CheXpert chest X-ray fine-tune
According to the model card, the model is a Vision Transformer fine-tuned on the CheXpert dataset for classifying lung conditions such as pneumonia and cardiomegaly.
Open Source Model Profile · codewithdark
vit-chest-xray
vit-chest-xray is an 85.80M-parameter ViT chest X-ray classifier from codewithdark. According to the model card, it was fine-tuned on CheXpert with 98.46% reported validation accuracy.
- Publisher
- codewithdark
- Task
- image-classification
- Model type
- vit
- License
- mit
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
vit-chest-xray is published by codewithdark as an image-classification model. The captured configuration identifies ViTForImageClassification with model type vit, and Safetensors metadata reports 85,802,501 parameters. According to the model card, it is a Vision Transformer fine-tuned on the CheXpert dataset for chest X-ray disease classification.
Recorded capabilities
Publisher-reported 98.46% validation accuracy
According to the model card, final validation accuracy is 98.46% with training loss 0.1069 and validation loss 0.0980.
Documented ten-epoch training setup
According to the model card, fine-tuning used AdamW, learning rate 3e-05, batch size 32, 10 epochs, binary cross-entropy with logits, and mixed precision.
Documented five-label inference workflow
According to the model card, inference loads the checkpoint with AutoImageProcessor and AutoModelForImageClassification and maps outputs to Cardiomegaly, Edema, Consolidation, Pneumonia, and No Finding.
Use cases in the source record
- Chest X-ray image classification research using the documented processor, model-loading, and five-label inference workflow.
- Dataset-methodology study of CheXpert fine-tuning, referencing the publisher-reported accuracy and multi-label annotation context.
Limitations and unknowns
- No independent evaluation results were extracted; accuracy and loss figures are publisher-reported.
- No clinical validation, intended-use clearance, or deployment safety evidence was extracted; medical use requires independent expert review.
- Provider state is historical snapshot data, not independently refreshed current availability.
Source and provenance
Source: codewithdark/vit-chest-xray
Captured: Unknown. Processed: 2026-09-07T19:35:20.501943+00:00.
Chest X-ray Image Classifier This repository contains a fine-tuned Vision Transformer (ViT) model for classifying chest X-ray images, utilizing the CheXpert dataset. The model is fine-tuned on the task of classifying various lung diseases from chest radiographs, achieving impressive accuracy in distinguishing between different conditions. Model Overview The fine-tuned Model Code on github is based on the Vision Transformer (ViT) architecture, which excels in handling image-based tasks by leveraging attention mechanisms for efficient feature extraction. The model was trained on the CheXpert dataset , which consists of labeled chest X…
F001F002F003F004F005F006F007F010F011F012F013F014F016F017