Instructions to use ecastera/eva-mistral-dolphin-7b-spanish with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ecastera/eva-mistral-dolphin-7b-spanish with Transformers:

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

pipe = pipeline("text-generation", model="ecastera/eva-mistral-dolphin-7b-spanish")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ecastera/eva-mistral-dolphin-7b-spanish")
model = AutoModelForCausalLM.from_pretrained("ecastera/eva-mistral-dolphin-7b-spanish")
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 ecastera/eva-mistral-dolphin-7b-spanish with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ecastera/eva-mistral-dolphin-7b-spanish"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ecastera/eva-mistral-dolphin-7b-spanish",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ecastera/eva-mistral-dolphin-7b-spanish

SGLang

How to use ecastera/eva-mistral-dolphin-7b-spanish 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 "ecastera/eva-mistral-dolphin-7b-spanish" \
    --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": "ecastera/eva-mistral-dolphin-7b-spanish",
		"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 "ecastera/eva-mistral-dolphin-7b-spanish" \
        --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": "ecastera/eva-mistral-dolphin-7b-spanish",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ecastera/eva-mistral-dolphin-7b-spanish with Docker Model Runner:
```
docker model run hf.co/ecastera/eva-mistral-dolphin-7b-spanish
```

¿Será posible proporcionar los modelos en formato GGUF?

by polysound - opened Jan 27, 2024

Discussion

polysound

Jan 27, 2024

Para usuarios que ejecutan modelos de lenguaje en software como llama.cpp (y otras que usan la librería como kobold.cpp), sería muy útil añadir los modelos en formato GGUF, opcionalmente proporcionar versiones cuantizadas a 4, 5, 6 y 8-bits (Qx_K_M) mediante quantize.exe de llama.cpp.
Estoy muy interesado en personas que hacen fine-tunes de modelos de lenguaje para tener una mejor coherencia en el español.
¡Gracias!

ecastera

Owner Jan 27, 2024

Si, estoy de acuerdo. Dejame que lo mire, no parece complicado conforme a las instrucciones:
https://github.com/ggerganov/llama.cpp/discussions/2948
Estoy trabajando en otro modelo en español mejor basado en udkai/Turdus. Es una pena que tengamos tan pocos modelos LLMs buenos en español.

ecastera changed discussion status to closed Jan 27, 2024

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment