Hugging Face's logo

jysh1023
/
bert_uncased_QAT

Text Classification
Transformers
PyTorch
TensorBoard
OpenVINO
bert
Generated from Trainer
Eval Results
Model card Files
xet
Metrics Community
1
jysh1023 commited on
Commit
8be076d
·
1 Parent(s): 5b6efc8

End of training

Browse files
Files changed (11) hide show
  1. README.md +5 -11
  2. compressed_graph.dot +0 -0
  3. logs/events.out.tfevents.1700312106.0a848c80699a.1576.0 +3 -0
  4. logs/events.out.tfevents.1700312559.0a848c80699a.1576.1 +3 -0
  5. nncf_output.log +188 -0
  6. openvino_config.json +60 -0
  7. openvino_model.bin +3 -0
  8. openvino_model.xml +0 -0
  9. original_graph.dot +0 -0
  10. pytorch_model.bin +2 -2
  11. training_args.bin +1 -1
README.md CHANGED
@@ -22,7 +22,7 @@ model-index:
22
  metrics:
23
  - name: Accuracy
24
  type: accuracy
25
- value: 0.9151376146788991
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -32,8 +32,8 @@ should probably proofread and complete it, then remove this comment. -->
32
 
33
  This model is a fine-tuned version of [google/bert_uncased_L-6_H-768_A-12](https://huggingface.co/google/bert_uncased_L-6_H-768_A-12) on the glue dataset.
34
  It achieves the following results on the evaluation set:
35
- - Loss: 0.3683
36
- - Accuracy: 0.9151
37
 
38
  ## Model description
39
 
@@ -58,20 +58,14 @@ The following hyperparameters were used during training:
58
  - seed: 33
59
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
60
  - lr_scheduler_type: linear
61
- - num_epochs: 7
62
  - mixed_precision_training: Native AMP
63
 
64
  ### Training results
65
 
66
  | Training Loss | Epoch | Step | Validation Loss | Accuracy |
67
  |:-------------:|:-----:|:----:|:---------------:|:--------:|
68
- | 0.2388 | 1.0 | 527 | 0.2575 | 0.9060 |
69
- | 0.1196 | 2.0 | 1054 | 0.2980 | 0.9002 |
70
- | 0.0789 | 3.0 | 1581 | 0.2825 | 0.9071 |
71
- | 0.0529 | 4.0 | 2108 | 0.3194 | 0.9071 |
72
- | 0.0364 | 5.0 | 2635 | 0.3683 | 0.9151 |
73
- | 0.0236 | 6.0 | 3162 | 0.4103 | 0.9094 |
74
- | 0.0154 | 7.0 | 3689 | 0.4751 | 0.9083 |
75
 
76
 
77
  ### Framework versions
 
22
  metrics:
23
  - name: Accuracy
24
  type: accuracy
25
+ value: 0.908256880733945
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
32
 
33
  This model is a fine-tuned version of [google/bert_uncased_L-6_H-768_A-12](https://huggingface.co/google/bert_uncased_L-6_H-768_A-12) on the glue dataset.
34
  It achieves the following results on the evaluation set:
35
+ - Loss: 0.2574
36
+ - Accuracy: 0.9083
37
 
38
  ## Model description
39
 
 
58
  - seed: 33
59
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
60
  - lr_scheduler_type: linear
61
+ - num_epochs: 1.0
62
  - mixed_precision_training: Native AMP
63
 
64
  ### Training results
65
 
66
  | Training Loss | Epoch | Step | Validation Loss | Accuracy |
67
  |:-------------:|:-----:|:----:|:---------------:|:--------:|
68
+ | 0.244 | 1.0 | 527 | 0.2574 | 0.9083 |
 
 
 
 
 
