-
Quantize FP32→INT8: 114.6MB→29.1MB
-
Upload FP32 ONNX + model card | acc_media=96.4% | val_loss=0.1163 | epoch=28
-
Upload FP32 ONNX + model card (114.6 MB) from Drive
-
Quantize FP32→INT8: 201.2MB→51.1MB
-
Upload model v3 — val_loss=1.6629 | acc_media=93.8%
-
Upload model v8 — val_loss=1.5758 | acc_media=92.8%
-
Upload model v9 — val_loss=1.8386 | acc_media=87.9%
-
Upload training_curves.png with huggingface_hub
-
Upload card_defect_classifier.onnx with huggingface_hub
-
Upload training_curves.png with huggingface_hub
-
Upload card_defect_classifier.onnx with huggingface_hub
-
Upload training_curves.png with huggingface_hub
-
Upload model_metadata.json with huggingface_hub
-
Upload card_defect_classifier.onnx with huggingface_hub
-
initial commit
