vip-llava-13b-hf

VipLLaVA Model Card

Below is the model card of VipLlava model 13b, which is copied from the original Llava model card that you can find here.

Check out also the Google Colab demo to run Llava on a free-tier Google Colab instance (the model works similarly as Llava):

Or check out our Spaces demo!

Model details

Model type: LLaVA is an open-source chatbot trained by fine-tuning LLaMA/Vicuna on GPT-generated multimodal instruction-following data. It is an auto-regressive language model, based on the transformer architecture.

Vip-LlaVa enhances the training protocol of Llava by marking images and interact with the model using natural cues like a “red bounding box” or “pointed arrow” during training.

Model date: ViP-LLaVa was released in December 2023.

Paper or resources for more information: https://vip-llava.github.io/

How to use the model

First, make sure to have transformers >= 4.35.3. The model supports multi-image and multi-prompt generation. Meaning that you can pass multiple images in your prompt. Make sure also to follow the correct prompt template and add the token <image> to the location where you want to query images:

According to the official code base, it is recommeneded to use this template:

A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.###Human: <image>\n<prompt>###Assistant:

Where <prompt> denotes the prompt asked by the user

Using pipeline:

from transformers import pipeline

pipe = pipeline("image-text-to-text", model="llava-hf/vip-llava-13b-hf")
messages = [
    {
      "role": "user",
      "content": [
          {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tasks/ai2d-demo.jpg"},
          {"type": "text", "text": "What does the label 15 represent? (1) lava (2) core (3) tunnel (4) ash cloud"},
        ],
    },
]

out = pipe(text=messages, max_new_tokens=20)
print(out)
>>> [{'input_text': [{'role': 'user', 'content': [{'type': 'image', 'url': 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tasks/ai2d-demo.jpg'}, {'type': 'text', 'text': 'What does the label 15 represent? (1) lava (2) core (3) tunnel (4) ash cloud'}]}], 'generated_text': 'Lava'}]

Using pure transformers:

Below is an example script to run generation in float16 precision on a GPU device:

import requests
from PIL import Image

import torch
from transformers import AutoProcessor, VipLlavaForConditionalGeneration

model_id = "llava-hf/vip-llava-13b-hf"
model = VipLlavaForConditionalGeneration.from_pretrained(
    model_id, 
    torch_dtype=torch.float16, 
    low_cpu_mem_usage=True, 
).to(0)

processor = AutoProcessor.from_pretrained(model_id)

# Define a chat histiry and use `apply_chat_template` to get correctly formatted prompt
# Each value in "content" has to be a list of dicts with types ("text", "image") 
conversation = [
    {

      "role": "user",
      "content": [
          {"type": "text", "text": "What are these?"},
          {"type": "image"},
        ],
    },
]
prompt = processor.apply_chat_template(conversation, add_generation_prompt=True)

image_file = "http://images.cocodataset.org/val2017/000000039769.jpg"
raw_image = Image.open(requests.get(image_file, stream=True).raw)
inputs = processor(images=raw_image, text=prompt, return_tensors='pt').to(0, torch.float16)

output = model.generate(**inputs, max_new_tokens=200, do_sample=False)
print(processor.decode(output[0][2:], skip_special_tokens=True))

---

From transformers>=v4.48, you can also pass image url or local path to the conversation history, and let the chat template handle the rest. Chat template will load the image for you and return inputs in torch.Tensor which you can pass directly to model.generate()

messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://www.ilankelman.org/stopsigns/australia.jpg"}
            {"type": "text", "text": "What is shown in this image?"},
        ],
    },
]

inputs = processor.apply_chat_template(messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors"pt")
output = model.generate(**inputs, max_new_tokens=50)

Model optimization

4-bit quantization through bitsandbytes library

First make sure to install bitsandbytes, pip install bitsandbytes and make sure to have access to a CUDA compatible GPU device. Simply change the snippet above with:

model = VipLlavaForConditionalGeneration.from_pretrained(
    model_id, 
    torch_dtype=torch.float16, 
    low_cpu_mem_usage=True,
+   load_in_4bit=True
)

Use Flash-Attention 2 to further speed-up generation

First make sure to install flash-attn. Refer to the original repository of Flash Attention regarding that package installation. Simply change the snippet above with:

model = VipLlavaForConditionalGeneration.from_pretrained(
    model_id, 
    torch_dtype=torch.float16, 
    low_cpu_mem_usage=True,
+   use_flash_attention_2=True
).to(0)

License

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

Citation

To cite this work please use

@misc{cai2023making,
      title={Making Large Multimodal Models Understand Arbitrary Visual Prompts}, 
      author={Mu Cai and Haotian Liu and Siva Karthik Mustikovela and Gregory P. Meyer and Yuning Chai and Dennis Park and Yong Jae Lee},
      year={2023},
      eprint={2312.00784},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}