 
69
 
70
 
71
  ### Framework versions
compressed_graph.dot ADDED
The diff for this file is too large to render. See raw diff
 
logs/events.out.tfevents.1700312106.0a848c80699a.1576.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c33d221ab0910be5d758fb04c7aebf629f244f1223c6b9112c2c2d89a6e89aa
3
+ size 5069
logs/events.out.tfevents.1700312559.0a848c80699a.1576.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:550e6d05743dca4c9e7b90c2cba8bda631811f82e98e62c4ea67e9d7cdc37b45
3
+ size 411
nncf_output.log ADDED
@@ -0,0 +1,188 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ INFO:nncf:Not adding activation input quantizer for operation: 7 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0
2
+ INFO:nncf:Not adding activation input quantizer for operation: 4 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0
3
+ INFO:nncf:Not adding activation input quantizer for operation: 5 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0
4
+ INFO:nncf:Not adding activation input quantizer for operation: 6 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/__add___0
5
+ INFO:nncf:Not adding activation input quantizer for operation: 8 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/__iadd___0
6
+ INFO:nncf:Not adding activation input quantizer for operation: 9 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/NNCFLayerNorm[LayerNorm]/layer_norm_0
7
+ INFO:nncf:Not adding activation input quantizer for operation: 10 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/Dropout[dropout]/dropout_0
8
+ INFO:nncf:Not adding activation input quantizer for operation: 23 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/__add___0
9
+ INFO:nncf:Not adding activation input quantizer for operation: 26 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
10
+ INFO:nncf:Not adding activation input quantizer for operation: 32 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/__add___0
11
+ INFO:nncf:Not adding activation input quantizer for operation: 33 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
12
+ INFO:nncf:Not adding activation input quantizer for operation: 37 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/__add___0
13
+ INFO:nncf:Not adding activation input quantizer for operation: 38 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
14
+ INFO:nncf:Not adding activation input quantizer for operation: 51 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/__add___0
15
+ INFO:nncf:Not adding activation input quantizer for operation: 54 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
16
+ INFO:nncf:Not adding activation input quantizer for operation: 60 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/__add___0
17
+ INFO:nncf:Not adding activation input quantizer for operation: 61 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
18
+ INFO:nncf:Not adding activation input quantizer for operation: 65 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/__add___0
19
+ INFO:nncf:Not adding activation input quantizer for operation: 66 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
20
+ INFO:nncf:Not adding activation input quantizer for operation: 79 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/__add___0
21
+ INFO:nncf:Not adding activation input quantizer for operation: 82 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
22
+ INFO:nncf:Not adding activation input quantizer for operation: 88 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/__add___0
23
+ INFO:nncf:Not adding activation input quantizer for operation: 89 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
24
+ INFO:nncf:Not adding activation input quantizer for operation: 93 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/__add___0
25
+ INFO:nncf:Not adding activation input quantizer for operation: 94 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
26
+ INFO:nncf:Not adding activation input quantizer for operation: 107 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/__add___0
27
+ INFO:nncf:Not adding activation input quantizer for operation: 110 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
28
+ INFO:nncf:Not adding activation input quantizer for operation: 116 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/__add___0
29
+ INFO:nncf:Not adding activation input quantizer for operation: 117 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
30
