Instructions to use explosion-testing/llama2-kv-sharing with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use explosion-testing/llama2-kv-sharing with Transformers:

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

pipe = pipeline("text-generation", model="explosion-testing/llama2-kv-sharing")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("explosion-testing/llama2-kv-sharing")
model = AutoModelForCausalLM.from_pretrained("explosion-testing/llama2-kv-sharing", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use explosion-testing/llama2-kv-sharing with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "explosion-testing/llama2-kv-sharing"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "explosion-testing/llama2-kv-sharing",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/explosion-testing/llama2-kv-sharing

SGLang

How to use explosion-testing/llama2-kv-sharing 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 "explosion-testing/llama2-kv-sharing" \
    --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": "explosion-testing/llama2-kv-sharing",
		"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 "explosion-testing/llama2-kv-sharing" \
        --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": "explosion-testing/llama2-kv-sharing",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use explosion-testing/llama2-kv-sharing with Docker Model Runner:
```
docker model run hf.co/explosion-testing/llama2-kv-sharing
```

llama2-kv-sharing

Commit History

Aug 1, 2023

05:36
Adding `safetensors` variant of this model⁠

SFconvertbot

ff09167

Jul 26, 2023

12:58
Causal LM

Daniël de Kok

5317cfa
12:54
Add model with kv head sharing

Daniël de Kok

568e54d
12:54
initial commit

Daniël de Kok

03eb517

Commit History

Aug 1, 2023

Adding `safetensors` variant of this model⁠

Jul 26, 2023

Causal LM

Add model with kv head sharing

initial commit