Instructions to use qnguyen3/nanoLLaVA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use qnguyen3/nanoLLaVA with Transformers:

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

pipe = pipeline("text-generation", model="qnguyen3/nanoLLaVA", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("qnguyen3/nanoLLaVA", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use qnguyen3/nanoLLaVA with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "qnguyen3/nanoLLaVA"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "qnguyen3/nanoLLaVA",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/qnguyen3/nanoLLaVA

SGLang

How to use qnguyen3/nanoLLaVA 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 "qnguyen3/nanoLLaVA" \
    --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": "qnguyen3/nanoLLaVA",
		"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 "qnguyen3/nanoLLaVA" \
        --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": "qnguyen3/nanoLLaVA",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use qnguyen3/nanoLLaVA with Docker Model Runner:
```
docker model run hf.co/qnguyen3/nanoLLaVA
```

Multi-round conversation w/ PKV cache example code

by Xenova - opened May 18, 2024

Discussion

Xenova

May 18, 2024

Hi there! As seen in your README, the model seemingly supports multi-round conversations. Does this also work with passing past key values? If so, could you provide example code for this, as it will dramatically improve performance? Thanks!

qnguyen3

Owner May 18, 2024

Hi @Xenova , i honestly do not know the answer, i will look into it to see if it is possible.

Xenova

May 18, 2024

Great! It will greatly speed up time-to-first-token for the web demo I'm working on. If it doesn't work, then it's alright, it will produce the same results, just a bit slower since it needs to recompute KV cache on second run.

Xenova

May 18, 2024

Okay I've got it working! Currently doesn't work in transformers due to a bug here (it always just looks at the last token when past KV cache is passed in by the user, even when user specifies > 1 new input token).

I've updated this in transformers.js and will put out a demo with this.

Xenova

May 18, 2024

I've updated the model card + released the demo! :)

Model: https://huggingface.co/Xenova/nanoLLaVA
Demo: https://huggingface.co/spaces/Xenova/experimental-nanollava-webgpu
Video:

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