Hugging Face's logo

raulgdp
/
roberta-large-ner-qlorafinetune-runs

TensorBoard
Safetensors
Generated from Trainer
Model card Files Metrics Community
raulgdp commited on
Commit
26db7aa
·
1 Parent(s): f181a0b

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -122
README.md CHANGED
@@ -1,16 +1,10 @@
1
  ---
2
- library_name: peft
3
  license: mit
4
  base_model: FacebookAI/xlm-roberta-large
5
  tags:
6
  - generated_from_trainer
7
  datasets:
8
  - conll2002
9
- metrics:
10
- - precision
11
- - recall
12
- - f1
13
- - accuracy
14
  model-index:
15
  - name: roberta-large-ner-qlorafinetune-runs
16
  results: []
@@ -22,12 +16,6 @@ should probably proofread and complete it, then remove this comment. -->
22
  # roberta-large-ner-qlorafinetune-runs
23
 
24
  This model is a fine-tuned version of [FacebookAI/xlm-roberta-large](https://huggingface.co/FacebookAI/xlm-roberta-large) on the conll2002 dataset.
25
- It achieves the following results on the evaluation set:
26
- - Loss: 0.0853
27
- - Precision: 0.8767
28
- - Recall: 0.8819
29
- - F1: 0.8793
30
- - Accuracy: 0.9814
31
 
32
  ## Model description
33
 
@@ -43,17 +31,6 @@ More information needed
43
 
44
  ## Training procedure
45
 
46
-
47
- The following `bitsandbytes` quantization config was used during training:
48
- - load_in_8bit: False
49
- - load_in_4bit: True
50
- - llm_int8_threshold: 6.0
51
- - llm_int8_skip_modules: ['classifier', 'pre_classifier']
52
- - llm_int8_enable_fp32_cpu_offload: False
53
- - llm_int8_has_fp16_weight: False
54
- - bnb_4bit_quant_type: nf4
55
- - bnb_4bit_use_double_quant: True
56
- - bnb_4bit_compute_dtype: bfloat16
57
  ### Training hyperparameters
58
 
59
  The following hyperparameters were used during training:
@@ -61,113 +38,17 @@ The following hyperparameters were used during training:
61
  - train_batch_size: 32
62
  - eval_batch_size: 32
63
  - seed: 42
64
- - optimizer: Use paged_adamw_8bit with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
65
  - lr_scheduler_type: linear
66
  - training_steps: 1820
67
- - mixed_precision_training: Native AMP
68
 
