Instructions to use Henil1/vit-axavision-2-ChestX with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Henil1/vit-axavision-2-ChestX with Transformers:

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

pipe = pipeline("image-text-to-text", model="Henil1/vit-axavision-2-ChestX")

# Load model directly
from transformers import AutoTokenizer, AutoModelForImageTextToText

tokenizer = AutoTokenizer.from_pretrained("Henil1/vit-axavision-2-ChestX")
model = AutoModelForImageTextToText.from_pretrained("Henil1/vit-axavision-2-ChestX")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Henil1/vit-axavision-2-ChestX with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Henil1/vit-axavision-2-ChestX"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Henil1/vit-axavision-2-ChestX",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Henil1/vit-axavision-2-ChestX

SGLang

How to use Henil1/vit-axavision-2-ChestX 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 "Henil1/vit-axavision-2-ChestX" \
    --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": "Henil1/vit-axavision-2-ChestX",
		"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 "Henil1/vit-axavision-2-ChestX" \
        --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": "Henil1/vit-axavision-2-ChestX",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Henil1/vit-axavision-2-ChestX with Docker Model Runner:
```
docker model run hf.co/Henil1/vit-axavision-2-ChestX
```

Vit-Axavision-2-ChestX 🩺

This model is a fine-tuned version of nlpconnect/vit-gpt2-image-captioning on a chest X-ray dataset. It is developed as part of the Axamine AI research efforts to explore medical vision-language applications. The model takes chest X-ray images as input and generates descriptive captions that may help in automated reporting, healthcare research, or AI-assisted diagnostics.

Model Details

Base model: nlpconnect/vit-gpt2-image-captioning
Architecture: VisionEncoderDecoderModel (ViT encoder + GPT2 decoder)
Fine-tuned on dataset: Shrey-1329/cxiu_hf_dataset
Model size: ~250M parameters
Developed by: Henilsinh Raj (Axamine AI)

Use Cases

Intended Use

Chest X-ray image captioning
Healthcare research
Medical AI experiments
Educational purposes

Limitations

This model does not provide medical diagnosis.
Captions are purely descriptive and may not fully reflect clinical accuracy.

Usage

Here’s how you can use the model for inference:

from transformers import VisionEncoderDecoderModel, ViTImageProcessor, AutoTokenizer
from PIL import Image
import torch
import requests

# Load model
model_id = "Henil1/vit-axavision-2-ChestX"
model = VisionEncoderDecoderModel.from_pretrained(model_id)
feature_extractor = ViTImageProcessor.from_pretrained(model_id)
tokenizer = AutoTokenizer.from_pretrained(model_id)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

# Preprocess image
image = Image.open("your_image_path.jpg").convert("RGB")
pixel_values = feature_extractor(images=image, return_tensors="pt").pixel_values.to(device)

# Generate caption
output_ids = model.generate(pixel_values, max_length=64, num_beams=4)
caption = tokenizer.decode(output_ids[0], skip_special_tokens=True)
print("Generated caption:", caption)

Citation

If you use this model, please cite:

@misc{henil2025axavision,
  author = {Henilsinh Raj},
  title = {Vit-Axavision-2-ChestX: Vision-Language Model for Chest X-Ray Captioning},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/Henil1/vit-axavision-2-ChestX}
}

Downloads last month: 1

Safetensors

Model size

0.2B params

Tensor type

F32

Inference Providers NEW

Image-Text-to-Text

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support