Instructions to use clibrain/Llama-2-ft-instruct-es with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use clibrain/Llama-2-ft-instruct-es with Transformers:

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

pipe = pipeline("text-generation", model="clibrain/Llama-2-ft-instruct-es")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("clibrain/Llama-2-ft-instruct-es")
model = AutoModelForCausalLM.from_pretrained("clibrain/Llama-2-ft-instruct-es", device_map="auto")

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use clibrain/Llama-2-ft-instruct-es with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "clibrain/Llama-2-ft-instruct-es"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "clibrain/Llama-2-ft-instruct-es",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/clibrain/Llama-2-ft-instruct-es

SGLang

How to use clibrain/Llama-2-ft-instruct-es 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 "clibrain/Llama-2-ft-instruct-es" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "clibrain/Llama-2-ft-instruct-es",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "clibrain/Llama-2-ft-instruct-es" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "clibrain/Llama-2-ft-instruct-es",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use clibrain/Llama-2-ft-instruct-es with Docker Model Runner:
```
docker model run hf.co/clibrain/Llama-2-ft-instruct-es
```

Inference Error

by eduiqe - opened Jul 24, 2023

Discussion

eduiqe

Jul 24, 2023

I tried to use your code snippet with transformers library version==4.31.0, but all the model generates is this character "\x0f".

I was wondering if I should be using another version of transformers.

Is there any fix to this problem?

mrm8488

CliBrAIn org Jul 26, 2023

Hi, @eduiqe . Have you tried to upgrade bitsandbytes to the highest version?

eduiqe

Jul 28, 2023

•

edited Jul 28, 2023

Hi, thanks for the response.
I´m using this versions of libraries:
torch == 1.11.0 and 2.0.0
transformers == 4.31.0
bitsandbytes == 0.41.0 (latest)
and I am running this code with cpu instead of cuda, which shouldn´t affect the result (?)

Is there anything else I can try to make the completion work?
Thanks again!

enric1296

Aug 1, 2023

same error here

PauFp20

Aug 6, 2023

Same

mrm8488

CliBrAIn org Aug 6, 2023

Thanks for your feedback. Working on fixing it

Napuh

Aug 7, 2023

Same error here. Using load_in_8bit and with auto device map:
model = AutoModelForCausalLM.from_pretrained(model_id, trust_remote_code=True, load_in_8bit = True, device_map='auto')

Response to all inputs is:
ed 10c30c30c30c300c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30c30-slash.

mrm8488

CliBrAIn org Aug 9, 2023

Hi, guys!
Problem solved here: https://huggingface.co/clibrain/Llama-2-7b-ft-instruct-es

mrm8488 changed discussion status to closed Aug 9, 2023

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