Upload FP32 ONNX + model card (114.6 MB) from Drive

README.md

@@ -21,17 +21,17 @@ Fine-tune de `microsoft/swin-small-patch4-window7-224` para clasificar defectos
 | `corner_wear` | none / light / medium / heavy |
 | `edge_whitening` | none / light / medium / heavy |
 
-## Métricas (mejor epoch: 3)
+## Métricas (mejor epoch: ?)
 
 | Métrica | Valor |
 |---|---|
-| val_loss | 1.6629 |
-| Exactitud surface | **90.3%** |
-| Exactitud corners | **93.9%** |
-| Exactitud edges | **97.0%** |
-| **Exactitud media** | **93.8%** |
+| val_loss | ? |
+| Exactitud surface | **?%** |
+| Exactitud corners | **?%** |
+| Exactitud edges | **?%** |
+| **Exactitud media** | **?%** |
 
-Entrenado con **2000 cartas** (30 épocas).
+Entrenado con **? cartas** (? épocas).
 
 ## Uso (ONNX)
 
@@ -44,4 +44,4 @@ sess = ort.InferenceSession('card_defect_classifier.onnx')
 surface, corners, edges = sess.run(None, {'pixel_values': img})
 ```
 
-Clases: `['none', 'micro', 'light', 'medium', 'heavy']` / `['none', 'light', 'medium', 'heavy']` / `['none', 'light', 'medium', 'heavy']`
+Clases: `['none','micro','light','medium','heavy']` / `['none','light','medium','heavy']` / `['none','light','medium','heavy']`

best_model.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:efd2baf1736755ebd4b10c44d09716af5702a317fac9af33bf231431616716c2
-size 198413218
+oid sha256:52c08192096f9c03f7a3a761ed83321dd3764d65ad56719a7f519a0b65f653e5
+size 112813386

card_defect_classifier.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3535f4206564766c5dc5472db03b6a5b616a2be10fcbe37ed24bcc6d144599e3
-size 51120570
+oid sha256:d191dc9f6d836afce9a41dc877bbae790c58084b56180107c6f23b9361bdf62e
+size 114605029

model_metadata.json

@@ -29,15 +29,5 @@
     0.224,
     0.225
   ],
-  "backbone": "microsoft/swin-small-patch4-window7-224",
-  "metrics": {
-    "best_epoch": 3,
-    "val_loss": 1.6629,
-    "acc_surface_pct": 90.3,
-    "acc_corners_pct": 93.9,
-    "acc_edges_pct": 97.0,
-    "acc_mean_pct": 93.8,
-    "num_cards": 2000,
-    "epochs_trained": 30
-  }
+  "backbone": "microsoft/swin-tiny-patch4-window7-224"
 }