Commit
•
ad9f351
1
Parent(s):
a4c8c46
update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,120 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- generated_from_trainer
|
4 |
+
datasets:
|
5 |
+
- fleurs
|
6 |
+
metrics:
|
7 |
+
- wer
|
8 |
+
model-index:
|
9 |
+
- name: microsoft-wavlm-fleurs-ur
|
10 |
+
results:
|
11 |
+
- task:
|
12 |
+
name: Automatic Speech Recognition
|
13 |
+
type: automatic-speech-recognition
|
14 |
+
dataset:
|
15 |
+
name: fleurs
|
16 |
+
type: fleurs
|
17 |
+
config: ur_pk
|
18 |
+
split: test
|
19 |
+
args: ur_pk
|
20 |
+
metrics:
|
21 |
+
- name: Wer
|
22 |
+
type: wer
|
23 |
+
value: 0.4026467344688151
|
24 |
+
---
|
25 |
+
|
26 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
27 |
+
should probably proofread and complete it, then remove this comment. -->
|
28 |
+
|
29 |
+
# microsoft-wavlm-fleurs-ur
|
30 |
+
|
31 |
+
This model is a fine-tuned version of [microsoft/wavlm-large](https://huggingface.co/microsoft/wavlm-large) on the fleurs dataset.
|
32 |
+
It achieves the following results on the evaluation set:
|
33 |
+
- Loss: 0.7294
|
34 |
+
- Wer: 0.4026
|
35 |
+
|
36 |
+
## Model description
|
37 |
+
|
38 |
+
More information needed
|
39 |
+
|
40 |
+
## Intended uses & limitations
|
41 |
+
|
42 |
+
More information needed
|
43 |
+
|
44 |
+
## Training and evaluation data
|
45 |
+
|
46 |
+
More information needed
|
47 |
+
|
48 |
+
## Training procedure
|
49 |
+
|
50 |
+
### Training hyperparameters
|
51 |
+
|
52 |
+
The following hyperparameters were used during training:
|
53 |
+
- learning_rate: 0.0003
|
54 |
+
- train_batch_size: 4
|
55 |
+
- eval_batch_size: 4
|
56 |
+
- seed: 42
|
57 |
+
- distributed_type: multi-GPU
|
58 |
+
- num_devices: 2
|
59 |
+
- total_train_batch_size: 8
|
60 |
+
- total_eval_batch_size: 8
|
61 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
62 |
+
- lr_scheduler_type: linear
|
63 |
+
- lr_scheduler_warmup_steps: 500
|
64 |
+
- num_epochs: 15.0
|
65 |
+
- mixed_precision_training: Native AMP
|
66 |
+
|
67 |
+
### Training results
|
68 |
+
|
69 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
70 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
71 |
+
| 3.911 | 0.35 | 100 | 3.7784 | 1.0 |
|
72 |
+
| 3.0833 | 0.71 | 200 | 3.0964 | 1.0 |
|
73 |
+
| 3.028 | 1.06 | 300 | 3.0377 | 1.0 |
|
74 |
+
| 2.5114 | 1.41 | 400 | 2.4941 | 0.9922 |
|
75 |
+
| 1.0583 | 1.77 | 500 | 1.0753 | 0.7579 |
|
76 |
+
| 0.715 | 2.12 | 600 | 0.8524 | 0.6410 |
|
77 |
+
| 0.6779 | 2.47 | 700 | 0.7711 | 0.6063 |
|
78 |
+
| 0.6123 | 2.83 | 800 | 0.7170 | 0.5706 |
|
79 |
+
| 0.8183 | 3.18 | 900 | 0.6897 | 0.5368 |
|
80 |
+
| 0.5195 | 3.53 | 1000 | 0.6586 | 0.5303 |
|
81 |
+
| 0.4774 | 3.89 | 1100 | 0.6306 | 0.5014 |
|
82 |
+
| 0.4242 | 4.24 | 1200 | 0.6138 | 0.4817 |
|
83 |
+
| 0.4549 | 4.59 | 1300 | 0.6027 | 0.4678 |
|
84 |
+
| 0.2576 | 4.95 | 1400 | 0.5878 | 0.4600 |
|
85 |
+
| 0.1578 | 5.3 | 1500 | 0.6144 | 0.4585 |
|
86 |
+
| 0.3556 | 5.65 | 1600 | 0.5884 | 0.4582 |
|
87 |
+
| 0.2427 | 6.01 | 1700 | 0.6071 | 0.4572 |
|
88 |
+
| 0.267 | 6.36 | 1800 | 0.6303 | 0.4514 |
|
89 |
+
| 0.2468 | 6.71 | 1900 | 0.6358 | 0.4495 |
|
90 |
+
| 0.159 | 7.07 | 2000 | 0.6242 | 0.4312 |
|
91 |
+
| 0.1527 | 7.42 | 2100 | 0.6372 | 0.4400 |
|
92 |
+
| 0.1401 | 7.77 | 2200 | 0.6252 | 0.4292 |
|
93 |
+
| 0.1211 | 8.13 | 2300 | 0.6358 | 0.4251 |
|
94 |
+
| 0.1022 | 8.48 | 2400 | 0.6529 | 0.4356 |
|
95 |
+
| 0.0818 | 8.83 | 2500 | 0.6773 | 0.4200 |
|
96 |
+
| 0.0918 | 9.19 | 2600 | 0.6879 | 0.4267 |
|
97 |
+
| 0.119 | 9.54 | 2700 | 0.6948 | 0.4254 |
|
98 |
+
| 0.1615 | 9.89 | 2800 | 0.6920 | 0.4259 |
|
99 |
+
| 0.0953 | 10.25 | 2900 | 0.7019 | 0.4218 |
|
100 |
+
| 0.1008 | 10.6 | 3000 | 0.6933 | 0.4133 |
|
101 |
+
| 0.0729 | 10.95 | 3100 | 0.6950 | 0.4164 |
|
102 |
+
| 0.0636 | 11.31 | 3200 | 0.7151 | 0.4121 |
|
103 |
+
| 0.0395 | 11.66 | 3300 | 0.7053 | 0.4098 |
|
104 |
+
| 0.0391 | 12.01 | 3400 | 0.7081 | 0.3984 |
|
105 |
+
| 0.0507 | 12.37 | 3500 | 0.7012 | 0.4111 |
|
106 |
+
| 0.0598 | 12.72 | 3600 | 0.7169 | 0.4035 |
|
107 |
+
| 0.0515 | 13.07 | 3700 | 0.7358 | 0.4102 |
|
108 |
+
| 0.0429 | 13.43 | 3800 | 0.7236 | 0.4013 |
|
109 |
+
| 0.0398 | 13.78 | 3900 | 0.7404 | 0.4026 |
|
110 |
+
| 0.0946 | 14.13 | 4000 | 0.7285 | 0.4029 |
|
111 |
+
| 0.0428 | 14.49 | 4100 | 0.7271 | 0.3991 |
|
112 |
+
| 0.0329 | 14.84 | 4200 | 0.7294 | 0.4026 |
|
113 |
+
|
114 |
+
|
115 |
+
### Framework versions
|
116 |
+
|
117 |
+
- Transformers 4.27.0.dev0
|
118 |
+
- Pytorch 1.13.1
|
119 |
+
- Datasets 2.8.0
|
120 |
+
- Tokenizers 0.13.2
|