TitanML/bakLlava-v1-hf

BakLLaVA Model Card

BakLlava is a model that is derived from the original Llava architecture, that uses Mistral-7b as a text backbone.

Below is the model card of BakLlava model 7b, which is copied from the original BakLlava model card that you can find here.

BakLLaVA 1 is a Mistral 7B base augmented with the LLaVA 1.5 architecture. In this first version, we showcase that a Mistral 7B base outperforms Llama 2 13B on several benchmarks. You can run BakLLaVA-1 on our repo. We are currently updating it to make it easier for you to finetune and inference. (https://github.com/SkunkworksAI/BakLLaVA).

Note: BakLLaVA-1 is fully open-source but was trained on certain data that includes LLaVA's corpus which is not commercially permissive. We will fix this in the upcoming release.

BakLLaVA 2 is cooking with a significantly larger (commercially viable) dataset and a novel architecture that expands beyond the current LLaVA method. BakLLaVA-2 will do away with the restrictions of BakLLaVA-1.

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 (USER: xxx\nASSISTANT:) and add the token <image> to the location where you want to query images:

Check out also the Google Colab demo to run Llava on a free-tier Google Colab instance:

Or check out our Spaces demo!

Using pipeline:

from transformers import pipeline
from PIL import Image    
import requests

model_id = "llava-hf/bakLlava-v1-hf"
pipe = pipeline("image-to-text", model=model_id)

url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tasks/ai2d-demo.jpg"
image = Image.open(requests.get(url, stream=True).raw)

# Define a chat history 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 does the label 15 represent? (1) lava (2) core (3) tunnel (4) ash cloud"},
          {"type": "image"},
        ],
    },
]
prompt = processor.apply_chat_template(conversation, add_generation_prompt=True)

outputs = pipe(image, prompt=prompt, generate_kwargs={"max_new_tokens": 200})
print(outputs)
>>> {"generated_text": "\nUSER: What does the label 15 represent? (1) lava (2) core (3) tunnel (4) ash cloud\nASSISTANT: 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, LlavaForConditionalGeneration

model_id = "llava-hf/bakLlava-v1-hf"
model = LlavaForConditionalGeneration.from_pretrained(
    model_id, 
    torch_dtype=torch.float16, 
    low_cpu_mem_usage=True, 
).to(0)

processor = AutoProcessor.from_pretrained(model_id)

# Define a chat history 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))

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

Evaluations

Training dataset

• 558K filtered image-text pairs from LAION/CC/SBU, captioned by BLIP.
• 158K GPT-generated multimodal instruction-following data.
• 450K academic-task-oriented VQA data mixture.
• 40K ShareGPT data.
• Additional private data (permissive)

License

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