Training in progress, epoch 0

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 19.73,
-    "total_flos": 1.0467617957139087e+18,
-    "train_loss": 0.13025596663355826,
-    "train_runtime": 778.539,
-    "train_samples_per_second": 17.571,
-    "train_steps_per_second": 0.051
+    "total_flos": 4.050283110270566e+16,
+    "train_loss": 0.24904835671186448,
+    "train_runtime": 1252.5827,
+    "train_samples_per_second": 10.921,
+    "train_steps_per_second": 0.032
 }

config.json

@@ -1,79 +1,56 @@
 {
-  "_name_or_path": "nvidia/mit-b0",
+  "_name_or_path": "microsoft/beit-base-patch16-224",
   "architectures": [
-    "SegformerForImageClassification"
+    "BeitForImageClassification"
   ],
   "attention_probs_dropout_prob": 0.0,
-  "classifier_dropout_prob": 0.1,
-  "decoder_hidden_size": 256,
-  "depths": [
-    2,
-    2,
-    2,
-    2
-  ],
-  "downsampling_rates": [
-    1,
-    4,
-    8,
-    16
-  ],
+  "auxiliary_channels": 256,
+  "auxiliary_concat_input": false,
+  "auxiliary_loss_weight": 0.4,
+  "auxiliary_num_convs": 1,
   "drop_path_rate": 0.1,
   "hidden_act": "gelu",
   "hidden_dropout_prob": 0.0,
-  "hidden_sizes": [
-    32,
-    64,
-    160,
-    256
-  ],
+  "hidden_size": 768,
   "id2label": {
     "0": "0",
     "1": "1"
   },
   "image_size": 224,
   "initializer_range": 0.02,
+  "intermediate_size": 3072,
   "label2id": {
     "0": "0",
     "1": "1"
   },
-  "layer_norm_eps": 1e-06,
-  "mlp_ratios": [
-    4,
-    4,
-    4,
-    4
+  "layer_norm_eps": 1e-12,
+  "layer_scale_init_value": 0.1,
+  "model_type": "beit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "out_indices": [
+    3,
+    5,
+    7,
+    11
   ],
-  "model_type": "segformer",
-  "num_attention_heads": [
+  "patch_size": 16,
+  "pool_scales": [
     1,
     2,
-    5,
-    8
-  ],
-  "num_channels": 3,
-  "num_encoder_blocks": 4,
-  "patch_sizes": [
-    7,
-    3,
     3,
-    3
+    6
   ],
   "problem_type": "single_label_classification",
-  "reshape_last_stage": true,
   "semantic_loss_ignore_index": 255,
-  "sr_ratios": [
-    8,
-    4,
-    2,
-    1
-  ],
-  "strides": [
-    4,
-    2,
-    2,
-    2
-  ],
   "torch_dtype": "float32",
-  "transformers_version": "4.24.0"
+  "transformers_version": "4.24.0",
+  "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
 }

preprocessor_config.json

@@ -1,18 +1,20 @@
 {
+  "crop_size": 224,
+  "do_center_crop": false,
   "do_normalize": true,
   "do_resize": true,
-  "feature_extractor_type": "SegformerFeatureExtractor",
+  "feature_extractor_type": "BeitFeatureExtractor",
   "image_mean": [
-    0.485,
-    0.456,
-    0.406
+    0.5,
+    0.5,
+    0.5
   ],
   "image_std": [
-    0.229,
-    0.224,
-    0.225
+    0.5,
+    0.5,
+    0.5
   ],
   "reduce_labels": false,
   "resample": 2,
-  "size": 512
+  "size": 224
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d798b3276d97dca241a8435aa05ecb4faa723183f363703fe99a3e247eeb25b4
-size 13346069
+oid sha256:151fabe70f8d1fb3f4234056b93b9d8dcffdf702ed3d9166b4fb9a75df1e8431
+size 346860409

runs/Jul10_18-44-31_teesta/events.out.tfevents.1688994879.teesta.6977.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:323c36aa6335515b2327f166b2a35a7bf4fcd58c4dfef2bc631e71821e4fe9dd
-size 13981
+oid sha256:34b139f5b3ab1d9f23ce6a990ec46edb0abc279d1630053d8ed01d11c1e0e2b6
+size 14329

runs/Jul10_19-19-56_teesta/1688997005.952017/events.out.tfevents.1688997005.teesta.35180.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a8c154c5e07a7cb89112973977c8483d9bccfe264b9391c4fde05f9724c0c7b
+size 5427

runs/Jul10_19-19-56_teesta/events.out.tfevents.1688997005.teesta.35180.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:252cfbf3c917b078cc875d1550fd22af76112b90e56ccf28cbe7e5eae772290f
+size 4659

