CheXpert-5-convnextv2-tiny-384
This model is a fine-tuned version of facebook/convnextv2-tiny-22k-384 on the CheXpert dataset. It achieves the following results on the evaluation set:
- Loss: 0.1009
- Auroc Atelectasis: 0.7943
- Auroc Cardiomegaly: 0.8187
- Auroc Consolidation: 0.9269
- Auroc Edema: 0.9233
- Auroc Pleural effusion: 0.9315
- Specificity Atelectasis: 0.7891
- Specificity Cardiomegaly: 1.0
- Specificity Consolidation: 0.9948
- Specificity Edema: 0.8407
- Specificity Pleural effusion: 0.8038
- Exact Match: 0.4464
- Hamming Distance: 0.1804
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 2500
- num_epochs: 6
Training results
Training Loss | Epoch | Step | Validation Loss | Auroc Atelectasis | Auroc Cardiomegaly | Auroc Consolidation | Auroc Edema | Auroc Pleural effusion | Specificity Atelectasis | Specificity Cardiomegaly | Specificity Consolidation | Specificity Edema | Specificity Pleural effusion | Exact Match | Hamming Distance |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0.0891 | 1.0 | 6120 | 0.0893 | 0.7323 | 0.8366 | 0.7020 | 0.8387 | 0.8702 | 0.4510 | 0.9661 | 1.0 | 0.6596 | 0.5392 | 0.2616 | 0.2444 |
0.0854 | 2.0 | 12240 | 0.0831 | 0.7535 | 0.8556 | 0.7350 | 0.8651 | 0.8881 | 0.7293 | 0.9042 | 0.9936 | 0.7083 | 0.6259 | 0.3571 | 0.1973 |
0.082 | 3.0 | 18360 | 0.0824 | 0.7683 | 0.8696 | 0.7473 | 0.8720 | 0.8961 | 0.6956 | 0.8196 | 0.9881 | 0.6087 | 0.6611 | 0.3298 | 0.2177 |
0.0799 | 4.0 | 24480 | 0.0802 | 0.7749 | 0.8720 | 0.7562 | 0.8783 | 0.9005 | 0.7450 | 0.8831 | 0.9608 | 0.7341 | 0.6984 | 0.3802 | 0.1880 |
0.0759 | 5.0 | 30600 | 0.0793 | 0.7795 | 0.8746 | 0.7583 | 0.8818 | 0.9030 | 0.7277 | 0.8948 | 0.9711 | 0.7618 | 0.7045 | 0.3869 | 0.1823 |
0.0739 | 6.0 | 36720 | 0.0798 | 0.7787 | 0.8727 | 0.7561 | 0.8812 | 0.9031 | 0.7461 | 0.8921 | 0.9690 | 0.7487 | 0.7074 | 0.3886 | 0.1824 |
Framework versions
- Transformers 4.41.0
- Pytorch 2.1.2
- Datasets 2.19.1
- Tokenizers 0.19.1
- Downloads last month
- 27
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Model tree for shreydan/CheXpert-5-convnextv2-tiny-384
Base model
facebook/convnextv2-tiny-22k-384