      "balance_dead_frac": 0.015625,
      "tokens": 5438177280
    },
    {
      "step": 21500,
      "t": 1785260370.219,
      "ntp": 1.4909950494766235,
      "zloss": 0.3707721531391144,
      "lr": 2.9999999999999997e-05,
      "grad_norm": 0.333984375,
      "ppl": 4.441512845942532,
      "balance_max_imbalance": 0.076904296875,
      "balance_dead_frac": 0.015625,
      "tokens": 5439488000
    }
  ]
}