Hugging Face's logo

karinegabsschon
/
classifier_adapter

TensorBoard
Generated from Trainer
Model card Files Metrics Community
karinegabsschon commited on
Commit
4209a31
·
verified ·
1 Parent(s): 9c1854d

karinegabsschon/classifier_adapter

Browse files
Files changed (10) hide show
  1. README.md +92 -0
  2. chinese/adapter_config.json +41 -0
  3. chinese/head_config.json +78 -0
  4. chinese/pytorch_adapter.bin +3 -0
  5. chinese/pytorch_model_head.bin +3 -0
  6. default/head_config.json +19 -0
  7. default/pytorch_model_head.bin +3 -0
  8. runs/Apr03_05-48-20_f9df232fefca/events.out.tfevents.1712123305.f9df232fefca.1407.0 +3 -0
  9. runs/Apr03_05-48-20_f9df232fefca/events.out.tfevents.1712125430.f9df232fefca.1407.1 +3 -0
  10. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: bert-base-chinese
3
+ tags:
4
+ - generated_from_trainer
5
+ metrics:
6
+ - accuracy
7
+ - precision
8
+ - recall
9
+ - f1
10
+ model-index:
11
+ - name: classifier_adapter
12
+ results: []
13
+ ---
14
+
15
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
16
+ should probably proofread and complete it, then remove this comment. -->
17
+
18
+ # classifier_adapter
19
+
20
+ This model is a fine-tuned version of [bert-base-chinese](https://huggingface.co/bert-base-chinese) on an unknown dataset.
21
+ It achieves the following results on the evaluation set:
22
+ - Loss: 0.0386
23
+ - Accuracy: 0.9875
24
+ - Precision: 0.8841
25
+ - Recall: 0.7947
26
+ - F1: 0.8283
27
+ - Ap: 0.8850
28
+
29
+ ## Model description
30
+
31
+ More information needed
32
+
33
+ ## Intended uses & limitations
34
+
35
+ More information needed
36
+
37
+ ## Training and evaluation data
38
+
39
+ More information needed
40
+
41
+ ## Training procedure
42
+
43
+ ### Training hyperparameters
44
+
45
+ The following hyperparameters were used during training:
46
+ - learning_rate: 0.0001
47
+ - train_batch_size: 16
48
+ - eval_batch_size: 16
49
+ - seed: 0
50
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
51
+ - lr_scheduler_type: linear
52
+ - num_epochs: 12
53
+
54
+ ### Training results
55
+
56
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | Ap |
57
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:------:|
58
+ | No log | 0.38 | 100 | 0.1590 | 0.9571 | 0.0 | 0.0 | 0.0 | 0.1046 |
59
+ | No log | 0.75 | 200 | 0.1578 | 0.9571 | 0.0 | 0.0 | 0.0 | 0.1808 |
60
+ | No log | 1.13 | 300 | 0.1185 | 0.9653 | 0.0899 | 0.0599 | 0.0680 | 0.4391 |
61
+ | No log | 1.51 | 400 | 0.0898 | 0.9724 | 0.2199 | 0.1409 | 0.1617 | 0.6479 |
62
+ | 0.1405 | 1.89 | 500 | 0.0774 | 0.9750 | 0.3319 | 0.2273 | 0.2575 | 0.7417 |
63
+ | 0.1405 | 2.26 | 600 | 0.0683 | 0.9771 | 0.4118 | 0.3002 | 0.3294 | 0.7791 |
64
+ | 0.1405 | 2.64 | 700 | 0.0616 | 0.9804 | 0.6207 | 0.4336 | 0.4810 | 0.8187 |
65
+ | 0.1405 | 3.02 | 800 | 0.0556 | 0.9821 | 0.7210 | 0.4875 | 0.5435 | 0.8380 |
66
+ | 0.1405 | 3.4 | 900 | 0.0519 | 0.9830 | 0.7329 | 0.5224 | 0.5839 | 0.8566 |
67
+ | 0.0598 | 3.77 | 1000 | 0.0486 | 0.9846 | 0.7818 | 0.6063 | 0.6615 | 0.8629 |