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 19.73,
-    "total_flos": 1.0467617957139087e+18,
-    "train_loss": 0.13025596663355826,
-    "train_runtime": 778.539,
-    "train_samples_per_second": 17.571,
-    "train_steps_per_second": 0.051
+    "total_flos": 4.050283110270566e+16,
+    "train_loss": 0.24904835671186448,
+    "train_runtime": 1252.5827,
+    "train_samples_per_second": 10.921,
+    "train_steps_per_second": 0.032
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_metric": 0.9444444444444444,
-  "best_model_checkpoint": "test/checkpoint-18",
+  "best_model_checkpoint": "test/checkpoint-14",
   "epoch": 19.727272727272727,
   "global_step": 40,
   "is_hyper_param_search": false,
@@ -9,281 +9,281 @@
   "log_history": [
     {
       "epoch": 0.73,
-      "eval_accuracy": 0.6388888888888888,
-      "eval_f1": 0.6513851917578626,
-      "eval_loss": 0.6436764597892761,
-      "eval_precision": 0.6720549981419547,
-      "eval_recall": 0.6388888888888888,
-      "eval_runtime": 6.0758,
-      "eval_samples_per_second": 5.925,
-      "eval_steps_per_second": 0.165,
+      "eval_accuracy": 0.6111111111111112,
+      "eval_f1": 0.5416666666666666,
+      "eval_loss": 0.7271409630775452,
+      "eval_precision": 0.5208333333333334,
+      "eval_recall": 0.6111111111111112,
+      "eval_runtime": 6.3413,
+      "eval_samples_per_second": 5.677,
+      "eval_steps_per_second": 0.158,
       "step": 2
     },
     {
       "epoch": 1.73,
-      "eval_accuracy": 0.8055555555555556,
-      "eval_f1": 0.7647399681297986,
-      "eval_loss": 0.3604269325733185,
-      "eval_precision": 0.8468013468013468,
-      "eval_recall": 0.8055555555555556,
-      "eval_runtime": 5.7608,
-      "eval_samples_per_second": 6.249,
-      "eval_steps_per_second": 0.174,
+      "eval_accuracy": 0.8333333333333334,
+      "eval_f1": 0.8148148148148149,
+      "eval_loss": 0.46258121728897095,
+      "eval_precision": 0.8666666666666667,
+      "eval_recall": 0.8333333333333334,
+      "eval_runtime": 5.8395,
+      "eval_samples_per_second": 6.165,
+      "eval_steps_per_second": 0.171,
       "step": 4
     },
     {
       "epoch": 2.73,
-      "eval_accuracy": 0.8888888888888888,
-      "eval_f1": 0.8888888888888888,
-      "eval_loss": 0.22702129185199738,
-      "eval_precision": 0.8888888888888888,
-      "eval_recall": 0.8888888888888888,
-      "eval_runtime": 4.8725,
-      "eval_samples_per_second": 7.388,
-      "eval_steps_per_second": 0.205,
+      "eval_accuracy": 0.9166666666666666,
+      "eval_f1": 0.9157054125998224,
+      "eval_loss": 0.3036719858646393,
+      "eval_precision": 0.9163636363636365,
+      "eval_recall": 0.9166666666666666,
+      "eval_runtime": 5.8411,
+      "eval_samples_per_second": 6.163,
+      "eval_steps_per_second": 0.171,
       "step": 6
     },
     {
       "epoch": 3.73,
-      "eval_accuracy": 0.8888888888888888,
-      "eval_f1": 0.8888888888888888,
-      "eval_loss": 0.16791754961013794,
-      "eval_precision": 0.8888888888888888,
-      "eval_recall": 0.8888888888888888,
-      "eval_runtime": 6.131,
-      "eval_samples_per_second": 5.872,
-      "eval_steps_per_second": 0.163,
+      "eval_accuracy": 0.9166666666666666,
+      "eval_f1": 0.9131652661064424,
+      "eval_loss": 0.24792712926864624,
+      "eval_precision": 0.9259259259259258,
+      "eval_recall": 0.9166666666666666,
+      "eval_runtime": 5.7041,
+      "eval_samples_per_second": 6.311,
+      "eval_steps_per_second": 0.175,
       "step": 8
     },
     {
       "epoch": 4.73,
       "learning_rate": 0.0001,
-      "loss": 0.4939,
+      "loss": 0.6338,
       "step": 10
     },
     {
       "epoch": 4.73,
-      "eval_accuracy": 0.8888888888888888,
