Update IR model to reshapeable version.

README.md

@@ -17,6 +17,8 @@ model-index:
     dataset:
       name: GLUE SST2
       type: glue
+      config: sst2
+      split: validation
       args: sst2
     metrics:
     - name: Accuracy
@@ -30,7 +32,7 @@ should probably proofread and complete it, then remove this comment. -->
 # jpqd-bert-base-ft-sst2
 
 > **Note**
-> This model was trained for only 1 epoch and is shared for testing purposes
+> This model was trained for only 1 epoch and is shared for testing purposes.
 
 This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the GLUE SST2 dataset.
 It achieves the following results on the evaluation set:
@@ -79,7 +81,7 @@ The following hyperparameters were used during training:
 
 ### Framework versions
 
-- Transformers 4.25.1
+- Transformers 4.26.1
 - Pytorch 1.13.1+cu117
 - Datasets 2.8.0
 - Tokenizers 0.13.2

all_results.json

@@ -2,13 +2,13 @@
     "epoch": 1.0,
     "eval_accuracy": 0.9254587155963303,
     "eval_loss": 0.21810248494148254,
-    "eval_runtime": 22.1504,
+    "eval_runtime": 22.3128,
     "eval_samples": 872,
-    "eval_samples_per_second": 39.367,
-    "eval_steps_per_second": 4.921,
+    "eval_samples_per_second": 39.081,
+    "eval_steps_per_second": 4.885,
     "train_loss": 0.40093172477146793,
-    "train_runtime": 1341.7612,
+    "train_runtime": 1354.5918,
     "train_samples": 67349,
-    "train_samples_per_second": 50.194,
-    "train_steps_per_second": 1.569
+    "train_samples_per_second": 49.719,
+    "train_steps_per_second": 1.554
 }

config.json

@@ -29,7 +29,7 @@
   "position_embedding_type": "absolute",
   "problem_type": "single_label_classification",
   "torch_dtype": "float32",
-  "transformers_version": "4.25.1",
+  "transformers_version": "4.26.1",
   "type_vocab_size": 2,
   "use_cache": true,
   "vocab_size": 30522

eval_results.json

@@ -2,8 +2,8 @@
     "epoch": 1.0,
     "eval_accuracy": 0.9254587155963303,
     "eval_loss": 0.21810248494148254,
-    "eval_runtime": 22.1504,
+    "eval_runtime": 22.3128,
     "eval_samples": 872,
-    "eval_samples_per_second": 39.367,
-    "eval_steps_per_second": 4.921
+    "eval_samples_per_second": 39.081,
+    "eval_steps_per_second": 4.885
 }

model.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:74220f4ab04e8b2cf0a8eb0ac647ab95ad8e9cf3abe9af47aee0f4d2b45dab3b
-size 439470129

nncf_output.log

@@ -75,6 +75,8 @@ INFO:nncf:Not adding activation input quantizer for operation: 352 BertForSequen
 INFO:nncf:Not adding activation input quantizer for operation: 357 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/__add___0
 INFO:nncf:Not adding activation input quantizer for operation: 358 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
 INFO:nncf:Collecting tensor statistics |████████████████| 1 / 1
+INFO:nncf:Compiling and loading torch extension: quantized_functions_cuda...
+INFO:nncf:Finished loading torch extension: quantized_functions_cuda
 INFO:nncf:BatchNorm statistics adaptation |██              | 1 / 7
 INFO:nncf:BatchNorm statistics adaptation |████            | 2 / 7
 INFO:nncf:BatchNorm statistics adaptation |██████          | 3 / 7
@@ -1547,4 +1549,3 @@ Epoch 0 |+==============+=====================+====================+============
 Epoch 0 || 8            | 100.00 % (77 / 77)  | 100.00 % (101 /    | 100.00 % (178 /    |
 Epoch 0 ||              |                     | 101)               | 178)               |
 Epoch 0 |+--------------+---------------------+--------------------+--------------------+
-INFO:nncf:Movement sparsity scheduler updates importance threshold and regularizationfactor per optimizer step, but steps_per_epoch was not set in config. Will measure the actual steps per epoch as signaled by a .epoch_step() call.

openvino_config.json

@@ -94,8 +94,8 @@
       "type": "long"
     }
   ],
-  "log_dir": "/tmp/jpqd-bert-base-ft-sst2",
-  "optimum_version": "1.6.1",
+  "log_dir": "jpqd-bert-base-ft-sst2",
+  "optimum_version": "1.6.4",
   "save_onnx_model": false,
-  "transformers_version": "4.25.1"
+  "transformers_version": "4.26.1"
 }

openvino_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e9cfad0e7bc7bf2932066a8a392d1315f9bbbaf4db18504ce184266bbdf288c8
-size 110187016
+oid sha256:7383489839e89fe2ad7daf09beed249bfc2dcb35476cf17776c9c2c95fdf9ec2
+size 438609916

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:99590d19bdea8d7f6dbf5a37e3947a09af3ef5805eb39e5238e5afa19452aeed
+oid sha256:fa61dca8594fa489f7d978ab5fd49ec7dc82397cfe93d54f751384de9e85dc2a
 size 779394143

tokenizer.json

@@ -1,21 +1,7 @@
 {
   "version": "1.0",
-  "truncation": {
-    "direction": "Right",
-    "max_length": 128,
-    "strategy": "LongestFirst",
-    "stride": 0
-  },
-  "padding": {
-    "strategy": {
-      "Fixed": 128
-    },
-    "direction": "Right",
-    "pad_to_multiple_of": null,
-    "pad_id": 0,
-    "pad_type_id": 0,
-    "pad_token": "[PAD]"
-  },
+  "truncation": null,
+  "padding": null,
   "added_tokens": [
     {
       "id": 0,

