wav2vec2-base-finetune-vi-v3
This model is a fine-tuned version of nguyenvulebinh/wav2vec2-base-vietnamese-250h on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.2019
- Wer: 0.1347
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 30
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
4.3444 | 0.59 | 500 | 3.0138 | 0.9938 |
1.6923 | 1.19 | 1000 | 0.6549 | 0.3873 |
0.8371 | 1.78 | 1500 | 0.3958 | 0.2492 |
0.6161 | 2.37 | 2000 | 0.3272 | 0.2161 |
0.5443 | 2.97 | 2500 | 0.2871 | 0.1967 |
0.4361 | 3.56 | 3000 | 0.2928 | 0.1852 |
0.3889 | 4.15 | 3500 | 0.2957 | 0.1745 |
0.3701 | 4.74 | 4000 | 0.2611 | 0.1702 |
0.353 | 5.34 | 4500 | 0.2514 | 0.1685 |
0.3299 | 5.93 | 5000 | 0.2239 | 0.1612 |
0.3235 | 6.52 | 5500 | 0.2362 | 0.1673 |
0.3 | 7.12 | 6000 | 0.2267 | 0.1593 |
0.285 | 7.71 | 6500 | 0.2397 | 0.1570 |
0.2766 | 8.3 | 7000 | 0.2415 | 0.1538 |
0.2652 | 8.9 | 7500 | 0.2265 | 0.1552 |
0.2396 | 9.49 | 8000 | 0.2299 | 0.1516 |
0.2442 | 10.08 | 8500 | 0.2136 | 0.1538 |
0.242 | 10.68 | 9000 | 0.2120 | 0.1510 |
0.2149 | 11.27 | 9500 | 0.2250 | 0.1494 |
0.2186 | 11.86 | 10000 | 0.2077 | 0.1491 |
0.2157 | 12.46 | 10500 | 0.2198 | 0.1461 |
0.2154 | 13.05 | 11000 | 0.2048 | 0.1484 |
0.2115 | 13.64 | 11500 | 0.2209 | 0.1444 |
0.2072 | 14.23 | 12000 | 0.2014 | 0.1414 |
0.1974 | 14.83 | 12500 | 0.2256 | 0.1451 |
0.1888 | 15.42 | 13000 | 0.1931 | 0.1445 |
0.1802 | 16.01 | 13500 | 0.2089 | 0.1452 |
0.1747 | 16.61 | 14000 | 0.2136 | 0.1441 |
0.1585 | 17.2 | 14500 | 0.2102 | 0.1411 |
0.1593 | 17.79 | 15000 | 0.1941 | 0.1413 |
0.1662 | 18.39 | 15500 | 0.2046 | 0.1426 |
0.165 | 18.98 | 16000 | 0.2077 | 0.1432 |
0.1696 | 19.57 | 16500 | 0.2062 | 0.1406 |
0.1499 | 20.17 | 17000 | 0.1997 | 0.1406 |
0.1459 | 20.76 | 17500 | 0.2063 | 0.1418 |
0.1381 | 21.35 | 18000 | 0.2040 | 0.1374 |
0.1504 | 21.95 | 18500 | 0.2013 | 0.1387 |
0.14 | 22.54 | 19000 | 0.2030 | 0.1393 |
0.1395 | 23.13 | 19500 | 0.2019 | 0.1357 |
0.1265 | 23.72 | 20000 | 0.2023 | 0.1364 |
0.1425 | 24.32 | 20500 | 0.1995 | 0.1362 |
0.1196 | 24.91 | 21000 | 0.2033 | 0.1368 |
0.1251 | 25.5 | 21500 | 0.2013 | 0.1357 |
0.1175 | 26.1 | 22000 | 0.2056 | 0.1360 |
0.1278 | 26.69 | 22500 | 0.2026 | 0.1353 |
0.1223 | 27.28 | 23000 | 0.1985 | 0.1365 |
0.1195 | 27.88 | 23500 | 0.2004 | 0.1358 |
0.1167 | 28.47 | 24000 | 0.1956 | 0.1364 |
0.1167 | 29.06 | 24500 | 0.2004 | 0.1350 |
0.1143 | 29.66 | 25000 | 0.2019 | 0.1347 |
Framework versions
- Transformers 4.30.2
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.13.3
- Downloads last month
- 64
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.