End of training

README.md

@@ -1,7 +1,7 @@
 ---
 library_name: transformers
-license: other
-base_model: google/mobilenet_v2_1.0_224
+license: apache-2.0
+base_model: google/vit-large-patch16-224
 tags:
 - generated_from_trainer
 metrics:
@@ -19,17 +19,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # ViT_L16
 
-This model is a fine-tuned version of [google/mobilenet_v2_1.0_224](https://huggingface.co/google/mobilenet_v2_1.0_224) on an unknown dataset.
+This model is a fine-tuned version of [google/vit-large-patch16-224](https://huggingface.co/google/vit-large-patch16-224) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3409
-- Accuracy: 0.8904
-- Precision: 0.8746
-- Recall: 0.8901
-- F1: 0.8823
-- Tp: 1458
-- Tn: 1701
-- Fp: 209
-- Fn: 180
+- Loss: 0.0933
+- Accuracy: 0.9746
+- Precision: 0.9844
+- Recall: 0.9603
+- F1: 0.9722
+- Tp: 1573
+- Tn: 1885
+- Fp: 25
+- Fn: 65
 
 ## Model description
 
@@ -49,38 +49,26 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 5e-06
-- train_batch_size: 128
-- eval_batch_size: 128
+- train_batch_size: 32
+- eval_batch_size: 32
 - seed: 42
 - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 137
-- num_epochs: 5
+- lr_scheduler_warmup_steps: 221
+- num_epochs: 2
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss | Accuracy | Precision | Recall | F1     | Tp   | Tn   | Fp  | Fn  |
-|:-------------:|:------:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:----:|:----:|:---:|:---:|
-| 0.6841        | 0.2432 | 27   | 0.6509          | 0.6088   | 0.5711    | 0.6129 | 0.5913 | 1004 | 1156 | 754 | 634 |
-| 0.6551        | 0.4865 | 54   | 0.6145          | 0.6700   | 0.6292    | 0.6941 | 0.6601 | 1137 | 1240 | 670 | 501 |
-| 0.5709        | 0.7297 | 81   | 0.6052          | 0.6750   | 0.6174    | 0.7784 | 0.6886 | 1275 | 1120 | 790 | 363 |
-| 0.4953        | 0.9730 | 108  | 0.5217          | 0.7709   | 0.6878    | 0.9225 | 0.7880 | 1511 | 1224 | 686 | 127 |
-| 0.4121        | 1.2162 | 135  | 0.4447          | 0.8588   | 0.8118    | 0.9035 | 0.8552 | 1480 | 1567 | 343 | 158 |
-| 0.3688        | 1.4595 | 162  | 0.4108          | 0.8557   | 0.8043    | 0.9084 | 0.8532 | 1488 | 1548 | 362 | 150 |
-| 0.3300        | 1.7027 | 189  | 0.4537          | 0.7889   | 0.7014    | 0.9451 | 0.8052 | 1548 | 1251 | 659 | 90  |
-| 0.3324        | 1.9459 | 216  | 0.3450          | 0.9039   | 0.8947    | 0.8974 | 0.8961 | 1470 | 1737 | 173 | 168 |
-| 0.3033        | 2.1892 | 243  | 0.4286          | 0.8081   | 0.7216    | 0.9512 | 0.8206 | 1558 | 1309 | 601 | 80  |
-| 0.3007        | 2.4324 | 270  | 0.3362          | 0.9056   | 0.9095    | 0.8834 | 0.8963 | 1447 | 1766 | 144 | 191 |
-| 0.2940        | 2.6757 | 297  | 0.3404          | 0.8867   | 0.8388    | 0.9341 | 0.8839 | 1530 | 1616 | 294 | 108 |
-| 0.2891        | 2.9189 | 324  | 0.2897          | 0.9228   | 0.9231    | 0.9084 | 0.9157 | 1488 | 1786 | 124 | 150 |
-| 0.2921        | 3.1622 | 351  | 0.3451          | 0.8881   | 0.9042    | 0.8474 | 0.8749 | 1388 | 1763 | 147 | 250 |
-| 0.2958        | 3.4054 | 378  | 0.3503          | 0.8816   | 0.8604    | 0.8877 | 0.8738 | 1454 | 1674 | 236 | 184 |
-| 0.2673        | 3.6486 | 405  | 0.3124          | 0.9073   | 0.8834    | 0.9206 | 0.9016 | 1508 | 1711 | 199 | 130 |
-| 0.2755        | 3.8919 | 432  | 0.3070          | 0.9104   | 0.9099    | 0.8944 | 0.9021 | 1465 | 1765 | 145 | 173 |
-| 0.2913        | 4.1351 | 459  | 0.3094          | 0.9101   | 0.8914    | 0.9170 | 0.9040 | 1502 | 1727 | 183 | 136 |
-| 0.2720        | 4.3784 | 486  | 0.3083          | 0.8952   | 0.9283    | 0.8376 | 0.8806 | 1372 | 1804 | 106 | 266 |
-| 0.2929        | 4.6216 | 513  | 0.2726          | 0.9225   | 0.9565    | 0.8718 | 0.9122 | 1428 | 1845 | 65  | 210 |
-| 0.2599        | 4.8649 | 540  | 0.3409          | 0.8904   | 0.8746    | 0.8901 | 0.8823 | 1458 | 1701 | 209 | 180 |
+| Training Loss | Epoch  | Step | Validation Loss | Accuracy | Precision | Recall | F1     | Tp   | Tn   | Fp | Fn  |
+|:-------------:|:------:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:----:|:----:|:--:|:---:|
+| 0.4292        | 0.2477 | 110  | 0.1793          | 0.9464   | 0.9925    | 0.8907 | 0.9389 | 1459 | 1899 | 11 | 179 |
+| 0.2230        | 0.4955 | 220  | 0.1211          | 0.9651   | 0.9980    | 0.9261 | 0.9607 | 1517 | 1907 | 3  | 121 |
+| 0.1974        | 0.7432 | 330  | 0.1342          | 0.9690   | 0.9811    | 0.9512 | 0.9659 | 1558 | 1880 | 30 | 80  |
+| 0.1879        | 0.9910 | 440  | 0.1397          | 0.9628   | 0.9591    | 0.9603 | 0.9597 | 1573 | 1843 | 67 | 65  |
+| 0.1643        | 1.2387 | 550  | 0.1083          | 0.9741   | 0.9856    | 0.9579 | 0.9715 | 1569 | 1887 | 23 | 69  |
+| 0.1654        | 1.4865 | 660  | 0.0963          | 0.9715   | 0.9936    | 0.9444 | 0.9684 | 1547 | 1900 | 10 | 91  |
+| 0.1664        | 1.7342 | 770  | 0.1130          | 0.9693   | 0.9693    | 0.9640 | 0.9666 | 1579 | 1860 | 50 | 59  |
+| 0.1637        | 1.9820 | 880  | 0.0933          | 0.9746   | 0.9844    | 0.9603 | 0.9722 | 1573 | 1885 | 25 | 65  |
 