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
     "train_loss": 0.40093172477146793,
-    "train_runtime": 1341.7612,
+    "train_runtime": 1354.5918,
     "train_samples": 67349,
-    "train_samples_per_second": 50.194,
-    "train_steps_per_second": 1.569
+    "train_samples_per_second": 49.719,
+    "train_steps_per_second": 1.554
 }

trainer_state.json

@@ -336,9 +336,9 @@
       "epoch": 0.12,
       "eval_accuracy": 0.8761467889908257,
       "eval_loss": 0.44158032536506653,
-      "eval_runtime": 21.7125,
-      "eval_samples_per_second": 40.161,
-      "eval_steps_per_second": 5.02,
+      "eval_runtime": 21.9879,
+      "eval_samples_per_second": 39.658,
+      "eval_steps_per_second": 4.957,
       "step": 250
     },
     {
@@ -670,9 +670,9 @@
       "epoch": 0.24,
       "eval_accuracy": 0.8899082568807339,
       "eval_loss": 0.49690014123916626,
-      "eval_runtime": 21.7581,
-      "eval_samples_per_second": 40.077,
-      "eval_steps_per_second": 5.01,
+      "eval_runtime": 22.004,
+      "eval_samples_per_second": 39.629,
+      "eval_steps_per_second": 4.954,
       "step": 500
     },
     {
@@ -1004,9 +1004,9 @@
       "epoch": 0.36,
       "eval_accuracy": 0.9162844036697247,
       "eval_loss": 0.2716875970363617,
-      "eval_runtime": 21.7417,
-      "eval_samples_per_second": 40.107,
-      "eval_steps_per_second": 5.013,
+      "eval_runtime": 21.9554,
+      "eval_samples_per_second": 39.717,
+      "eval_steps_per_second": 4.965,
       "step": 750
     },
     {
@@ -1338,9 +1338,9 @@
       "epoch": 0.48,
       "eval_accuracy": 0.911697247706422,
       "eval_loss": 0.24315589666366577,
-      "eval_runtime": 21.9413,
-      "eval_samples_per_second": 39.742,
-      "eval_steps_per_second": 4.968,
+      "eval_runtime": 22.0177,
+      "eval_samples_per_second": 39.604,
+      "eval_steps_per_second": 4.951,
       "step": 1000
     },
     {
@@ -1672,9 +1672,9 @@
       "epoch": 0.59,
       "eval_accuracy": 0.9243119266055045,
       "eval_loss": 0.20327819883823395,
-      "eval_runtime": 21.7723,
-      "eval_samples_per_second": 40.051,
-      "eval_steps_per_second": 5.006,
+      "eval_runtime": 22.009,
+      "eval_samples_per_second": 39.62,
+      "eval_steps_per_second": 4.953,
       "step": 1250
     },
     {
@@ -2006,9 +2006,9 @@
       "epoch": 0.71,
       "eval_accuracy": 0.9243119266055045,
       "eval_loss": 0.2382841557264328,
-      "eval_runtime": 21.7644,
-      "eval_samples_per_second": 40.065,
-      "eval_steps_per_second": 5.008,
+      "eval_runtime": 22.0278,
+      "eval_samples_per_second": 39.586,
+      "eval_steps_per_second": 4.948,
       "step": 1500
     },
     {
@@ -2340,9 +2340,9 @@
       "epoch": 0.83,
       "eval_accuracy": 0.9254587155963303,
       "eval_loss": 0.22330859303474426,
-      "eval_runtime": 21.7241,
-      "eval_samples_per_second": 40.14,
-      "eval_steps_per_second": 5.017,
+      "eval_runtime": 21.9928,
+      "eval_samples_per_second": 39.649,
+      "eval_steps_per_second": 4.956,
       "step": 1750
     },
     {
@@ -2674,9 +2674,9 @@
       "epoch": 0.95,
       "eval_accuracy": 0.9254587155963303,
       "eval_loss": 0.22065171599388123,
-      "eval_runtime": 21.7453,
-      "eval_samples_per_second": 40.101,
-      "eval_steps_per_second": 5.013,
+      "eval_runtime": 21.9957,
+      "eval_samples_per_second": 39.644,
+      "eval_steps_per_second": 4.956,
       "step": 2000
     },
     {
@@ -2814,9 +2814,9 @@
       "step": 2105,
       "total_flos": 4441630972486656.0,
       "train_loss": 0.40093172477146793,
-      "train_runtime": 1341.7612,
-      "train_samples_per_second": 50.194,
-      "train_steps_per_second": 1.569
+      "train_runtime": 1354.5918,
+      "train_samples_per_second": 49.719,
+      "train_steps_per_second": 1.554
     }
   ],
   "max_steps": 2105,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3fbb005c34586cdf1d9dfa3fa4f3e0e2b7209a1984b065821a0ac11dbbf8e40b
-size 3643
+oid sha256:512755fad1ec8d16b5f8c62bc77ce4f3655b8f55c801d2452ac80202929c67a1
+size 3579