-      "eval_f1": 0.8888888888888888,
-      "eval_loss": 0.15902230143547058,
-      "eval_precision": 0.8888888888888888,
-      "eval_recall": 0.8888888888888888,
-      "eval_runtime": 4.9896,
-      "eval_samples_per_second": 7.215,
-      "eval_steps_per_second": 0.2,
+      "eval_accuracy": 0.9166666666666666,
+      "eval_f1": 0.9174468085106382,
+      "eval_loss": 0.23573820292949677,
+      "eval_precision": 0.919732441471572,
+      "eval_recall": 0.9166666666666666,
+      "eval_runtime": 5.6473,
+      "eval_samples_per_second": 6.375,
+      "eval_steps_per_second": 0.177,
       "step": 10
     },
     {
       "epoch": 5.73,
-      "eval_accuracy": 0.9166666666666666,
-      "eval_f1": 0.9152598477325389,
-      "eval_loss": 0.15024107694625854,
-      "eval_precision": 0.9156378600823045,
-      "eval_recall": 0.9166666666666666,
-      "eval_runtime": 4.9812,
-      "eval_samples_per_second": 7.227,
-      "eval_steps_per_second": 0.201,
+      "eval_accuracy": 0.8888888888888888,
+      "eval_f1": 0.890746934225195,
+      "eval_loss": 0.21319927275180817,
+      "eval_precision": 0.8982683982683982,
+      "eval_recall": 0.8888888888888888,
+      "eval_runtime": 6.2525,
+      "eval_samples_per_second": 5.758,
+      "eval_steps_per_second": 0.16,
       "step": 12
     },
     {
       "epoch": 6.73,
-      "eval_accuracy": 0.9166666666666666,
-      "eval_f1": 0.9152598477325389,
-      "eval_loss": 0.13303342461585999,
-      "eval_precision": 0.9156378600823045,
-      "eval_recall": 0.9166666666666666,
-      "eval_runtime": 4.8591,
-      "eval_samples_per_second": 7.409,
-      "eval_steps_per_second": 0.206,
+      "eval_accuracy": 0.9444444444444444,
+      "eval_f1": 0.9430303030303031,
+      "eval_loss": 0.17125670611858368,
+      "eval_precision": 0.9487179487179487,
+      "eval_recall": 0.9444444444444444,
+      "eval_runtime": 5.7878,
+      "eval_samples_per_second": 6.22,
+      "eval_steps_per_second": 0.173,
       "step": 14
     },
     {
       "epoch": 7.73,
-      "eval_accuracy": 0.8888888888888888,
-      "eval_f1": 0.8888888888888888,
-      "eval_loss": 0.12413659691810608,
-      "eval_precision": 0.8888888888888888,
-      "eval_recall": 0.8888888888888888,
-      "eval_runtime": 5.4858,
-      "eval_samples_per_second": 6.562,
-      "eval_steps_per_second": 0.182,
+      "eval_accuracy": 0.9444444444444444,
+      "eval_f1": 0.9444444444444444,
+      "eval_loss": 0.15498512983322144,
+      "eval_precision": 0.9444444444444444,
+      "eval_recall": 0.9444444444444444,
+      "eval_runtime": 6.2641,
+      "eval_samples_per_second": 5.747,
+      "eval_steps_per_second": 0.16,
       "step": 16
     },
     {
       "epoch": 8.73,
       "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.0989239364862442,
-      "eval_precision": 0.9484126984126983,
+      "eval_f1": 0.9444444444444444,
+      "eval_loss": 0.14064311981201172,
+      "eval_precision": 0.9444444444444444,
       "eval_recall": 0.9444444444444444,
-      "eval_runtime": 5.6045,
-      "eval_samples_per_second": 6.423,
-      "eval_steps_per_second": 0.178,
+      "eval_runtime": 5.34,
+      "eval_samples_per_second": 6.742,
+      "eval_steps_per_second": 0.187,
       "step": 18
     },
     {
       "epoch": 9.73,
       "learning_rate": 6.666666666666667e-05,
-      "loss": 0.0234,
+      "loss": 0.217,
       "step": 20
     },
     {
       "epoch": 9.73,
-      "eval_accuracy": 0.9166666666666666,
-      "eval_f1": 0.9152598477325389,
-      "eval_loss": 0.10008219629526138,
-      "eval_precision": 0.9156378600823045,
-      "eval_recall": 0.9166666666666666,
-      "eval_runtime": 6.5249,
-      "eval_samples_per_second": 5.517,
