Commit
·
32ea87c
1
Parent(s):
db64d3a
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
model-index:
|
| 6 |
+
- name: wav2vec2-tcrs
|
| 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-tcrs
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned version of [facebook/wav2vec2-large-lv60](https://huggingface.co/facebook/wav2vec2-large-lv60) on the None dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 2.9550
|
| 18 |
+
- Wer: 1.0657
|
| 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: 0.0001
|
| 38 |
+
- train_batch_size: 1
|
| 39 |
+
- eval_batch_size: 8
|
| 40 |
+
- seed: 42
|
| 41 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 42 |
+
- lr_scheduler_type: linear
|
| 43 |
+
- lr_scheduler_warmup_steps: 1000
|
| 44 |
+
- num_epochs: 100
|
| 45 |
+
- mixed_precision_training: Native AMP
|
| 46 |
+
|
| 47 |
+
### Training results
|
| 48 |
+
|
| 49 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 50 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 51 |
+
| 13.6613 | 3.38 | 500 | 3.2415 | 1.0 |
|
| 52 |
+
| 2.9524 | 6.76 | 1000 | 3.0199 | 1.0 |
|
| 53 |
+
| 2.9425 | 10.14 | 1500 | 3.0673 | 1.0 |
|
| 54 |
+
| 2.9387 | 13.51 | 2000 | 3.0151 | 1.0 |
|
| 55 |
+
| 2.9384 | 16.89 | 2500 | 3.0320 | 1.0 |
|
| 56 |
+
| 2.929 | 20.27 | 3000 | 2.9691 | 1.0 |
|
| 57 |
+
| 2.9194 | 23.65 | 3500 | 2.9596 | 1.0 |
|
| 58 |
+
| 2.9079 | 27.03 | 4000 | 2.9279 | 1.0 |
|
| 59 |
+
| 2.8957 | 30.41 | 4500 | 2.9647 | 1.0 |
|
| 60 |
+
| 2.8385 | 33.78 | 5000 | 2.8114 | 1.0193 |
|
| 61 |
+
| 2.6546 | 37.16 | 5500 | 2.6744 | 1.0983 |
|
| 62 |
+
| 2.5866 | 40.54 | 6000 | 2.6192 | 1.1071 |
|
| 63 |
+
| 2.5475 | 43.92 | 6500 | 2.5777 | 1.0950 |
|
| 64 |
+
| 2.5177 | 47.3 | 7000 | 2.5845 | 1.1220 |
|
| 65 |
+
| 2.482 | 50.68 | 7500 | 2.5730 | 1.1264 |
|
| 66 |
+
| 2.4343 | 54.05 | 8000 | 2.5722 | 1.0955 |
|
| 67 |
+
| 2.3754 | 57.43 | 8500 | 2.5781 | 1.1353 |
|
| 68 |
+
| 2.3055 | 60.81 | 9000 | 2.6177 | 1.0972 |
|
| 69 |
+
| 2.2446 | 64.19 | 9500 | 2.6351 | 1.1027 |
|
| 70 |
+
| 2.1625 | 67.57 | 10000 | 2.6924 | 1.0756 |
|
| 71 |
+
| 2.1078 | 70.95 | 10500 | 2.6817 | 1.0795 |
|
| 72 |
+
| 2.0366 | 74.32 | 11000 | 2.7629 | 1.0657 |
|
| 73 |
+
| 1.9899 | 77.7 | 11500 | 2.7972 | 1.0845 |
|
| 74 |
+
| 1.9309 | 81.08 | 12000 | 2.8450 | 1.0734 |
|
| 75 |
+
| 1.8861 | 84.46 | 12500 | 2.8703 | 1.0668 |
|
| 76 |
+
| 1.8437 | 87.84 | 13000 | 2.9308 | 1.0917 |
|
| 77 |
+
| 1.8192 | 91.22 | 13500 | 2.9298 | 1.0701 |
|
| 78 |
+
| 1.7952 | 94.59 | 14000 | 2.9488 | 1.0685 |
|
| 79 |
+
| 1.7745 | 97.97 | 14500 | 2.9550 | 1.0657 |
|
| 80 |
+
|
| 81 |
+
|
| 82 |
+
### Framework versions
|
| 83 |
+
|
| 84 |
+
- Transformers 4.11.3
|
| 85 |
+
- Pytorch 1.9.1
|
| 86 |
+
- Datasets 1.18.3
|
| 87 |
+
- Tokenizers 0.10.3
|