69
  ### Training results
70
 
71
- | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
72
- |:-------------:|:------:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
73
- | 0.9933 | 0.0766 | 20 | 0.4461 | 0.1132 | 0.0565 | 0.0754 | 0.8691 |
74
- | 0.2822 | 0.1533 | 40 | 0.2098 | 0.5212 | 0.6337 | 0.5720 | 0.9451 |
75
- | 0.146 | 0.2299 | 60 | 0.1251 | 0.7357 | 0.7652 | 0.7502 | 0.9657 |
76
- | 0.0986 | 0.3065 | 80 | 0.1222 | 0.7928 | 0.8134 | 0.8030 | 0.9720 |
77
- | 0.1087 | 0.3831 | 100 | 0.1080 | 0.7738 | 0.7953 | 0.7844 | 0.9714 |
78
- | 0.0874 | 0.4598 | 120 | 0.0974 | 0.8118 | 0.8205 | 0.8161 | 0.9742 |
79
- | 0.0876 | 0.5364 | 140 | 0.0791 | 0.8374 | 0.8532 | 0.8452 | 0.9791 |
80
- | 0.0909 | 0.6130 | 160 | 0.0831 | 0.8333 | 0.8486 | 0.8408 | 0.9779 |
81
- | 0.0835 | 0.6897 | 180 | 0.0802 | 0.8452 | 0.8479 | 0.8465 | 0.9773 |
82
- | 0.0706 | 0.7663 | 200 | 0.0879 | 0.8234 | 0.8359 | 0.8296 | 0.9757 |
83
- | 0.0634 | 0.8429 | 220 | 0.0812 | 0.8368 | 0.8504 | 0.8435 | 0.9782 |
84
- | 0.0548 | 0.9195 | 240 | 0.0918 | 0.8067 | 0.8235 | 0.8150 | 0.9741 |
85
- | 0.0602 | 0.9962 | 260 | 0.0903 | 0.8328 | 0.8458 | 0.8393 | 0.9762 |
86
- | 0.0528 | 1.0728 | 280 | 0.0878 | 0.8125 | 0.8456 | 0.8287 | 0.9760 |
87
- | 0.0513 | 1.1494 | 300 | 0.0991 | 0.8269 | 0.8332 | 0.8300 | 0.9754 |
88
- | 0.0565 | 1.2261 | 320 | 0.0805 | 0.8575 | 0.8683 | 0.8629 | 0.9788 |
89
- | 0.0514 | 1.3027 | 340 | 0.0864 | 0.8104 | 0.8359 | 0.8230 | 0.9751 |
90
- | 0.0499 | 1.3793 | 360 | 0.1011 | 0.8012 | 0.8194 | 0.8102 | 0.9740 |
91
- | 0.0585 | 1.4559 | 380 | 0.0945 | 0.8179 | 0.8258 | 0.8219 | 0.9739 |
92
- | 0.0489 | 1.5326 | 400 | 0.0887 | 0.8177 | 0.8389 | 0.8282 | 0.9755 |
93
- | 0.0534 | 1.6092 | 420 | 0.0747 | 0.8283 | 0.8523 | 0.8401 | 0.9784 |
94
- | 0.0612 | 1.6858 | 440 | 0.0743 | 0.8336 | 0.8532 | 0.8433 | 0.9775 |
95
- | 0.0531 | 1.7625 | 460 | 0.0722 | 0.8632 | 0.8699 | 0.8666 | 0.9802 |
96
- | 0.0453 | 1.8391 | 480 | 0.0752 | 0.8346 | 0.8502 | 0.8423 | 0.9776 |
97
- | 0.0504 | 1.9157 | 500 | 0.0754 | 0.8373 | 0.8562 | 0.8466 | 0.9783 |
98
- | 0.0468 | 1.9923 | 520 | 0.0734 | 0.8534 | 0.8612 | 0.8573 | 0.9794 |
99
- | 0.0573 | 2.0690 | 540 | 0.0728 | 0.8412 | 0.8653 | 0.8531 | 0.9785 |
100
- | 0.0344 | 2.1456 | 560 | 0.0741 | 0.8549 | 0.8690 | 0.8619 | 0.9794 |
101
- | 0.0303 | 2.2222 | 580 | 0.0767 | 0.8428 | 0.8658 | 0.8541 | 0.9783 |
102
- | 0.0421 | 2.2989 | 600 | 0.0753 | 0.8370 | 0.8587 | 0.8477 | 0.9786 |
103
- | 0.0335 | 2.3755 | 620 | 0.0762 | 0.8523 | 0.8594 | 0.8558 | 0.9788 |
104
- | 0.0376 | 2.4521 | 640 | 0.0735 | 0.8511 | 0.8732 | 0.8620 | 0.9797 |
105
- | 0.0366 | 2.5287 | 660 | 0.0764 | 0.8647 | 0.8619 | 0.8633 | 0.9792 |
106
- | 0.0351 | 2.6054 | 680 | 0.0771 | 0.8711 | 0.8713 | 0.8712 | 0.9795 |
107
- | 0.0389 | 2.6820 | 700 | 0.0734 | 0.8563 | 0.8640 | 0.8601 | 0.9788 |
108
- | 0.037 | 2.7586 | 720 | 0.0708 | 0.8475 | 0.8660 | 0.8567 | 0.9804 |
109
- | 0.0434 | 2.8352 | 740 | 0.0719 | 0.8441 | 0.8587 | 0.8513 | 0.9783 |
110
- | 0.0399 | 2.9119 | 760 | 0.0681 | 0.8572 | 0.8773 | 0.8671 | 0.9807 |
111
- | 0.0381 | 2.9885 | 780 | 0.0679 | 0.8615 | 0.8759 | 0.8686 | 0.9809 |
112
- | 0.0343 | 3.0651 | 800 | 0.0680 | 0.8593 | 0.8844 | 0.8717 | 0.9816 |
113
- | 0.0283 | 3.1418 | 820 | 0.0708 | 0.8760 | 0.8780 | 0.8770 | 0.9812 |
114
- | 0.0286 | 3.2184 | 840 | 0.0757 | 0.8688 | 0.8796 | 0.8742 | 0.9802 |
115
- | 0.0279 | 3.2950 | 860 | 0.0721 | 0.875 | 0.8798 | 0.8774 | 0.9810 |
116
- | 0.0274 | 3.3716 | 880 | 0.0717 | 0.8651 | 0.8766 | 0.8708 | 0.9796 |
117
- | 0.0296 | 3.4483 | 900 | 0.0697 | 0.8558 | 0.8782 | 0.8669 | 0.9796 |
