Instructions to use TinyPixel/Llama-2-7B-bf16-sharded with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TinyPixel/Llama-2-7B-bf16-sharded with Transformers:

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

pipe = pipeline("text-generation", model="TinyPixel/Llama-2-7B-bf16-sharded")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("TinyPixel/Llama-2-7B-bf16-sharded")
model = AutoModelForCausalLM.from_pretrained("TinyPixel/Llama-2-7B-bf16-sharded")

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use TinyPixel/Llama-2-7B-bf16-sharded with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TinyPixel/Llama-2-7B-bf16-sharded"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TinyPixel/Llama-2-7B-bf16-sharded",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/TinyPixel/Llama-2-7B-bf16-sharded

SGLang

How to use TinyPixel/Llama-2-7B-bf16-sharded 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 "TinyPixel/Llama-2-7B-bf16-sharded" \
    --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": "TinyPixel/Llama-2-7B-bf16-sharded",
		"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 "TinyPixel/Llama-2-7B-bf16-sharded" \
        --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": "TinyPixel/Llama-2-7B-bf16-sharded",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use TinyPixel/Llama-2-7B-bf16-sharded with Docker Model Runner:
```
docker model run hf.co/TinyPixel/Llama-2-7B-bf16-sharded
```

I want to shard another model how can I shard?

by mandeepbagga - opened Aug 21, 2023

Discussion

mandeepbagga

Aug 21, 2023

Hi, hope you are doing well. I'm trying to use another model to finetune it further, but on google collab, memory is getting filled. So, I want to shard the model first I know it will cost me some money, but I'll be able to finetune it on collab later.
It will be cool if someone can tell me if we can use CPU ram to shard a model.

Thanks

TinyPixel

Owner Aug 21, 2023

Just load the model and do
model.push_to_hub("", max_shard_size="2000MB", use_auth_token=True)
Also push the tokenizer
tokenizer.push_to_hub("", use_auth_token=True)

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