wav2vec2-large-mms-1b-turkish-colab
This model is a fine-tuned version of facebook/mms-1b-all on the common_voice_11_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.1588
- Wer: 0.2274
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: 0.001
- train_batch_size: 32
- 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: 100
- num_epochs: 4
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
7.9572 | 0.09 | 100 | 0.2279 | 0.2821 |
0.316 | 0.18 | 200 | 0.2011 | 0.2632 |
0.282 | 0.27 | 300 | 0.2027 | 0.2555 |
0.285 | 0.35 | 400 | 0.1978 | 0.2580 |
0.2741 | 0.44 | 500 | 0.1956 | 0.2596 |
0.2643 | 0.53 | 600 | 0.1790 | 0.2487 |
0.2758 | 0.62 | 700 | 0.1791 | 0.2463 |
0.2766 | 0.71 | 800 | 0.1791 | 0.2499 |
0.2733 | 0.8 | 900 | 0.1854 | 0.2610 |
0.2611 | 0.89 | 1000 | 0.1793 | 0.2464 |
0.2582 | 0.97 | 1100 | 0.1734 | 0.2460 |
0.2421 | 1.06 | 1200 | 0.1711 | 0.2408 |
0.2424 | 1.15 | 1300 | 0.1737 | 0.2434 |
0.2455 | 1.24 | 1400 | 0.1761 | 0.2489 |
0.2588 | 1.33 | 1500 | 0.1720 | 0.2410 |
0.2591 | 1.42 | 1600 | 0.1761 | 0.2444 |
0.2497 | 1.51 | 1700 | 0.1696 | 0.2381 |
0.254 | 1.59 | 1800 | 0.1728 | 0.2391 |
0.2479 | 1.68 | 1900 | 0.1724 | 0.2402 |
0.2395 | 1.77 | 2000 | 0.1726 | 0.2389 |
0.237 | 1.86 | 2100 | 0.1710 | 0.2378 |
0.2427 | 1.95 | 2200 | 0.1682 | 0.2348 |
0.2399 | 2.04 | 2300 | 0.1699 | 0.2371 |
0.2457 | 2.13 | 2400 | 0.1695 | 0.2357 |
0.2432 | 2.21 | 2500 | 0.1707 | 0.2387 |
0.229 | 2.3 | 2600 | 0.1687 | 0.2324 |
0.2413 | 2.39 | 2700 | 0.1681 | 0.2354 |
0.2286 | 2.48 | 2800 | 0.1664 | 0.2329 |
0.2405 | 2.57 | 2900 | 0.1646 | 0.2337 |
0.2266 | 2.66 | 3000 | 0.1668 | 0.2341 |
0.2337 | 2.75 | 3100 | 0.1642 | 0.2325 |
0.233 | 2.83 | 3200 | 0.1635 | 0.2301 |
0.2235 | 2.92 | 3300 | 0.1639 | 0.2342 |
0.2395 | 3.01 | 3400 | 0.1630 | 0.2305 |
0.2165 | 3.1 | 3500 | 0.1622 | 0.2305 |
0.2258 | 3.19 | 3600 | 0.1617 | 0.2296 |
0.2288 | 3.28 | 3700 | 0.1608 | 0.2307 |
0.218 | 3.37 | 3800 | 0.1610 | 0.2301 |
0.2242 | 3.45 | 3900 | 0.1604 | 0.2304 |
0.2248 | 3.54 | 4000 | 0.1603 | 0.2273 |
0.2223 | 3.63 | 4100 | 0.1595 | 0.2282 |
0.2161 | 3.72 | 4200 | 0.1593 | 0.2283 |
0.2281 | 3.81 | 4300 | 0.1592 | 0.2278 |
0.2236 | 3.9 | 4400 | 0.1593 | 0.2281 |
0.2277 | 3.99 | 4500 | 0.1588 | 0.2274 |
Framework versions
- Transformers 4.31.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
- Downloads last month
- 8
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.