You can deploy this model and start earning money today!

gradeeye/
four-ch-tversky

Name: four-ch-tversky
Availability: InStock

domain-generalization

image-classification

cc-by-nc-4.0

Views

No views yet

GradeEye four-ch-tversky

This repository contains GradeEye CORN ordinal diabetic-retinopathy classifiers using the convnext_tiny backbone at 384x384 resolution, with 4-channel input and a 4-threshold ordinal head. The _ema.safetensors file is the primary artifact: the paper's reported evaluation metrics were generated using the EMA state dict. The unsuffixed .safetensors file is the corresponding raw model_state_dict secondary artifact.

Checkpoints

Primary EMA weights	Raw secondary weights	Architecture	Held-out fold	Best QWK	Epoch
`lodo_aptos_convnext_tiny_best_ema.safetensors`	`lodo_aptos_convnext_tiny_best.safetensors`	`convnext_tiny`	varies	0.7539	4
`lodo_ddr_convnext_tiny_best_ema.safetensors`	`lodo_ddr_convnext_tiny_best.safetensors`	`convnext_tiny`	varies	0.7433	7
`lodo_eyepacs_convnext_tiny_best_ema.safetensors`	`lodo_eyepacs_convnext_tiny_best.safetensors`	`convnext_tiny`	varies	0.7808	3
`lodo_messidor2_convnext_tiny_best_ema.safetensors`	`lodo_messidor2_convnext_tiny_best.safetensors`	`convnext_tiny`	varies	0.7307	7

Preprocessing

Resize the RGB fundus image to 384x384 using the same offline preprocessing pipeline.
Convert RGB to float in [0,1] and apply ImageNet normalization: mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225).
Append one auxiliary channel in [0,1] after RGB normalization. It is not ImageNet-normalized.

Fourth-channel construction

Source pool: segmentation_pooled_tversky.
Producer: gradeeye/seg-unet-tversky.
Exact method: Raw sigmoid probability map from the Tversky-trained U-Net; uint16 PNG divided by 65535 to [0,1]. No morphological post-processing.
The resulting tensor is (4, 384, 384) in channel-first layout.
No opening, closing, or Gaussian blur is applied in this variant; only the segmenter training loss differs from the BCE+Dice producer.

Loading

python

import json
from modeling import load_model

config = json.load(open('config.json'))
model = load_model('lodo_eyepacs_convnext_tiny_best_ema.safetensors', config)
# model(x) returns CORN logits with shape (batch, 4)

Install torch, timm, and safetensors, and make the GradeEye source repository available on PYTHONPATH.

Intended use and limitations

These weights are released for research and reproducibility only. They are not validated for clinical diagnosis or treatment decisions. Performance varies substantially by held-out dataset and should not be interpreted as clinical-grade generalization.

Source code and paper materials: https://github.com/ahmed-farhanur-rashid/gradeeye.

gradeeye/four-ch-tversky