Hugging Face's logo

Rolv-Arild
/
xls-r-300m-npsc-seq2seq

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
speech-encoder-decoder
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
Rolv-Arild commited on
Commit
0020927
•
1 Parent(s): 50254ac

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +98 -0
README.md ADDED
@@ -0,0 +1,98 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_trainer
4
+ model-index:
5
+ - name: ''
6
+ results: []
7
+ ---
8
+
9
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
10
+ should probably proofread and complete it, then remove this comment. -->
11
+
12
+ #
13
+
14
+ This model was trained from scratch on the None dataset.
15
+ It achieves the following results on the evaluation set:
16
+ - Loss: 0.3350
17
+ - Wer: 0.3441
18
+
19
+ ## Model description
20
+
21
+ More information needed
22
+
23
+ ## Intended uses & limitations
24
+
25
+ More information needed
26
+
27
+ ## Training and evaluation data
28
+
29
+ More information needed
30
+
31
+ ## Training procedure
32
+
33
+ ### Training hyperparameters
34
+
35
+ The following hyperparameters were used during training:
36
+ - learning_rate: 0.0003
37
+ - train_batch_size: 8
38
+ - eval_batch_size: 8
39
+ - seed: 42
40
+ - gradient_accumulation_steps: 8
41
+ - total_train_batch_size: 64
42
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
43
+ - lr_scheduler_type: linear
44
+ - lr_scheduler_warmup_steps: 400
45
+ - num_epochs: 20.0
46
+ - mixed_precision_training: Native AMP
47
+
48
+ ### Training results
49
+
50
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
51
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
52
+ | 2.2943 | 0.51 | 400 | 4.4367 | 0.9745 |
53
+ | 2.8816 | 1.02 | 800 | 2.8103 | 1.2576 |
54
+ | 2.7842 | 1.53 | 1200 | 2.6832 | 1.1330 |
55
+ | 2.313 | 2.04 | 1600 | 2.6678 | 1.1348 |
56
+ | 2.2897 | 2.55 | 2000 | 2.5744 | 1.2102 |
57
+ | 1.3821 | 3.06 | 2400 | 2.9908 | 0.9377 |
58
+ | 1.074 | 3.57 | 2800 | 2.6649 | 0.8966 |
59
+ | 1.3643 | 4.08 | 3200 | 1.0064 | 0.7236 |
60
+ | 0.8286 | 4.59 | 3600 | 0.6339 | 0.5454 |
61
+ | 0.3872 | 5.1 | 4000 | 0.5170 | 0.4718 |
62
+ | 0.3654 | 5.61 | 4400 | 0.4386 | 0.4420 |
63
+ | 0.2672 | 6.12 | 4800 | 0.5186 | 0.4679 |
64
+ | 0.2519 | 6.63 | 5200 | 0.4238 | 0.4177 |
65
+ | 0.3293 | 7.14 | 5600 | 0.3584 | 0.3970 |
66
+ | 0.314 | 7.65 | 6000 | 0.3325 | 0.3911 |
67
+ | 0.1698 | 8.16 | 6400 | 0.3411 | 0.3855 |
68
+ | 0.1682 | 8.67 | 6800 | 0.3239 | 0.3801 |
69
+ | 0.1325 | 9.18 | 7200 | 0.3474 | 0.3832 |
70
+ | 0.1577 | 9.69 | 7600 | 0.3289 | 0.3839 |
71
+ | 0.2259 | 10.2 | 8000 | 0.3183 | 0.3756 |
72
+ | 0.2473 | 10.71 | 8400 | 0.3132 | 0.3654 |
73
+ | 0.1136 | 11.22 | 8800 | 0.3242 | 0.3670 |
74
+ | 0.108 | 11.73 | 9200 | 0.3201 | 0.3601 |
75
+ | 0.0806 | 12.24 | 9600 | 0.3223 | 0.3609 |
76
+ | 0.0896 | 12.75 | 10000 | 0.3228 | 0.3584 |
77
+ | 0.1642 | 13.27 | 10400 | 0.3140 | 0.3547 |
78
+ | 0.1442 | 13.77 | 10800 | 0.3235 | 0.3539 |
79
+ | 0.0802 | 14.29 | 11200 | 0.3175 | 0.3553 |
80
+ | 0.0747 | 14.8 | 11600 | 0.3126 | 0.3512 |
81
+ | 0.0488 | 15.31 | 12000 | 0.3292 | 0.3525 |
82
+ | 0.0469 | 15.82 | 12400 | 0.3231 | 0.3504 |
83
+ | 0.1021 | 16.33 | 12800 | 0.3230 | 0.3502 |
84
+ | 0.0841 | 16.84 | 13200 | 0.3348 | 0.3513 |
85
+ | 0.0502 | 17.35 | 13600 | 0.3318 | 0.3486 |
86
+ | 0.059 | 17.86 | 14000 | 0.3359 | 0.3462 |
87
+ | 0.0417 | 18.37 | 14400 | 0.3310 | 0.3467 |
88
+ | 0.0322 | 18.88 | 14800 | 0.3325 | 0.3467 |
89
+ | 0.0531 | 19.39 | 15200 | 0.3357 | 0.3449 |
90
+ | 0.0623 | 19.9 | 15600 | 0.3350 | 0.3441 |
91
+
92
+
93
+ ### Framework versions
94
+
95
+ - Transformers 4.17.0.dev0
96
+ - Pytorch 1.10.0+cu113
97
+ - Datasets 1.18.1
98
+ - Tokenizers 0.11.0