metadata
library_name: transformers
license: cc-by-nc-4.0
base_model: facebook/mms-1b-all
tags:
- automatic-speech-recognition
- genbed
- mms
- generated_from_trainer
metrics:
- wer
model-index:
- name: mms-1b-all-bem-genbed-combined-adapter-test
results: []
mms-1b-all-bem-genbed-combined-adapter-test
This model is a fine-tuned version of facebook/mms-1b-all on the GENBED - BEM dataset. It achieves the following results on the evaluation set:
- Loss: 0.2293
- Wer: 0.3752
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: 0.0003
- 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: 100
- num_epochs: 30.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 6.4783 | 0.1374 | 100 | 0.6403 | 0.69 |
| 0.5413 | 0.2747 | 200 | 0.3260 | 0.4931 |
| 0.437 | 0.4121 | 300 | 0.3014 | 0.4678 |
| 0.4402 | 0.5495 | 400 | 0.2982 | 0.4818 |
| 0.4153 | 0.6868 | 500 | 0.2936 | 0.4702 |
| 0.4154 | 0.8242 | 600 | 0.2884 | 0.4493 |
| 0.3789 | 0.9615 | 700 | 0.2806 | 0.4521 |
| 0.3667 | 1.0989 | 800 | 0.2764 | 0.4352 |
| 0.3929 | 1.2363 | 900 | 0.2763 | 0.4654 |
| 0.3847 | 1.3736 | 1000 | 0.2705 | 0.4406 |
| 0.3833 | 1.5110 | 1100 | 0.2697 | 0.4246 |
| 0.3742 | 1.6484 | 1200 | 0.2668 | 0.4250 |
| 0.3694 | 1.7857 | 1300 | 0.2690 | 0.4189 |
| 0.3494 | 1.9231 | 1400 | 0.2635 | 0.416 |
| 0.3724 | 2.0604 | 1500 | 0.2626 | 0.4323 |
| 0.3723 | 2.1978 | 1600 | 0.2598 | 0.4247 |
| 0.3505 | 2.3352 | 1700 | 0.2583 | 0.412 |
| 0.3393 | 2.4725 | 1800 | 0.2563 | 0.4128 |
| 0.3352 | 2.6099 | 1900 | 0.2545 | 0.4156 |
| 0.3516 | 2.7473 | 2000 | 0.2551 | 0.4315 |
| 0.3489 | 2.8846 | 2100 | 0.2560 | 0.4270 |
| 0.3512 | 3.0220 | 2200 | 0.2536 | 0.4039 |
| 0.339 | 3.1593 | 2300 | 0.2490 | 0.3989 |
| 0.3374 | 3.2967 | 2400 | 0.2495 | 0.3964 |
| 0.3295 | 3.4341 | 2500 | 0.2518 | 0.4037 |
| 0.3391 | 3.5714 | 2600 | 0.2491 | 0.4077 |
| 0.3373 | 3.7088 | 2700 | 0.2445 | 0.3989 |
| 0.3097 | 3.8462 | 2800 | 0.2462 | 0.4118 |
| 0.3458 | 3.9835 | 2900 | 0.2443 | 0.4034 |
| 0.313 | 4.1209 | 3000 | 0.2433 | 0.3882 |
| 0.3171 | 4.2582 | 3100 | 0.2426 | 0.3968 |
| 0.3122 | 4.3956 | 3200 | 0.2430 | 0.3936 |
| 0.3255 | 4.5330 | 3300 | 0.2404 | 0.3822 |
| 0.3253 | 4.6703 | 3400 | 0.2356 | 0.3946 |
| 0.3341 | 4.8077 | 3500 | 0.2369 | 0.3872 |
| 0.3183 | 4.9451 | 3600 | 0.2345 | 0.3854 |
| 0.3461 | 5.0824 | 3700 | 0.2395 | 0.3828 |
| 0.3147 | 5.2198 | 3800 | 0.2359 | 0.3775 |
| 0.317 | 5.3571 | 3900 | 0.2320 | 0.3808 |
| 0.3094 | 5.4945 | 4000 | 0.2366 | 0.3797 |
| 0.2913 | 5.6319 | 4100 | 0.2357 | 0.3749 |
| 0.3195 | 5.7692 | 4200 | 0.2332 | 0.3694 |
| 0.3189 | 5.9066 | 4300 | 0.2313 | 0.3870 |
| 0.3105 | 6.0440 | 4400 | 0.2326 | 0.3806 |
| 0.2937 | 6.1813 | 4500 | 0.2346 | 0.3784 |
| 0.3088 | 6.3187 | 4600 | 0.2313 | 0.3726 |
| 0.2852 | 6.4560 | 4700 | 0.2307 | 0.3709 |
| 0.3083 | 6.5934 | 4800 | 0.2293 | 0.3752 |
| 0.3194 | 6.7308 | 4900 | 0.2292 | 0.3711 |
| 0.297 | 6.8681 | 5000 | 0.2303 | 0.3715 |
| 0.3086 | 7.0055 | 5100 | 0.2340 | 0.4027 |
| 0.3058 | 7.1429 | 5200 | 0.2294 | 0.3665 |
Framework versions
- Transformers 4.46.0.dev0
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0