118
- | 0.0355 | 3.5249 | 920 | 0.0624 | 0.8741 | 0.8856 | 0.8798 | 0.9820 |
119
- | 0.0223 | 3.6015 | 940 | 0.0751 | 0.8693 | 0.8803 | 0.8748 | 0.9814 |
120
- | 0.0265 | 3.6782 | 960 | 0.0785 | 0.8694 | 0.875 | 0.8722 | 0.9798 |
121
- | 0.0261 | 3.7548 | 980 | 0.0701 | 0.8768 | 0.8830 | 0.8799 | 0.9808 |
122
- | 0.0326 | 3.8314 | 1000 | 0.0693 | 0.8708 | 0.8842 | 0.8774 | 0.9820 |
123
- | 0.028 | 3.9080 | 1020 | 0.0719 | 0.8579 | 0.8752 | 0.8665 | 0.9795 |
124
- | 0.032 | 3.9847 | 1040 | 0.0726 | 0.8766 | 0.8801 | 0.8783 | 0.9809 |
125
- | 0.0176 | 4.0613 | 1060 | 0.0766 | 0.8753 | 0.8821 | 0.8787 | 0.9809 |
126
- | 0.0218 | 4.1379 | 1080 | 0.0808 | 0.8681 | 0.8805 | 0.8743 | 0.9810 |
127
- | 0.0163 | 4.2146 | 1100 | 0.0833 | 0.8682 | 0.875 | 0.8716 | 0.9803 |
128
- | 0.0201 | 4.2912 | 1120 | 0.0882 | 0.8702 | 0.8752 | 0.8727 | 0.9802 |
129
- | 0.0193 | 4.3678 | 1140 | 0.0838 | 0.8678 | 0.8778 | 0.8727 | 0.9801 |
130
- | 0.023 | 4.4444 | 1160 | 0.0855 | 0.8648 | 0.8761 | 0.8704 | 0.9802 |
131
- | 0.0241 | 4.5211 | 1180 | 0.0793 | 0.8585 | 0.8768 | 0.8676 | 0.9795 |
132
- | 0.0224 | 4.5977 | 1200 | 0.0806 | 0.8792 | 0.8863 | 0.8827 | 0.9814 |
133
- | 0.0263 | 4.6743 | 1220 | 0.0737 | 0.8674 | 0.8775 | 0.8724 | 0.9816 |
134
- | 0.0266 | 4.7510 | 1240 | 0.0824 | 0.8501 | 0.8706 | 0.8603 | 0.9794 |
135
- | 0.0212 | 4.8276 | 1260 | 0.0770 | 0.8657 | 0.8709 | 0.8683 | 0.9811 |
136
- | 0.0228 | 4.9042 | 1280 | 0.0724 | 0.8673 | 0.8771 | 0.8722 | 0.9810 |
137
- | 0.0216 | 4.9808 | 1300 | 0.0703 | 0.8747 | 0.8842 | 0.8794 | 0.9817 |
138
- | 0.0155 | 5.0575 | 1320 | 0.0782 | 0.8799 | 0.8904 | 0.8851 | 0.9824 |
139
- | 0.015 | 5.1341 | 1340 | 0.0792 | 0.8822 | 0.8842 | 0.8832 | 0.9822 |
140
- | 0.0159 | 5.2107 | 1360 | 0.0787 | 0.8771 | 0.8872 | 0.8821 | 0.9818 |
141
- | 0.0182 | 5.2874 | 1380 | 0.0766 | 0.8767 | 0.8821 | 0.8794 | 0.9816 |
142
- | 0.0147 | 5.3640 | 1400 | 0.0756 | 0.8753 | 0.8778 | 0.8765 | 0.9814 |
143
- | 0.0182 | 5.4406 | 1420 | 0.0813 | 0.8755 | 0.8794 | 0.8775 | 0.9809 |
144
- | 0.0163 | 5.5172 | 1440 | 0.0822 | 0.8823 | 0.8835 | 0.8829 | 0.9815 |
145
- | 0.0172 | 5.5939 | 1460 | 0.0819 | 0.8767 | 0.8810 | 0.8789 | 0.9813 |
146
- | 0.015 | 5.6705 | 1480 | 0.0777 | 0.8714 | 0.8796 | 0.8755 | 0.9811 |
147
- | 0.0108 | 5.7471 | 1500 | 0.0801 | 0.8801 | 0.8835 | 0.8818 | 0.9814 |
148
- | 0.0169 | 5.8238 | 1520 | 0.0798 | 0.8793 | 0.8853 | 0.8823 | 0.9818 |
149
- | 0.016 | 5.9004 | 1540 | 0.0832 | 0.8723 | 0.8771 | 0.8747 | 0.9810 |
150
- | 0.0128 | 5.9770 | 1560 | 0.0829 | 0.8730 | 0.8780 | 0.8755 | 0.9809 |
151
- | 0.0101 | 6.0536 | 1580 | 0.0834 | 0.8711 | 0.8773 | 0.8742 | 0.9808 |
152
- | 0.0124 | 6.1303 | 1600 | 0.0831 | 0.8710 | 0.8796 | 0.8753 | 0.9808 |
153
- | 0.012 | 6.2069 | 1620 | 0.0855 | 0.8702 | 0.8761 | 0.8731 | 0.9805 |
154
- | 0.0124 | 6.2835 | 1640 | 0.0822 | 0.8719 | 0.8805 | 0.8762 | 0.9811 |
155
- | 0.0126 | 6.3602 | 1660 | 0.0823 | 0.8762 | 0.8830 | 0.8796 | 0.9812 |
156
- | 0.0123 | 6.4368 | 1680 | 0.0814 | 0.8769 | 0.8853 | 0.8811 | 0.9812 |
157
- | 0.0098 | 6.5134 | 1700 | 0.0827 | 0.8732 | 0.8828 | 0.8780 | 0.9813 |
158
- | 0.0126 | 6.5900 | 1720 | 0.0820 | 0.8741 | 0.8810 | 0.8775 | 0.9810 |
159
- | 0.0108 | 6.6667 | 1740 | 0.0840 | 0.8778 | 0.8830 | 0.8804 | 0.9815 |
160
- | 0.0118 | 6.7433 | 1760 | 0.0852 | 0.8757 | 0.8824 | 0.8790 | 0.9814 |
161
- | 0.0103 | 6.8199 | 1780 | 0.0854 | 0.8773 | 0.8824 | 0.8798 | 0.9814 |
162
- | 0.0103 | 6.8966 | 1800 | 0.0854 | 0.8782 | 0.8828 | 0.8805 | 0.9815 |
163
- | 0.0081 | 6.9732 | 1820 | 0.0853 | 0.8767 | 0.8819 | 0.8793 | 0.9814 |
164
 
