update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
datasets:
|
6 |
+
- esnli
|
7 |
+
metrics:
|
8 |
+
- f1
|
9 |
+
- accuracy
|
10 |
+
model-index:
|
11 |
+
- name: roberta-large-e-snli-classification-nli-base
|
12 |
+
results:
|
13 |
+
- task:
|
14 |
+
name: Text Classification
|
15 |
+
type: text-classification
|
16 |
+
dataset:
|
17 |
+
name: esnli
|
18 |
+
type: esnli
|
19 |
+
config: plain_text
|
20 |
+
split: validation
|
21 |
+
args: plain_text
|
22 |
+
metrics:
|
23 |
+
- name: F1
|
24 |
+
type: f1
|
25 |
+
value: 0.9258678577111056
|
26 |
+
- name: Accuracy
|
27 |
+
type: accuracy
|
28 |
+
value: 0.9260312944523471
|
29 |
+
---
|
30 |
+
|
31 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
32 |
+
should probably proofread and complete it, then remove this comment. -->
|
33 |
+
|
34 |
+
# roberta-large-e-snli-classification-nli-base
|
35 |
+
|
36 |
+
This model is a fine-tuned version of [roberta-large](https://huggingface.co/roberta-large) on the esnli dataset.
|
37 |
+
It achieves the following results on the evaluation set:
|
38 |
+
- Loss: 0.2221
|
39 |
+
- F1: 0.9259
|
40 |
+
- Accuracy: 0.9260
|
41 |
+
|
42 |
+
## Model description
|
43 |
+
|
44 |
+
More information needed
|
45 |
+
|
46 |
+
## Intended uses & limitations
|
47 |
+
|
48 |
+
More information needed
|
49 |
+
|
50 |
+
## Training and evaluation data
|
51 |
+
|
52 |
+
More information needed
|
53 |
+
|
54 |
+
## Training procedure
|
55 |
+
|
56 |
+
### Training hyperparameters
|
57 |
+
|
58 |
+
The following hyperparameters were used during training:
|
59 |
+
- learning_rate: 1e-05
|
60 |
+
- train_batch_size: 64
|
61 |
+
- eval_batch_size: 64
|
62 |
+
- seed: 42
|
63 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
64 |
+
- lr_scheduler_type: linear
|
65 |
+
- lr_scheduler_warmup_ratio: 0.05
|
66 |
+
- num_epochs: 3
|
67 |
+
- mixed_precision_training: Native AMP
|
68 |
+
|
69 |
+
### Training results
|
70 |
+
|
71 |
+
| Training Loss | Epoch | Step | Validation Loss | F1 | Accuracy |
|
72 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|:--------:|
|
73 |
+
| 0.9995 | 0.05 | 400 | 0.4236 | 0.8437 | 0.8465 |
|
74 |
+
| 0.4089 | 0.09 | 800 | 0.2961 | 0.8926 | 0.8933 |
|
75 |
+
| 0.3681 | 0.14 | 1200 | 0.2980 | 0.8914 | 0.8924 |
|
76 |
+
| 0.3467 | 0.19 | 1600 | 0.2872 | 0.8977 | 0.8990 |
|
77 |
+
| 0.324 | 0.23 | 2000 | 0.2506 | 0.9106 | 0.9110 |
|
78 |
+
| 0.3222 | 0.28 | 2400 | 0.2552 | 0.9132 | 0.9128 |
|
79 |
+
| 0.3138 | 0.33 | 2800 | 0.2379 | 0.9183 | 0.9183 |
|
80 |
+
| 0.3107 | 0.37 | 3200 | 0.2396 | 0.9152 | 0.9156 |
|
81 |
+
| 0.304 | 0.42 | 3600 | 0.2354 | 0.9174 | 0.9177 |
|
82 |
+
| 0.3027 | 0.47 | 4000 | 0.2360 | 0.9191 | 0.9191 |
|
83 |
+
| 0.2968 | 0.51 | 4400 | 0.2329 | 0.9182 | 0.9187 |
|
84 |
+
| 0.2888 | 0.56 | 4800 | 0.2462 | 0.9189 | 0.9196 |
|
85 |
+
| 0.2898 | 0.61 | 5200 | 0.2335 | 0.9206 | 0.9212 |
|
86 |
+
| 0.288 | 0.65 | 5600 | 0.2350 | 0.9220 | 0.9223 |
|
87 |
+
| 0.2746 | 0.7 | 6000 | 0.2208 | 0.9275 | 0.9278 |
|
88 |
+
| 0.2756 | 0.75 | 6400 | 0.2304 | 0.9209 | 0.9216 |
|
89 |
+
| 0.272 | 0.79 | 6800 | 0.2243 | 0.9237 | 0.9238 |
|
90 |
+
| 0.2809 | 0.84 | 7200 | 0.2176 | 0.9259 | 0.9261 |
|
91 |
+
| 0.2733 | 0.89 | 7600 | 0.2194 | 0.9271 | 0.9273 |
|
92 |
+
| 0.2723 | 0.93 | 8000 | 0.2221 | 0.9259 | 0.9260 |
|
93 |
+
|
94 |
+
|
95 |
+
### Framework versions
|
96 |
+
|
97 |
+
- Transformers 4.27.1
|
98 |
+
- Pytorch 1.12.1+cu113
|
99 |
+
- Datasets 2.10.1
|
100 |
+
- Tokenizers 0.13.2
|