---
license: creativeml-openrail-m
language:
- en
metrics:
- bleu
tags:
  - endpoints
  - text-generation-inference
inference: true
---

<h3 align='center' style='font-size: 24px;'>Blazzing Fast Tiny Vision Language Model</h3>


<p align='center', style='font-size: 16px;' >A Custom 3B parameter Model. Built by <a href="https://www.linkedin.com/in/manishkumarthota/">@Manish</a> The model is released for research purposes only, commercial use is not allowed. </p>

## How to use


**Install dependencies**
```bash
pip install transformers # latest version is ok, but we recommend v4.31.0
pip install -q pillow accelerate einops
```

You can use the following code for model inference. The format of text instruction is similar to [LLaVA](https://github.com/haotian-liu/LLaVA).

```Python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from PIL import Image

torch.set_default_device("cuda")

#Create model
model = AutoModelForCausalLM.from_pretrained(
    "ManishThota/CustomModel", 
    torch_dtype=torch.float16, 
    device_map="auto",
    trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("ManishThota/CustomModel", trust_remote_code=True)

#function to generate the answer
def predict(question, image_path):
    #Set inputs
    text = f"USER: <image>\n{question}? ASSISTANT:"
    image = Image.open(image_path)
    
    input_ids = tokenizer(text, return_tensors='pt').input_ids.to('cuda')
    image_tensor = model.image_preprocess(image)
    
    #Generate the answer
    output_ids = model.generate(
        input_ids,
        max_new_tokens=25,
        images=image_tensor,
        use_cache=True)[0]
    
    return tokenizer.decode(output_ids[input_ids.shape[1]:], skip_special_tokens=True).strip()

```