 
 ### Framework versions

config.json

@@ -1,33 +1,34 @@
 {
   "architectures": [
-    "MobileNetV2ForImageClassification"
+    "ViTForImageClassification"
   ],
-  "classifier_dropout_prob": 0.2,
-  "depth_divisible_by": 8,
-  "depth_multiplier": 1.0,
+  "attention_probs_dropout_prob": 0.0,
   "dtype": "float32",
-  "expand_ratio": 6,
-  "finegrained_output": true,
-  "first_layer_is_expansion": true,
-  "hidden_act": "relu6",
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 1024,
   "id2label": {
     "0": "0",
     "1": "1"
   },
   "image_size": 224,
   "initializer_range": 0.02,
+  "intermediate_size": 4096,
   "label2id": {
     "0": 0,
     "1": 1
   },
-  "layer_norm_eps": 0.001,
-  "min_depth": 8,
-  "model_type": "mobilenet_v2",
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 16,
   "num_channels": 3,
-  "output_stride": 32,
+  "num_hidden_layers": 24,
+  "patch_size": 16,
+  "pooler_act": "tanh",
+  "pooler_output_size": 1024,
   "problem_type": "single_label_classification",
-  "semantic_loss_ignore_index": 255,
-  "tf_padding": true,
+  "qkv_bias": true,
   "transformers_version": "5.0.0",
   "use_cache": false
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d6e9644d68cf097846ddac5c08ca36f23f8e7ad9770058bd4d332fc127e8bce2
-size 9080216
+oid sha256:779aeafe9a53b2a1b8a4b18e8a6a8bf637f7940dbd277f1883680150cb5fee1e
+size 1213261264

preprocessor_config.json

@@ -1,12 +1,6 @@
 {
-  "crop_size": {
-    "height": 224,
-    "width": 224
-  },
-  "data_format": "channels_first",
-  "do_center_crop": true,
+  "do_convert_rgb": null,
   "do_normalize": true,
-  "do_reduce_labels": false,
   "do_rescale": true,
   "do_resize": true,
   "image_mean": [
@@ -14,7 +8,7 @@
     0.5,
     0.5
   ],
-  "image_processor_type": "MobileNetV2ImageProcessorFast",
+  "image_processor_type": "ViTImageProcessor",
   "image_std": [
     0.5,
     0.5,
@@ -23,6 +17,7 @@
   "resample": 2,
   "rescale_factor": 0.00392156862745098,
   "size": {
-    "shortest_edge": 256
+    "height": 224,
+    "width": 224
   }
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9718d31a58ba43fb20b11cd760cd66a129ccd3519ce7e56f71efe4a0dad8c9a2
+oid sha256:a20650d184cd21a5ecf7b63db317856b5992087aa5fe9cbe6ebb9da5ccee247d
 size 5137