update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- generated_from_trainer
|
4 |
+
metrics:
|
5 |
+
- accuracy
|
6 |
+
model-index:
|
7 |
+
- name: dit-small_rvl_cdip_100_examples_per_class_simkd_CEKD_t1_aNone
|
8 |
+
results: []
|
9 |
+
---
|
10 |
+
|
11 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
12 |
+
should probably proofread and complete it, then remove this comment. -->
|
13 |
+
|
14 |
+
# dit-small_rvl_cdip_100_examples_per_class_simkd_CEKD_t1_aNone
|
15 |
+
|
16 |
+
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
|
17 |
+
It achieves the following results on the evaluation set:
|
18 |
+
- Loss: 0.1481
|
19 |
+
- Accuracy: 0.08
|
20 |
+
- Brier Loss: 0.9369
|
21 |
+
- Nll: 9.2883
|
22 |
+
- F1 Micro: 0.08
|
23 |
+
- F1 Macro: 0.0357
|
24 |
+
- Ece: 0.1153
|
25 |
+
- Aurc: 0.8531
|
26 |
+
|
27 |
+
## Model description
|
28 |
+
|
29 |
+
More information needed
|
30 |
+
|
31 |
+
## Intended uses & limitations
|
32 |
+
|
33 |
+
More information needed
|
34 |
+
|
35 |
+
## Training and evaluation data
|
36 |
+
|
37 |
+
More information needed
|
38 |
+
|
39 |
+
## Training procedure
|
40 |
+
|
41 |
+
### Training hyperparameters
|
42 |
+
|
43 |
+
The following hyperparameters were used during training:
|
44 |
+
- learning_rate: 2e-05
|
45 |
+
- train_batch_size: 4
|
46 |
+
- eval_batch_size: 4
|
47 |
+
- seed: 42
|
48 |
+
- gradient_accumulation_steps: 16
|
49 |
+
- total_train_batch_size: 64
|
50 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
51 |
+
- lr_scheduler_type: linear
|
52 |
+
- lr_scheduler_warmup_ratio: 0.1
|
53 |
+
- num_epochs: 25
|
54 |
+
|
55 |
+
### Training results
|
56 |
+
|
57 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|
58 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:------:|:--------:|:--------:|:------:|:------:|
|
59 |
+
| No log | 0.96 | 12 | 0.1528 | 0.0625 | 0.9377 | 9.9656 | 0.0625 | 0.0074 | 0.1025 | 0.9319 |
|
60 |
+
| No log | 1.96 | 24 | 0.1507 | 0.06 | 0.9377 | 9.9434 | 0.06 | 0.0074 | 0.1036 | 0.9537 |
|
61 |
+
| No log | 2.96 | 36 | 0.1500 | 0.0625 | 0.9376 | 8.6216 | 0.0625 | 0.0074 | 0.1019 | 0.9383 |
|
62 |
+
| No log | 3.96 | 48 | 0.1498 | 0.0625 | 0.9376 | 9.2776 | 0.0625 | 0.0074 | 0.1032 | 0.9438 |
|
63 |
+
| No log | 4.96 | 60 | 0.1496 | 0.0625 | 0.9375 | 9.3105 | 0.0625 | 0.0074 | 0.1017 | 0.9421 |
|
64 |
+
| No log | 5.96 | 72 | 0.1495 | 0.0625 | 0.9375 | 9.7276 | 0.0625 | 0.0074 | 0.1029 | 0.9380 |
|
65 |
+
| No log | 6.96 | 84 | 0.1494 | 0.0625 | 0.9374 | 9.6348 | 0.0625 | 0.0074 | 0.1017 | 0.9347 |
|
66 |
+
| No log | 7.96 | 96 | 0.1493 | 0.0625 | 0.9374 | 9.6145 | 0.0625 | 0.0074 | 0.1008 | 0.9359 |
|
67 |
+
| No log | 8.96 | 108 | 0.1492 | 0.0625 | 0.9374 | 9.5748 | 0.0625 | 0.0074 | 0.1019 | 0.9371 |
|
68 |
+
| No log | 9.96 | 120 | 0.1491 | 0.0625 | 0.9373 | 9.5551 | 0.0625 | 0.0074 | 0.1005 | 0.9372 |
|
69 |
+
| No log | 10.96 | 132 | 0.1490 | 0.065 | 0.9373 | 9.5267 | 0.065 | 0.0122 | 0.1047 | 0.9315 |
|
70 |
+
| No log | 11.96 | 144 | 0.1489 | 0.065 | 0.9373 | 9.5165 | 0.065 | 0.0122 | 0.1043 | 0.9284 |
|
71 |
+
| No log | 12.96 | 156 | 0.1488 | 0.065 | 0.9372 | 9.5162 | 0.065 | 0.0123 | 0.1068 | 0.9302 |
|
72 |
+
| No log | 13.96 | 168 | 0.1488 | 0.07 | 0.9372 | 9.5139 | 0.07 | 0.0213 | 0.1070 | 0.9275 |
|
73 |
+
| No log | 14.96 | 180 | 0.1487 | 0.0725 | 0.9371 | 9.4579 | 0.0725 | 0.0253 | 0.1095 | 0.9174 |
|
74 |
+
| No log | 15.96 | 192 | 0.1486 | 0.075 | 0.9371 | 9.3950 | 0.075 | 0.0286 | 0.1106 | 0.9161 |
|
75 |
+
| No log | 16.96 | 204 | 0.1485 | 0.075 | 0.9371 | 9.3347 | 0.075 | 0.0280 | 0.1055 | 0.9014 |
|
76 |
+
| No log | 17.96 | 216 | 0.1484 | 0.0775 | 0.9370 | 9.3157 | 0.0775 | 0.0315 | 0.1089 | 0.8695 |
|
77 |
+
| No log | 18.96 | 228 | 0.1483 | 0.08 | 0.9370 | 9.3125 | 0.08 | 0.0362 | 0.1133 | 0.8526 |
|
78 |
+
| No log | 19.96 | 240 | 0.1483 | 0.08 | 0.9370 | 9.2915 | 0.08 | 0.0360 | 0.1113 | 0.8554 |
|
79 |
+
| No log | 20.96 | 252 | 0.1482 | 0.0775 | 0.9370 | 9.2937 | 0.0775 | 0.0374 | 0.1118 | 0.8475 |
|
80 |
+
| No log | 21.96 | 264 | 0.1482 | 0.08 | 0.9369 | 9.2903 | 0.08 | 0.0357 | 0.1167 | 0.8526 |
|
81 |
+
| No log | 22.96 | 276 | 0.1482 | 0.08 | 0.9369 | 9.2888 | 0.08 | 0.0357 | 0.1099 | 0.8540 |
|
82 |
+
| No log | 23.96 | 288 | 0.1481 | 0.08 | 0.9369 | 9.2877 | 0.08 | 0.0357 | 0.1126 | 0.8531 |
|
83 |
+
| No log | 24.96 | 300 | 0.1481 | 0.08 | 0.9369 | 9.2883 | 0.08 | 0.0357 | 0.1153 | 0.8531 |
|
84 |
+
|
85 |
+
|
86 |
+
### Framework versions
|
87 |
+
|
88 |
+
- Transformers 4.26.1
|
89 |
+
- Pytorch 1.13.1.post200
|
90 |
+
- Datasets 2.9.0
|
91 |
+
- Tokenizers 0.13.2
|