+ INFO:nncf:Not adding activation input quantizer for operation: 121 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/__add___0
31
+ INFO:nncf:Not adding activation input quantizer for operation: 122 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
32
+ INFO:nncf:Not adding activation input quantizer for operation: 135 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/__add___0
33
+ INFO:nncf:Not adding activation input quantizer for operation: 138 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
34
+ INFO:nncf:Not adding activation input quantizer for operation: 144 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/__add___0
35
+ INFO:nncf:Not adding activation input quantizer for operation: 145 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
36
+ INFO:nncf:Not adding activation input quantizer for operation: 149 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/__add___0
37
+ INFO:nncf:Not adding activation input quantizer for operation: 150 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
38
+ INFO:nncf:Not adding activation input quantizer for operation: 163 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/__add___0
39
+ INFO:nncf:Not adding activation input quantizer for operation: 166 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
40
+ INFO:nncf:Not adding activation input quantizer for operation: 172 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/__add___0
41
+ INFO:nncf:Not adding activation input quantizer for operation: 173 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
42
+ INFO:nncf:Not adding activation input quantizer for operation: 177 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/__add___0
43
+ INFO:nncf:Not adding activation input quantizer for operation: 178 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
44
+ INFO:nncf:Collecting tensor statistics |█████ | 1 / 3
45
+ INFO:nncf:Collecting tensor statistics |██████████ | 2 / 3
46
+ INFO:nncf:Collecting tensor statistics |████████████████| 3 / 3
47
+ INFO:nncf:Compiling and loading torch extension: quantized_functions_cuda...
48
+ INFO:nncf:Not adding activation input quantizer for operation: 7 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0
49
+ INFO:nncf:Not adding activation input quantizer for operation: 4 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0
50
+ INFO:nncf:Not adding activation input quantizer for operation: 5 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0
51
+ INFO:nncf:Not adding activation input quantizer for operation: 6 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/__add___0
52
+ INFO:nncf:Not adding activation input quantizer for operation: 8 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/__iadd___0
53
+ INFO:nncf:Not adding activation input quantizer for operation: 9 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/NNCFLayerNorm[LayerNorm]/layer_norm_0
54
+ INFO:nncf:Not adding activation input quantizer for operation: 10 BertForSequenceClassification/BertModel[bert]/BertEmbeddings[embeddings]/Dropout[dropout]/dropout_0
55
+ INFO:nncf:Not adding activation input quantizer for operation: 23 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/__add___0
56
+ INFO:nncf:Not adding activation input quantizer for operation: 26 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
57
+ INFO:nncf:Not adding activation input quantizer for operation: 32 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/__add___0
58
+ INFO:nncf:Not adding activation input quantizer for operation: 33 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
59
+ INFO:nncf:Not adding activation input quantizer for operation: 37 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/__add___0
60
+ INFO:nncf:Not adding activation input quantizer for operation: 38 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
61
+ INFO:nncf:Not adding activation input quantizer for operation: 51 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/__add___0
62
+ INFO:nncf:Not adding activation input quantizer for operation: 54 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
63
+ INFO:nncf:Not adding activation input quantizer for operation: 60 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/__add___0
64
+ INFO:nncf:Not adding activation input quantizer for operation: 61 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
65
+ INFO:nncf:Not adding activation input quantizer for operation: 65 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/__add___0
66
+ INFO:nncf:Not adding activation input quantizer for operation: 66 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
67
+ INFO:nncf:Not adding activation input quantizer for operation: 79 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/__add___0
68
+ INFO:nncf:Not adding activation input quantizer for operation: 82 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
69
+ INFO:nncf:Not adding activation input quantizer for operation: 88 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/__add___0
70
+ INFO:nncf:Not adding activation input quantizer for operation: 89 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
71
+ INFO:nncf:Not adding activation input quantizer for operation: 93 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/__add___0
72
+ INFO:nncf:Not adding activation input quantizer for operation: 94 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
73
+ INFO:nncf:Not adding activation input quantizer for operation: 107 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/__add___0
74
+ INFO:nncf:Not adding activation input quantizer for operation: 110 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
75
+ INFO:nncf:Not adding activation input quantizer for operation: 116 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/__add___0
76
+ INFO:nncf:Not adding activation input quantizer for operation: 117 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
77
+ INFO:nncf:Not adding activation input quantizer for operation: 121 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/__add___0
78
+ INFO:nncf:Not adding activation input quantizer for operation: 122 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
79
+ INFO:nncf:Not adding activation input quantizer for operation: 135 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/__add___0
80
+ INFO:nncf:Not adding activation input quantizer for operation: 138 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
81
+ INFO:nncf:Not adding activation input quantizer for operation: 144 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/__add___0
82
+ INFO:nncf:Not adding activation input quantizer for operation: 145 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
83
+ INFO:nncf:Not adding activation input quantizer for operation: 149 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/__add___0
84
+ INFO:nncf:Not adding activation input quantizer for operation: 150 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
85
+ INFO:nncf:Not adding activation input quantizer for operation: 163 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/__add___0
86
+ INFO:nncf:Not adding activation input quantizer for operation: 166 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/matmul_1
87
+ INFO:nncf:Not adding activation input quantizer for operation: 172 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/__add___0
88
+ INFO:nncf:Not adding activation input quantizer for operation: 173 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
89
+ INFO:nncf:Not adding activation input quantizer for operation: 177 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/__add___0
90
+ INFO:nncf:Not adding activation input quantizer for operation: 178 BertForSequenceClassification/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/NNCFLayerNorm[LayerNorm]/layer_norm_0
91
+ INFO:nncf:Collecting tensor statistics |█████ | 1 / 3
92
+ INFO:nncf:Collecting tensor statistics |██████████ | 2 / 3
93
+ INFO:nncf:Collecting tensor statistics |████████████████| 3 / 3
94
+ INFO:nncf:Compiling and loading torch extension: quantized_functions_cuda...
95
+ INFO:nncf:Finished loading torch extension: quantized_functions_cuda
96
+ WARNING:nncf:You are setting `forward` on an NNCF-processed model object.
97
+ NNCF relies on custom-wrapping the `forward` call in order to function properly.
98
+ Arbitrary adjustments to the forward function on an NNCFNetwork object have undefined behavior.
99
+ If you need to replace the underlying forward function of the original model so that NNCF should be using that instead of the original forward function that NNCF saved during the compressed model creation, you can do this by calling:
100
+ model.nncf.set_original_unbound_forward(fn)
101
+ if `fn` has an unbound 0-th `self` argument, or
102
+ with model.nncf.temporary_bound_original_forward(fn): ...
103
+ if `fn` already had 0-th `self` argument bound or never had it in the first place.
104
+ WARNING:nncf:You are setting `forward` on an NNCF-processed model object.
105
+ NNCF relies on custom-wrapping the `forward` call in order to function properly.
106
+ Arbitrary adjustments to the forward function on an NNCFNetwork object have undefined behavior.
107
+ If you need to replace the underlying forward function of the original model so that NNCF should be using that instead of the original forward function that NNCF saved during the compressed model creation, you can do this by calling:
108
+ model.nncf.set_original_unbound_forward(fn)
109
+ if `fn` has an unbound 0-th `self` argument, or
110
+ with model.nncf.temporary_bound_original_forward(fn): ...
111
+ if `fn` already had 0-th `self` argument bound or never had it in the first place.
112
+ INFO:nncf:Statistics of the quantization algorithm:
113
+ Epoch 0 |+--------------------------------+-------+
114
+ Epoch 0 || Statistic's name | Value |
115
+ Epoch 0 |+================================+=======+
116
+ Epoch 0 || Ratio of enabled quantizations | 100 |
117
+ Epoch 0 |+--------------------------------+-------+
118
+ Epoch 0 |
119
+ Epoch 0 |Statistics of the quantization share:
120
+ Epoch 0 |+----------------------------------+--------------------+
121
+ Epoch 0 || Statistic's name | Value |
122
+ Epoch 0 |+==================================+====================+
123
+ Epoch 0 || Symmetric WQs / All placed WQs | 100.00 % (38 / 38) |
124
+ Epoch 0 |+----------------------------------+--------------------+
125
+ Epoch 0 || Asymmetric WQs / All placed WQs | 0.00 % (0 / 38) |
126
+ Epoch 0 |+----------------------------------+--------------------+
127
+ Epoch 0 || Signed WQs / All placed WQs | 100.00 % (38 / 38) |
128
+ Epoch 0 |+----------------------------------+--------------------+
129
+ Epoch 0 || Unsigned WQs / All placed WQs | 0.00 % (0 / 38) |
130
+ Epoch 0 |+----------------------------------+--------------------+
131
+ Epoch 0 || Per-tensor WQs / All placed WQs | 0.00 % (0 / 38) |
132
+ Epoch 0 |+----------------------------------+--------------------+
133
+ Epoch 0 || Per-channel WQs / All placed WQs | 100.00 % (38 / 38) |
134
+ Epoch 0 |+----------------------------------+--------------------+
135
+ Epoch 0 || Placed WQs / Potential WQs | 70.37 % (38 / 54) |
136
+ Epoch 0 |+----------------------------------+--------------------+
137
+ Epoch 0 || Symmetric AQs / All placed AQs | 27.27 % (12 / 44) |
138
+ Epoch 0 |+----------------------------------+--------------------+
139
+ Epoch 0 || Asymmetric AQs / All placed AQs | 72.73 % (32 / 44) |
140
+ Epoch 0 |+----------------------------------+--------------------+
141
+ Epoch 0 || Signed AQs / All placed AQs | 100.00 % (44 / 44) |
142
+ Epoch 0 |+----------------------------------+--------------------+
143
+ Epoch 0 || Unsigned AQs / All placed AQs | 0.00 % (0 / 44) |
144
+ Epoch 0 |+----------------------------------+--------------------+
145
+ Epoch 0 || Per-tensor AQs / All placed AQs | 100.00 % (44 / 44) |
146
+ Epoch 0 |+----------------------------------+--------------------+
147
+ Epoch 0 || Per-channel AQs / All placed AQs | 0.00 % (0 / 44) |
148
+ Epoch 0 |+----------------------------------+--------------------+
149
+ Epoch 0 |
150
+ Epoch 0 |Statistics of the bitwidth distribution:
151
+ Epoch 0 |+--------------+---------------------+--------------------+--------------------+
152
+ Epoch 0 || Num bits (N) | N-bits WQs / Placed | N-bits AQs / | N-bits Qs / Placed |
153
+ Epoch 0 || | WQs | Placed AQs | Qs |
154
+ Epoch 0 |+==============+=====================+====================+====================+
155
+ Epoch 0 || 8 | 100.00 % (38 / 38) | 100.00 % (44 / 44) | 100.00 % (82 / 82) |
156
+ Epoch 0 |+--------------+---------------------+--------------------+--------------------+
157
+ WARNING:nncf:You are setting `forward` on an NNCF-processed model object.
158
+ NNCF relies on custom-wrapping the `forward` call in order to function properly.
159
+ Arbitrary adjustments to the forward function on an NNCFNetwork object have undefined behavior.
160
+ If you need to replace the underlying forward function of the original model so that NNCF should be using that instead of the original forward function that NNCF saved during the compressed model creation, you can do this by calling:
161
+ model.nncf.set_original_unbound_forward(fn)
162
+ if `fn` has an unbound 0-th `self` argument, or
163