68
+ | 0.0598 | 4.15 | 1100 | 0.0469 | 0.9853 | 0.8223 | 0.6807 | 0.7248 | 0.8633 |
69
+ | 0.0598 | 4.53 | 1200 | 0.0457 | 0.9856 | 0.8521 | 0.7235 | 0.7663 | 0.8666 |
70
+ | 0.0598 | 4.91 | 1300 | 0.0439 | 0.9859 | 0.8436 | 0.6955 | 0.7435 | 0.8753 |
71
+ | 0.0598 | 5.28 | 1400 | 0.0424 | 0.9862 | 0.8715 | 0.6964 | 0.7496 | 0.8739 |
72
+ | 0.0399 | 5.66 | 1500 | 0.0415 | 0.9869 | 0.8695 | 0.7621 | 0.7994 | 0.8772 |
73
+ | 0.0399 | 6.04 | 1600 | 0.0416 | 0.9865 | 0.8700 | 0.7670 | 0.8039 | 0.8853 |
74
+ | 0.0399 | 6.42 | 1700 | 0.0401 | 0.9871 | 0.8687 | 0.7686 | 0.8047 | 0.8846 |
75
+ | 0.0399 | 6.79 | 1800 | 0.0405 | 0.9867 | 0.8734 | 0.7851 | 0.8167 | 0.8848 |
76
+ | 0.0399 | 7.17 | 1900 | 0.0410 | 0.9865 | 0.8600 | 0.7708 | 0.8057 | 0.8770 |
77
+ | 0.0315 | 7.55 | 2000 | 0.0393 | 0.9873 | 0.8869 | 0.7718 | 0.8158 | 0.8819 |
78
+ | 0.0315 | 7.92 | 2100 | 0.0385 | 0.9871 | 0.8747 | 0.7861 | 0.8196 | 0.8856 |
79
+ | 0.0315 | 8.3 | 2200 | 0.0386 | 0.9877 | 0.8863 | 0.7856 | 0.8227 | 0.8857 |
80
+ | 0.0315 | 8.68 | 2300 | 0.0390 | 0.9869 | 0.8695 | 0.7949 | 0.8221 | 0.8830 |
81
+ | 0.0315 | 9.06 | 2400 | 0.0391 | 0.9872 | 0.8685 | 0.8081 | 0.8311 | 0.8830 |
82
+ | 0.026 | 9.43 | 2500 | 0.0386 | 0.9875 | 0.8841 | 0.7947 | 0.8283 | 0.8850 |
83
+ | 0.026 | 9.81 | 2600 | 0.0390 | 0.9871 | 0.8615 | 0.8064 | 0.8264 | 0.8840 |
84
+ | 0.026 | 10.19 | 2700 | 0.0386 | 0.9873 | 0.8689 | 0.8023 | 0.8264 | 0.8859 |
85
+ | 0.026 | 10.57 | 2800 | 0.0386 | 0.9873 | 0.8737 | 0.7986 | 0.8265 | 0.8860 |
86
+
87
+
88
+ ### Framework versions
89
+
90
+ - Transformers 4.36.2
91
+ - Pytorch 2.2.1+cu121
92
+ - Tokenizers 0.15.2
chinese/adapter_config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "adapter_residual_before_ln": false,
4
+ "cross_adapter": false,
5
+ "factorized_phm_W": true,
6
+ "factorized_phm_rule": false,
7
+ "hypercomplex_nonlinearity": "glorot-uniform",
8
+ "init_weights": "bert",
9
+ "inv_adapter": null,
10
+ "inv_adapter_reduction_factor": null,
11
+ "is_parallel": false,
12
+ "learn_phm": true,
13
+ "leave_out": [],
14
+ "ln_after": false,
15
+ "ln_before": false,
16
+ "mh_adapter": false,
17
+ "non_linearity": "relu",
18
+ "original_ln_after": true,
19
+ "original_ln_before": true,
20
+ "output_adapter": true,
21
+ "phm_bias": true,
22
+ "phm_c_init": "normal",
23
+ "phm_dim": 4,
24
+ "phm_init_range": 0.0001,
25
+ "phm_layer": false,
26
+ "phm_rank": 1,
27
+ "reduction_factor": 2,
28
+ "residual_before_ln": true,
29
+ "scaling": 1.0,
30
+ "shared_W_phm": false,
31
+ "shared_phm_rule": true,
32
+ "use_gating": false
33
+ },
34
+ "config_id": "e15331acc3f0a5c0",
35
+ "hidden_size": 768,
36
+ "model_class": "BertAdapterModel",
37
+ "model_name": "bert-base-chinese",
38
+ "model_type": "bert",
39
+ "name": "chinese",
40
+ "version": "0.1.2"
41
+ }
chinese/head_config.json ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "activation_function": "tanh",
4
+ "bias": true,
5
+ "dropout_prob": null,
6
+ "head_type": "multilabel_classification",
7
+ "label2id": {
