Instructions to use Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact with Transformers:

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

pipe = pipeline("text-generation", model="Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact")
model = AutoModelForCausalLM.from_pretrained("Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact")
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

vLLM

How to use Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact

SGLang

How to use Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact 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 "Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact" \
    --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": "Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact",
		"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 "Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact" \
        --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": "Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact with Docker Model Runner:
```
docker model run hf.co/Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

tiny-llama-ultra-compact

Este repositorio fue generado automáticamente por el demo ultra compacto.

Resumen del entrenamiento

Dataset utilizado: HuggingFaceFW/finewiki (config: sr)
Las métricas detalladas se omiten para mantener el entrenamiento acumulativo sin exponer valores crudos.
Huella agregada de métricas: 25942bef49702bbc1bd6b3f0549e7ebc93c44505266f2a5f5bc33dbaa67726cd (campos rastreados: 27)

Muestra generada

La muestra completa se omite para ahorrar espacio. Longitud: 300 caracteres. Huella SHA256: f7291e1c4f8bd374e29cd6307789fe437ce3dae7a67a40a9cfffcfa76bb1e0be.

Reproducción

Ejecuta python Xd.py con las dependencias indicadas para volver a entrenar y subir los artefactos automáticamente.

Downloads last month: 103

Safetensors

Model size

289k params

Tensor type

F32

Model tree for Ignaciohhhhggfgjfrffd/tiny-llama-ultra-compact

Finetunes

3 models

Quantizations

1 model