Skip to content
Other

hybrid-radiograph-feature-fusion

hybrid models for chest X-ray analysis

2 min readupdated

A set of pretrained PyTorch checkpoints for medical image classification, combining modern AI model architectures (ConvNeXt-Tiny and Swin Transformer-T) with classical computer vision descriptors (HOG, LBP, Gabor, and SIFT-BoVW). The model tackles binary classification of chest X-rays using the Kermany Pediatric Chest X-Ray Cohort dataset. This hybrid approach improves accuracy by extracting both deep and textural image features together. The solution will be useful for researchers and developers in medical diagnostic imaging.

Key features

  • Combination of AI model backbones (ConvNeXt-Tiny and Swin-T) with classical image descriptors
  • Pretrained weights for various feature fusion strategies (concatenation, Gated Cross-Modal, Token Attention)
  • Tested on the pediatric chest X-ray dataset (225 test images)
  • Full compatibility with the PyTorch ecosystem and Hugging Face Hub utilities

What you can do

  • Integrate pretrained weights into automated X-ray analysis pipelines
  • Conduct a comparative analysis of the effectiveness of deep and hybrid features in medical CV
  • Run inference for the desired checkpoint using the provided loading script

Example

import torch
from huggingface_hub import hf_hub_download
from src.models import HybridMedicalClassifier

ckpt_path = hf_hub_download(
    repo_id="subhan1501/hybrid-radiograph-feature-fusion",
    filename="exp6_swin_sift.pth",
)

model = HybridMedicalClassifier(
    backbone_name="swin_t",
    fusion_type="token_attention",
    hand_dim=208,
    embed_dim=256,
    num_classes=2,
    pretrained=False,
)
model.load_state_dict(torch.load(ckpt_path, map_location="cpu"))
model.eval()
Read moreDetails
• Test accuracy: up to 94.22% (exp1_convnext_baseline.pth)
• Dataset: Kermany Pediatric Chest X-Ray Cohort (1500 images: 70% Train / 15% Val / 15% Test)
• Preprocessing: Symmetric CLAHE (clipLimit=2.0, tileGrid=(8, 8))
• License: MIT License
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.1 GB · ready files0.6 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 GB
  • 4-bitweights 0.0 GB0.5 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # subhan1501/hybrid-radiograph-feature-fusion
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download subhan1501/hybrid-radiograph-feature-fusion --local-dir ~/models/hybrid-radiograph-feature-fusion
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of hybrid-radiograph-feature-fusion (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

subhan1501/hybrid-radiograph-feature-fusion
Author
subhan1501
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments