Jungwoo4021
commited on
Commit
•
31c9927
1
Parent(s):
ade7e4e
update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
datasets:
|
6 |
+
- superb
|
7 |
+
metrics:
|
8 |
+
- accuracy
|
9 |
+
model-index:
|
10 |
+
- name: wav2vec2-base-ks-linear_lrX100
|
11 |
+
results: []
|
12 |
+
---
|
13 |
+
|
14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
15 |
+
should probably proofread and complete it, then remove this comment. -->
|
16 |
+
|
17 |
+
# wav2vec2-base-ks-linear_lrX100
|
18 |
+
|
19 |
+
This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the superb dataset.
|
20 |
+
It achieves the following results on the evaluation set:
|
21 |
+
- Loss: 0.7289
|
22 |
+
- Accuracy: 0.7870
|
23 |
+
|
24 |
+
## Model description
|
25 |
+
|
26 |
+
More information needed
|
27 |
+
|
28 |
+
## Intended uses & limitations
|
29 |
+
|
30 |
+
More information needed
|
31 |
+
|
32 |
+
## Training and evaluation data
|
33 |
+
|
34 |
+
More information needed
|
35 |
+
|
36 |
+
## Training procedure
|
37 |
+
|
38 |
+
### Training hyperparameters
|
39 |
+
|
40 |
+
The following hyperparameters were used during training:
|
41 |
+
- learning_rate: 0.003
|
42 |
+
- train_batch_size: 256
|
43 |
+
- eval_batch_size: 256
|
44 |
+
- seed: 0
|
45 |
+
- gradient_accumulation_steps: 4
|
46 |
+
- total_train_batch_size: 1024
|
47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
48 |
+
- lr_scheduler_type: linear
|
49 |
+
- lr_scheduler_warmup_ratio: 0.1
|
50 |
+
- num_epochs: 10.0
|
51 |
+
- mixed_precision_training: Native AMP
|
52 |
+
|
53 |
+
### Training results
|
54 |
+
|
55 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
56 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
57 |
+
| 1.1789 | 1.0 | 50 | 1.3621 | 0.6225 |
|
58 |
+
| 0.636 | 2.0 | 100 | 0.9176 | 0.6912 |
|
59 |
+
| 0.5575 | 3.0 | 150 | 0.8543 | 0.7376 |
|
60 |
+
| 0.5289 | 4.0 | 200 | 0.6970 | 0.8001 |
|
61 |
+
| 0.4926 | 5.0 | 250 | 0.8232 | 0.7548 |
|
62 |
+
| 0.4831 | 6.0 | 300 | 0.7442 | 0.7755 |
|
63 |
+
| 0.4539 | 7.0 | 350 | 0.7484 | 0.7785 |
|
64 |
+
| 0.4816 | 8.0 | 400 | 0.7038 | 0.7982 |
|
65 |
+
| 0.4666 | 9.0 | 450 | 0.7277 | 0.7764 |
|
66 |
+
| 0.4417 | 10.0 | 500 | 0.7289 | 0.7870 |
|
67 |
+
|
68 |
+
|
69 |
+
### Framework versions
|
70 |
+
|
71 |
+
- Transformers 4.22.0.dev0
|
72 |
+
- Pytorch 1.11.0+cu115
|
73 |
+
- Datasets 2.4.0
|
74 |
+
- Tokenizers 0.12.1
|