-      "eval_steps_per_second": 0.153,
+      "eval_accuracy": 0.9444444444444444,
+      "eval_f1": 0.9444444444444444,
+      "eval_loss": 0.16372796893119812,
+      "eval_precision": 0.9444444444444444,
+      "eval_recall": 0.9444444444444444,
+      "eval_runtime": 5.2833,
+      "eval_samples_per_second": 6.814,
+      "eval_steps_per_second": 0.189,
       "step": 20
     },
     {
       "epoch": 10.73,
-      "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.09186337888240814,
-      "eval_precision": 0.9484126984126983,
-      "eval_recall": 0.9444444444444444,
-      "eval_runtime": 5.01,
-      "eval_samples_per_second": 7.186,
-      "eval_steps_per_second": 0.2,
+      "eval_accuracy": 0.9166666666666666,
+      "eval_f1": 0.9157054125998224,
+      "eval_loss": 0.16719864308834076,
+      "eval_precision": 0.9163636363636365,
+      "eval_recall": 0.9166666666666666,
+      "eval_runtime": 6.4055,
+      "eval_samples_per_second": 5.62,
+      "eval_steps_per_second": 0.156,
       "step": 22
     },
     {
       "epoch": 11.73,
       "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.09773090481758118,
-      "eval_precision": 0.9484126984126983,
+      "eval_f1": 0.9444444444444444,
+      "eval_loss": 0.13150224089622498,
+      "eval_precision": 0.9444444444444444,
       "eval_recall": 0.9444444444444444,
-      "eval_runtime": 5.7387,
-      "eval_samples_per_second": 6.273,
-      "eval_steps_per_second": 0.174,
+      "eval_runtime": 5.1218,
+      "eval_samples_per_second": 7.029,
+      "eval_steps_per_second": 0.195,
       "step": 24
     },
     {
       "epoch": 12.73,
       "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.10761526226997375,
-      "eval_precision": 0.9484126984126983,
+      "eval_f1": 0.9444444444444444,
+      "eval_loss": 0.1597292125225067,
+      "eval_precision": 0.9444444444444444,
       "eval_recall": 0.9444444444444444,
-      "eval_runtime": 5.1262,
-      "eval_samples_per_second": 7.023,
+      "eval_runtime": 5.1268,
+      "eval_samples_per_second": 7.022,
       "eval_steps_per_second": 0.195,
       "step": 26
     },
     {
       "epoch": 13.73,
       "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.09590016305446625,
-      "eval_precision": 0.9484126984126983,
+      "eval_f1": 0.9444444444444444,
+      "eval_loss": 0.1412082463502884,
+      "eval_precision": 0.9444444444444444,
       "eval_recall": 0.9444444444444444,
-      "eval_runtime": 5.0649,
-      "eval_samples_per_second": 7.108,
-      "eval_steps_per_second": 0.197,
+      "eval_runtime": 5.6714,
+      "eval_samples_per_second": 6.348,
+      "eval_steps_per_second": 0.176,
       "step": 28
     },
     {
       "epoch": 14.73,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.0024,
+      "loss": 0.0906,
       "step": 30
     },
     {
       "epoch": 14.73,
       "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.11737969517707825,
-      "eval_precision": 0.9484126984126983,
+      "eval_f1": 0.9444444444444444,
+      "eval_loss": 0.1776580661535263,
+      "eval_precision": 0.9444444444444444,
       "eval_recall": 0.9444444444444444,
-      "eval_runtime": 10.5713,
-      "eval_samples_per_second": 3.405,
-      "eval_steps_per_second": 0.095,
+      "eval_runtime": 5.8702,
+      "eval_samples_per_second": 6.133,
+      "eval_steps_per_second": 0.17,
       "step": 30
     },
     {
       "epoch": 15.73,
       "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.10555433481931686,
-      "eval_precision": 0.9484126984126983,
+      "eval_f1": 0.9444444444444444,
+      "eval_loss": 0.14607585966587067,
+      "eval_precision": 0.9444444444444444,
       "eval_recall": 0.9444444444444444,
