update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc0-1.0
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
model-index:
|
6 |
+
- name: wav2vec2-large-voxrex-npsc-nynorsk
|
7 |
+
results: []
|
8 |
+
---
|
9 |
+
|
10 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
11 |
+
should probably proofread and complete it, then remove this comment. -->
|
12 |
+
|
13 |
+
# wav2vec2-large-voxrex-npsc-nynorsk
|
14 |
+
|
15 |
+
This model is a fine-tuned version of [KBLab/wav2vec2-large-voxrex](https://huggingface.co/KBLab/wav2vec2-large-voxrex) on the None dataset.
|
16 |
+
It achieves the following results on the evaluation set:
|
17 |
+
- Loss: 0.3580
|
18 |
+
- Wer: 0.2155
|
19 |
+
|
20 |
+
## Model description
|
21 |
+
|
22 |
+
More information needed
|
23 |
+
|
24 |
+
## Intended uses & limitations
|
25 |
+
|
26 |
+
More information needed
|
27 |
+
|
28 |
+
## Training and evaluation data
|
29 |
+
|
30 |
+
More information needed
|
31 |
+
|
32 |
+
## Training procedure
|
33 |
+
|
34 |
+
### Training hyperparameters
|
35 |
+
|
36 |
+
The following hyperparameters were used during training:
|
37 |
+
- learning_rate: 7.5e-05
|
38 |
+
- train_batch_size: 16
|
39 |
+
- eval_batch_size: 16
|
40 |
+
- seed: 42
|
41 |
+
- gradient_accumulation_steps: 2
|
42 |
+
- total_train_batch_size: 32
|
43 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
44 |
+
- lr_scheduler_type: linear
|
45 |
+
- lr_scheduler_warmup_steps: 2000
|
46 |
+
- num_epochs: 40.0
|
47 |
+
- mixed_precision_training: Native AMP
|
48 |
+
|
49 |
+
### Training results
|
50 |
+
|
51 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
52 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
53 |
+
| 3.086 | 2.17 | 500 | 3.0773 | 1.0 |
|
54 |
+
| 2.8532 | 4.35 | 1000 | 2.8393 | 1.0 |
|
55 |
+
| 0.9738 | 6.52 | 1500 | 0.7283 | 0.4890 |
|
56 |
+
| 0.6763 | 8.7 | 2000 | 0.5340 | 0.3662 |
|
57 |
+
| 0.5303 | 10.87 | 2500 | 0.4521 | 0.3140 |
|
58 |
+
| 0.4765 | 13.04 | 3000 | 0.4181 | 0.2853 |
|
59 |
+
| 0.4219 | 15.22 | 3500 | 0.4156 | 0.2934 |
|
60 |
+
| 0.3564 | 17.39 | 4000 | 0.3925 | 0.2509 |
|
61 |
+
| 0.3282 | 19.57 | 4500 | 0.3824 | 0.2420 |
|
62 |
+
| 0.3118 | 21.74 | 5000 | 0.3636 | 0.2354 |
|
63 |
+
| 0.2919 | 23.91 | 5500 | 0.3615 | 0.2281 |
|
64 |
+
| 0.2961 | 26.09 | 6000 | 0.3548 | 0.2255 |
|
65 |
+
| 0.284 | 28.26 | 6500 | 0.3526 | 0.2209 |
|
66 |
+
| 0.2566 | 30.43 | 7000 | 0.3526 | 0.2205 |
|
67 |
+
| 0.2422 | 32.61 | 7500 | 0.3569 | 0.2173 |
|
68 |
+
| 0.2472 | 34.78 | 8000 | 0.3592 | 0.2166 |
|
69 |
+
| 0.2337 | 36.96 | 8500 | 0.3625 | 0.2172 |
|
70 |
+
| 0.2315 | 39.13 | 9000 | 0.3580 | 0.2155 |
|
71 |
+
|
72 |
+
|
73 |
+
### Framework versions
|
74 |
+
|
75 |
+
- Transformers 4.17.0.dev0
|
76 |
+
- Pytorch 1.10.0+cu113
|
77 |
+
- Datasets 1.18.3
|
78 |
+
- Tokenizers 0.10.3
|