+ with model.nncf.temporary_bound_original_forward(fn): ...
164
+ if `fn` already had 0-th `self` argument bound or never had it in the first place.
165
+ WARNING:nncf:You are setting `forward` on an NNCF-processed model object.
166
+ NNCF relies on custom-wrapping the `forward` call in order to function properly.
167
+ Arbitrary adjustments to the forward function on an NNCFNetwork object have undefined behavior.
168
+ If you need to replace the underlying forward function of the original model so that NNCF should be using that instead of the original forward function that NNCF saved during the compressed model creation, you can do this by calling:
169
+ model.nncf.set_original_unbound_forward(fn)
170
+ if `fn` has an unbound 0-th `self` argument, or
171
+ with model.nncf.temporary_bound_original_forward(fn): ...
172
+ if `fn` already had 0-th `self` argument bound or never had it in the first place.
173
+ WARNING:nncf:You are setting `forward` on an NNCF-processed model object.
174
+ NNCF relies on custom-wrapping the `forward` call in order to function properly.
175
+ Arbitrary adjustments to the forward function on an NNCFNetwork object have undefined behavior.
176
+ If you need to replace the underlying forward function of the original model so that NNCF should be using that instead of the original forward function that NNCF saved during the compressed model creation, you can do this by calling:
177
+ model.nncf.set_original_unbound_forward(fn)
178
+ if `fn` has an unbound 0-th `self` argument, or
179
+ with model.nncf.temporary_bound_original_forward(fn): ...
180
+ if `fn` already had 0-th `self` argument bound or never had it in the first place.
181
+ WARNING:nncf:You are setting `forward` on an NNCF-processed model object.
182
+ NNCF relies on custom-wrapping the `forward` call in order to function properly.
183
+ Arbitrary adjustments to the forward function on an NNCFNetwork object have undefined behavior.
184
+ If you need to replace the underlying forward function of the original model so that NNCF should be using that instead of the original forward function that NNCF saved during the compressed model creation, you can do this by calling:
185
+ model.nncf.set_original_unbound_forward(fn)
186
+ if `fn` has an unbound 0-th `self` argument, or
187
+ with model.nncf.temporary_bound_original_forward(fn): ...
188
+ if `fn` already had 0-th `self` argument bound or never had it in the first place.
openvino_config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "compression": {
3
+ "algorithm": "quantization",
4
+ "export_to_onnx_standard_ops": false,
5
+ "ignored_scopes": [
6
+ "{re}.*Embedding.*",
7
+ "{re}.*add___.*",
8
+ "{re}.*layer_norm_.*",
9
+ "{re}.*matmul_1",
10
+ "{re}.*__truediv__.*"
11
+ ],
12
+ "initializer": {
13
+ "batchnorm_adaptation": {
14
+ "num_bn_adaptation_samples": 0
15
+ },
16
+ "range": {
17
+ "num_init_samples": 300,
18
+ "type": "mean_min_max"
19
+ }
20
+ },
21
+ "overflow_fix": "disable",
22
+ "preset": "mixed",
23
+ "scope_overrides": {
24
+ "activations": {
25
+ "{re}.*matmul_0": {
26
+ "mode": "symmetric"
27
+ }
28
+ }
29
+ }
30
+ },
31
+ "input_info": [
32
+ {
33
+ "keyword": "input_ids",
34
+ "sample_size": [
35
+ 128,
36
+ 66
37
+ ],
38
+ "type": "long"
39
+ },
40
+ {
41
+ "keyword": "token_type_ids",
42
+ "sample_size": [
43
+ 128,
44
+ 66
45
+ ],
46
+ "type": "long"
47
+ },
48
+ {
49
+ "keyword": "attention_mask",
50
+ "sample_size": [
51
+ 128,
52
+ 66
53
+ ],
54
+ "type": "long"
55
+ }
56
+ ],
57
+ "optimum_version": "1.14.1",
58
+ "save_onnx_model": false,
59
+ "transformers_version": "4.35.2"
60
+ }
openvino_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14e7a63cb6ed4e71f0fb41442750168d21fe4cc1f4a932ab908a126c9ffc3058
3
+ size 138739212
openvino_model.xml ADDED
The diff for this file is too large to render. See raw diff
 
original_graph.dot ADDED
The diff for this file is too large to render. See raw diff
 
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0f7f5f1d23dd59c0a5ff0fbe1f57c2504e1a2a954ab8cee1440ae102e9640adf
3
- size 267862062
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6286ab040d463864d3a8e9d171df02ea044a878ecd16646b9aef131ecbea578
3
+ size 268172814
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3df994f15823187005e047b4e22092c87282c7d500e76747325096ed2b14a250
3
  size 4600
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:77fea7f47d6f91af7370043dfc2a8acdf21fdd576a23651c0002ed739a6042b2
3
  size 4600