-      "eval_runtime": 5.3198,
-      "eval_samples_per_second": 6.767,
-      "eval_steps_per_second": 0.188,
+      "eval_runtime": 7.113,
+      "eval_samples_per_second": 5.061,
+      "eval_steps_per_second": 0.141,
       "step": 32
     },
     {
       "epoch": 16.73,
-      "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.1085939034819603,
-      "eval_precision": 0.9484126984126983,
-      "eval_recall": 0.9444444444444444,
-      "eval_runtime": 5.2418,
-      "eval_samples_per_second": 6.868,
-      "eval_steps_per_second": 0.191,
+      "eval_accuracy": 0.9166666666666666,
+      "eval_f1": 0.9174468085106382,
+      "eval_loss": 0.19147923588752747,
+      "eval_precision": 0.919732441471572,
+      "eval_recall": 0.9166666666666666,
+      "eval_runtime": 5.3662,
+      "eval_samples_per_second": 6.709,
+      "eval_steps_per_second": 0.186,
       "step": 34
     },
     {
       "epoch": 17.73,
-      "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.11273058503866196,
-      "eval_precision": 0.9484126984126983,
-      "eval_recall": 0.9444444444444444,
-      "eval_runtime": 6.5423,
-      "eval_samples_per_second": 5.503,
-      "eval_steps_per_second": 0.153,
+      "eval_accuracy": 0.9166666666666666,
+      "eval_f1": 0.9174468085106382,
+      "eval_loss": 0.13753658533096313,
+      "eval_precision": 0.919732441471572,
+      "eval_recall": 0.9166666666666666,
+      "eval_runtime": 5.3487,
+      "eval_samples_per_second": 6.731,
+      "eval_steps_per_second": 0.187,
       "step": 36
     },
     {
       "epoch": 18.73,
-      "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.1172882616519928,
-      "eval_precision": 0.9484126984126983,
-      "eval_recall": 0.9444444444444444,
-      "eval_runtime": 5.9599,
-      "eval_samples_per_second": 6.04,
-      "eval_steps_per_second": 0.168,
+      "eval_accuracy": 0.9166666666666666,
+      "eval_f1": 0.9174468085106382,
+      "eval_loss": 0.1448293775320053,
+      "eval_precision": 0.919732441471572,
+      "eval_recall": 0.9166666666666666,
+      "eval_runtime": 5.5619,
+      "eval_samples_per_second": 6.473,
+      "eval_steps_per_second": 0.18,
       "step": 38
     },
     {
       "epoch": 19.73,
       "learning_rate": 0.0,
-      "loss": 0.0013,
+      "loss": 0.0548,
       "step": 40
     },
     {
       "epoch": 19.73,
       "eval_accuracy": 0.9444444444444444,
-      "eval_f1": 0.9423868312757201,
-      "eval_loss": 0.11149459332227707,
-      "eval_precision": 0.9484126984126983,
+      "eval_f1": 0.9444444444444444,
+      "eval_loss": 0.12921589612960815,
+      "eval_precision": 0.9444444444444444,
       "eval_recall": 0.9444444444444444,
-      "eval_runtime": 5.9908,
-      "eval_samples_per_second": 6.009,
-      "eval_steps_per_second": 0.167,
+      "eval_runtime": 5.8187,
+      "eval_samples_per_second": 6.187,
+      "eval_steps_per_second": 0.172,
       "step": 40
     },
     {
       "epoch": 19.73,
       "step": 40,
-      "total_flos": 1.0467617957139087e+18,
-      "train_loss": 0.13025596663355826,
-      "train_runtime": 778.539,
-      "train_samples_per_second": 17.571,
-      "train_steps_per_second": 0.051
+      "total_flos": 4.050283110270566e+16,
+      "train_loss": 0.24904835671186448,
+      "train_runtime": 1252.5827,
+      "train_samples_per_second": 10.921,
+      "train_steps_per_second": 0.032
     }
   ],
   "max_steps": 40,
   "num_train_epochs": 20,
-  "total_flos": 1.0467617957139087e+18,
+  "total_flos": 4.050283110270566e+16,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:49aab5e5341ad6773a4a7b366c080f1cd6f47aa4da51bb4235c3e6f7333bdf36
+oid sha256:6e79370bcf724d04d0deeebfd7cb730102e73af0e10cf634a2778707017455f8
 size 3387