Instructions to use ConvLLaVA/ConvLLaVA-pretrain-1024 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ConvLLaVA/ConvLLaVA-pretrain-1024 with Transformers:

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

pipe = pipeline("text-generation", model="ConvLLaVA/ConvLLaVA-pretrain-1024")

# Load model directly
from transformers import AutoProcessor, AutoModelForCausalLM

processor = AutoProcessor.from_pretrained("ConvLLaVA/ConvLLaVA-pretrain-1024")
model = AutoModelForCausalLM.from_pretrained("ConvLLaVA/ConvLLaVA-pretrain-1024", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ConvLLaVA/ConvLLaVA-pretrain-1024 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ConvLLaVA/ConvLLaVA-pretrain-1024"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ConvLLaVA/ConvLLaVA-pretrain-1024",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ConvLLaVA/ConvLLaVA-pretrain-1024

SGLang

How to use ConvLLaVA/ConvLLaVA-pretrain-1024 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 "ConvLLaVA/ConvLLaVA-pretrain-1024" \
    --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": "ConvLLaVA/ConvLLaVA-pretrain-1024",
		"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 "ConvLLaVA/ConvLLaVA-pretrain-1024" \
        --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": "ConvLLaVA/ConvLLaVA-pretrain-1024",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ConvLLaVA/ConvLLaVA-pretrain-1024 with Docker Model Runner:
```
docker model run hf.co/ConvLLaVA/ConvLLaVA-pretrain-1024
```

ConvLLaVA-pretrain-1024 / README.md

ChunjiangGe

Update README.md

fcb9047 verified over 2 years ago

preview code

Raw

History Blame Contribute Delete

1.41 kB

	---
	datasets:
	- Lin-Chen/ShareGPT4V
	- FreedomIntelligence/ALLaVA-4V
	- Vision-Flan/vision-flan_191-task_1k
	---
	# ConvLLaVA Model Card

	## Model details

	Model type: ConvLLaVA is an open-source chatbot trained by fine-tuning LLM on multimodal instruction-following data. It is an auto-regressive language model, based on the transformer architecture. Base LLM: lmsys/vicuna-7b-v1.5

	Model date: ConvLLaVA-pretrain-1024 was trained in March 2024.

	Paper or resources for more information: https://github.com/alibaba/conv-llava/

	## License
	Llama 2 is licensed under the LLAMA 2 Community License, Copyright (c) Meta Platforms, Inc. All Rights Reserved.

	Where to send questions or comments about the model: https://github.com/alibaba/conv-llava/issues

	## Intended use
	Primary intended uses: The primary use of ConvLLaVA is research on large multimodal models and chatbots.

	Primary intended users: The primary intended users of the model are researchers and hobbyists in computer vision, natural language processing, machine learning, and artificial intelligence.

	## Training dataset

	- 1.2M ShareGPT4V-PT caption data.
	- 100K ShareGPT4V caption data.
	- 1.4M ALLaVA caption and instruction data.
	- 186K VFLAN multitask data.
	- 158K GPT-generated multimodal instruction-following data.
	- 500K academic-task-oriented VQA data mixture.
	- 40K ShareGPT data.

	## Paper
	arxiv.org/abs/2405.15738