8
+ "LABEL_0": 0,
9
+ "LABEL_1": 1,
10
+ "LABEL_10": 10,
11
+ "LABEL_11": 11,
12
+ "LABEL_12": 12,
13
+ "LABEL_13": 13,
14
+ "LABEL_14": 14,
15
+ "LABEL_15": 15,
16
+ "LABEL_16": 16,
17
+ "LABEL_17": 17,
18
+ "LABEL_18": 18,
19
+ "LABEL_19": 19,
20
+ "LABEL_2": 2,
21
+ "LABEL_20": 20,
22
+ "LABEL_21": 21,
23
+ "LABEL_22": 22,
24
+ "LABEL_23": 23,
25
+ "LABEL_24": 24,
26
+ "LABEL_25": 25,
27
+ "LABEL_26": 26,
28
+ "LABEL_27": 27,
29
+ "LABEL_28": 28,
30
+ "LABEL_29": 29,
31
+ "LABEL_3": 3,
32
+ "LABEL_30": 30,
33
+ "LABEL_31": 31,
34
+ "LABEL_32": 32,
35
+ "LABEL_33": 33,
36
+ "LABEL_34": 34,
37
+ "LABEL_35": 35,
38
+ "LABEL_36": 36,
39
+ "LABEL_37": 37,
40
+ "LABEL_38": 38,
41
+ "LABEL_39": 39,
42
+ "LABEL_4": 4,
43
+ "LABEL_40": 40,
44
+ "LABEL_41": 41,
45
+ "LABEL_42": 42,
46
+ "LABEL_43": 43,
47
+ "LABEL_44": 44,
48
+ "LABEL_45": 45,
49
+ "LABEL_46": 46,
50
+ "LABEL_47": 47,
51
+ "LABEL_48": 48,
52
+ "LABEL_49": 49,
53
+ "LABEL_5": 5,
54
+ "LABEL_50": 50,
55
+ "LABEL_51": 51,
56
+ "LABEL_52": 52,
57
+ "LABEL_53": 53,
58
+ "LABEL_54": 54,
59
+ "LABEL_55": 55,
60
+ "LABEL_56": 56,
61
+ "LABEL_57": 57,
62
+ "LABEL_58": 58,
63
+ "LABEL_6": 6,
64
+ "LABEL_7": 7,
65
+ "LABEL_8": 8,
66
+ "LABEL_9": 9
67
+ },
68
+ "layers": 2,
69
+ "num_labels": 59,
70
+ "use_pooler": false
71
+ },
72
+ "hidden_size": 768,
73
+ "model_class": "BertAdapterModel",
74
+ "model_name": "bert-base-chinese",
75
+ "model_type": "bert",
76
+ "name": "chinese",
77
+ "version": "0.1.2"
78
+ }
chinese/pytorch_adapter.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e760a84bfe8ec1e51e27d19c67e872df6e030d8e62fab207f9ba194780fe750
3
+ size 28384358
chinese/pytorch_model_head.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf87d7edff3b9b6a8f3327bf72d3a14dab541775c37cee36cbaa198682c529f8
3
+ size 2545960
default/head_config.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "activation_function": "gelu",
4
+ "bias": true,
5
+ "embedding_size": 768,
6
+ "head_type": "masked_lm",
7
+ "label2id": null,
8
+ "layer_norm": true,
9
+ "layers": 2,
10
+ "shift_labels": false,
11
+ "vocab_size": 21128
12
+ },
13
+ "hidden_size": 768,
14
+ "model_class": "BertAdapterModel",
15
+ "model_name": "bert-base-chinese",
16
+ "model_type": "bert",
17
+ "name": "default",
18
+ "version": "0.1.2"
19
+ }
default/pytorch_model_head.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3c5d613ac771aa2cb6415fd5d0d26e5ccd78a903951bd0b81f198839eb6affa
3
+ size 67360950
runs/Apr03_05-48-20_f9df232fefca/events.out.tfevents.1712123305.f9df232fefca.1407.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9647f1862c19f35eb789df6a2cad5c6c394215e5094106bd21208d28e96b6412
3
+ size 24698
runs/Apr03_05-48-20_f9df232fefca/events.out.tfevents.1712125430.f9df232fefca.1407.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:71004c51b4effde01c68ab1f5d1373348946362042b5da131515595e9a81736b
3
+ size 606
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b05603fa23bd802794b61ac3e0a6953653325c8206618806f5fa9d0840662c6
3
+ size 4664