End of training

README.md

@@ -1,6 +1,6 @@
 ---
 license: apache-2.0
-base_model: google/vit-base-patch16-224-in21k
+base_model: microsoft/beit-base-patch16-224-pt22k-ft22k
 tags:
 - image-classification
 - vision
@@ -17,9 +17,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # donut-base-beans
 
-This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
+This model is a fine-tuned version of [microsoft/beit-base-patch16-224-pt22k-ft22k](https://huggingface.co/microsoft/beit-base-patch16-224-pt22k-ft22k) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0548
+- Loss: 0.0517
 - F1: 0.3303
 
 ## Model description
@@ -45,22 +45,17 @@ The following hyperparameters were used during training:
 - seed: 1337
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- num_epochs: 10.0
+- num_epochs: 5.0
 
 ### Training results
 
-| Training Loss | Epoch | Step  | Accuracy | Validation Loss |
-|:-------------:|:-----:|:-----:|:--------:|:---------------:|
-| 0.0769        | 1.0   | 6733  | 0.4973   | 0.0657          |
-| 0.0729        | 2.0   | 13466 | 0.4973   | 0.0613          |
-| 0.0564        | 3.0   | 20199 | 0.4973   | 0.0621          |
-| 0.0534        | 4.0   | 26932 | 0.4973   | 0.0572          |
-| 0.0511        | 5.0   | 33665 | 0.4973   | 0.0562          |
-| 0.0659        | 6.0   | 40398 | 0.4973   | 0.0577          |
-| 0.0623        | 7.0   | 47131 | 0.4973   | 0.0567          |
-| 0.0575        | 8.0   | 53864 | 0.4973   | 0.0555          |
-| 0.048         | 9.0   | 60597 | 0.4973   | 0.0548          |
-| 0.0389        | 10.0  | 67330 | 0.4973   | 0.0548          |
+| Training Loss | Epoch | Step  | F1  | Validation Loss |
+|:-------------:|:-----:|:-----:|:---:|:---------------:|
+| 0.0672        | 1.0   | 13466 | 0.0 | 0.0696          |
+| 0.0612        | 2.0   | 26932 | 0.0 | 0.0588          |
+| 0.0496        | 3.0   | 40398 | 0.0 | 0.0556          |
+| 0.0443        | 4.0   | 53864 | 0.0 | 0.0535          |
+| 0.0683        | 5.0   | 67330 | 0.0 | 0.0517          |
 
 
 ### Framework versions

all_results.json

@@ -1,14 +1,13 @@
 {
-    "epoch": 10.0,
-    "eval_accuracy": 0.49727767695099817,
+    "epoch": 5.0,
     "eval_f1": 0.33031292965957215,
-    "eval_loss": 0.05480470508337021,
-    "eval_runtime": 512.1974,
-    "eval_samples_per_second": 74.227,
-    "eval_steps_per_second": 4.641,
-    "total_flos": 1.6694802102985576e+20,
+    "eval_loss": 0.05167969688773155,
+    "eval_runtime": 516.9756,
+    "eval_samples_per_second": 73.541,
+    "eval_steps_per_second": 4.598,
+    "total_flos": 8.34383324330106e+19,
     "train_loss": 0.0,
-    "train_runtime": 0.0767,
-    "train_samples_per_second": 28073613.463,
-    "train_steps_per_second": 877361.502
+    "train_runtime": 0.0747,
+    "train_samples_per_second": 14413399.944,
+    "train_steps_per_second": 900900.217
 }

config.json

@@ -1,10 +1,15 @@
 {
-  "_name_or_path": "google/vit-base-patch16-224-in21k",
+  "_name_or_path": "microsoft/beit-base-patch16-224-pt22k-ft22k",
+  "add_fpn": false,
   "architectures": [
-    "ViTForImageClassification"
+    "BeitForImageClassification"
   ],
   "attention_probs_dropout_prob": 0.0,
-  "encoder_stride": 16,
+  "auxiliary_channels": 256,
+  "auxiliary_concat_input": false,
+  "auxiliary_loss_weight": 0.4,
+  "auxiliary_num_convs": 1,
+  "drop_path_rate": 0.1,
   "finetuning_task": "image-classification",
   "hidden_act": "gelu",
   "hidden_dropout_prob": 0.0,
@@ -19,13 +24,49 @@
     "Value(dtype='int64', id=None)": "0"
   },
   "layer_norm_eps": 1e-12,
-  "model_type": "vit",
+  "layer_scale_init_value": 0.1,
+  "model_type": "beit",
   "num_attention_heads": 12,
   "num_channels": 3,
   "num_hidden_layers": 12,
+  "out_features": [
+    "stage12"
+  ],
+  "out_indices": [
+    12
+  ],
   "patch_size": 16,
+  "pool_scales": [
+    1,
+    2,
+    3,
+    6
+  ],
   "problem_type": "regression",
-  "qkv_bias": true,
+  "reshape_hidden_states": true,
+  "semantic_loss_ignore_index": 255,
+  "stage_names": [
+    "stem",
+    "stage1",
+    "stage2",
+    "stage3",
+    "stage4",
+    "stage5",
+    "stage6",
+    "stage7",
+    "stage8",
+    "stage9",
+    "stage10",
+    "stage11",
+    "stage12"
+  ],
   "torch_dtype": "float32",
-  "transformers_version": "4.43.0.dev0"
+  "transformers_version": "4.43.0.dev0",
+  "use_absolute_position_embeddings": false,
+  "use_auxiliary_head": true,
+  "use_mask_token": false,
+  "use_mean_pooling": true,
+  "use_relative_position_bias": true,
+  "use_shared_relative_position_bias": false,
+  "vocab_size": 8192
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 10.0,
+    "epoch": 5.0,
     "eval_f1": 0.33031292965957215,
-    "eval_loss": 0.05480470508337021,
-    "eval_runtime": 512.1974,
-    "eval_samples_per_second": 74.227,
-    "eval_steps_per_second": 4.641
+    "eval_loss": 0.05167969688773155,
+    "eval_runtime": 516.9756,
+    "eval_samples_per_second": 73.541,
+    "eval_steps_per_second": 4.598
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2561cfd7419a353fba1d1368529d633d6fb11d901bddcfe413ee3478cec81425
-size 343220892
+oid sha256:87a3e53f87c26251c9485be8d2e6b4896fb3ca5fbc1fe742353ffe4c57e091e8
+size 343077252

preprocessor_config.json

@@ -1,5 +1,11 @@
 {
+  "crop_size": {
+    "height": 224,
+    "width": 224
+  },
+  "do_center_crop": false,
   "do_normalize": true,
+  "do_reduce_labels": false,
   "do_rescale": true,
   "do_resize": true,
   "image_mean": [
@@ -7,7 +13,7 @@
     0.5,
     0.5
   ],
-  "image_processor_type": "ViTImageProcessor",
+  "image_processor_type": "BeitImageProcessor",
   "image_std": [
     0.5,
     0.5,

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 10.0,
-    "total_flos": 1.6694802102985576e+20,
+    "epoch": 5.0,
+    "total_flos": 8.34383324330106e+19,
     "train_loss": 0.0,
-    "train_runtime": 0.0767,
-    "train_samples_per_second": 28073613.463,
-    "train_steps_per_second": 877361.502
+    "train_runtime": 0.0747,
+    "train_samples_per_second": 14413399.944,
+    "train_steps_per_second": 900900.217
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7d8a6f533c72550417f1e253380ae12ad72153532482192012f764515f190bbf
+oid sha256:a0a2d72729478c9c56a6673fbc1a93a59085bf2829c446ff056e0e39458bcde9
 size 5304