165
 
166
  ### Framework versions
167
 
168
- - PEFT 0.4.0
169
- - PEFT 0.13.2
170
- - Transformers 4.46.3
171
  - Pytorch 2.5.1
172
  - Datasets 3.1.0
173
- - Tokenizers 0.20.3
 
1
  ---
 
2
  license: mit
3
  base_model: FacebookAI/xlm-roberta-large
4
  tags:
5
  - generated_from_trainer
6
  datasets:
7
  - conll2002
 
 
 
 
 
8
  model-index:
9
  - name: roberta-large-ner-qlorafinetune-runs
10
  results: []
 
16
  # roberta-large-ner-qlorafinetune-runs
17
 
18
  This model is a fine-tuned version of [FacebookAI/xlm-roberta-large](https://huggingface.co/FacebookAI/xlm-roberta-large) on the conll2002 dataset.
 
 
 
 
 
 
19
 
20
  ## Model description
21
 
 
31
 
32
  ## Training procedure
33
 
 
 
 
 
 
 
 
 
 
 
 
34
  ### Training hyperparameters
35
 
36
  The following hyperparameters were used during training:
 
38
  - train_batch_size: 32
39
  - eval_batch_size: 32
40
  - seed: 42
41
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
42
  - lr_scheduler_type: linear
43
  - training_steps: 1820
 
44
 
45
  ### Training results
46
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
47
 
48
 
49
  ### Framework versions
50
 
51
+ - Transformers 4.31.0
 
 
52
  - Pytorch 2.5.1
53
  - Datasets 3.1.0
54
+ - Tokenizers 0.13.3