File size: 4,040 Bytes
3f35920 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 |
---
library_name: transformers
license: cc-by-nc-4.0
base_model: facebook/mms-1b-all
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: mms-1b-bemgen-combined-model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# mms-1b-bemgen-combined-model
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2478
- Wer: 0.3899
## 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: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- 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 |
|:-------------:|:------:|:----:|:---------------:|:------:|
| 6.8762 | 0.0516 | 100 | 0.9801 | 0.9386 |
| 0.5788 | 0.1031 | 200 | 0.3466 | 0.5014 |
| 0.4891 | 0.1547 | 300 | 0.3220 | 0.4820 |
| 0.4386 | 0.2063 | 400 | 0.3071 | 0.4802 |
| 0.4272 | 0.2579 | 500 | 0.3056 | 0.4988 |
| 0.3982 | 0.3094 | 600 | 0.2981 | 0.4626 |
| 0.425 | 0.3610 | 700 | 0.2977 | 0.4631 |
| 0.4036 | 0.4126 | 800 | 0.2897 | 0.4438 |
| 0.3903 | 0.4642 | 900 | 0.2878 | 0.4627 |
| 0.3758 | 0.5157 | 1000 | 0.2926 | 0.4523 |
| 0.3861 | 0.5673 | 1100 | 0.2807 | 0.4410 |
| 0.3763 | 0.6189 | 1200 | 0.2790 | 0.4331 |
| 0.3984 | 0.6704 | 1300 | 0.2803 | 0.4312 |
| 0.373 | 0.7220 | 1400 | 0.2802 | 0.4246 |
| 0.3848 | 0.7736 | 1500 | 0.2759 | 0.4752 |
| 0.4235 | 0.8252 | 1600 | 0.2738 | 0.4268 |
| 0.3704 | 0.8767 | 1700 | 0.2688 | 0.4219 |
| 0.3911 | 0.9283 | 1800 | 0.2653 | 0.4201 |
| 0.3954 | 0.9799 | 1900 | 0.2697 | 0.4482 |
| 0.352 | 1.0315 | 2000 | 0.2654 | 0.4154 |
| 0.3808 | 1.0830 | 2100 | 0.2631 | 0.4051 |
| 0.3681 | 1.1346 | 2200 | 0.2610 | 0.4219 |
| 0.3355 | 1.1862 | 2300 | 0.2608 | 0.4098 |
| 0.342 | 1.2378 | 2400 | 0.2602 | 0.4082 |
| 0.347 | 1.2893 | 2500 | 0.2628 | 0.4055 |
| 0.3409 | 1.3409 | 2600 | 0.2588 | 0.4129 |
| 0.3423 | 1.3925 | 2700 | 0.2617 | 0.4192 |
| 0.3341 | 1.4440 | 2800 | 0.2578 | 0.4055 |
| 0.3425 | 1.4956 | 2900 | 0.2580 | 0.3988 |
| 0.337 | 1.5472 | 3000 | 0.2568 | 0.4071 |
| 0.3412 | 1.5988 | 3100 | 0.2552 | 0.3993 |
| 0.3837 | 1.6503 | 3200 | 0.2622 | 0.4084 |
| 0.3372 | 1.7019 | 3300 | 0.2548 | 0.3991 |
| 0.3394 | 1.7535 | 3400 | 0.2535 | 0.4061 |
| 0.3542 | 1.8051 | 3500 | 0.2512 | 0.3927 |
| 0.3368 | 1.8566 | 3600 | 0.2580 | 0.4004 |
| 0.3807 | 1.9082 | 3700 | 0.2490 | 0.3975 |
| 0.3454 | 1.9598 | 3800 | 0.2514 | 0.4002 |
| 0.3456 | 2.0113 | 3900 | 0.2457 | 0.3931 |
| 0.3202 | 2.0629 | 4000 | 0.2466 | 0.3916 |
| 0.3233 | 2.1145 | 4100 | 0.2495 | 0.3975 |
| 0.3052 | 2.1661 | 4200 | 0.2478 | 0.3899 |
### Framework versions
- Transformers 4.47.1
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0
|