mms-1b-swagen-combined-5hrs-model
This model is a fine-tuned version of facebook/mms-1b-all on the SWAGEN - SWA dataset. It achieves the following results on the evaluation set:
- Loss: 0.2261
- Wer: 0.1896
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
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- 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 |
---|---|---|---|---|
15.7298 | 0.2392 | 100 | 0.9582 | 0.6895 |
0.6897 | 0.4785 | 200 | 0.2673 | 0.1978 |
0.5013 | 0.7177 | 300 | 0.2438 | 0.1978 |
0.5306 | 0.9569 | 400 | 0.2348 | 0.1948 |
0.4662 | 1.1962 | 500 | 0.2367 | 0.1990 |
0.486 | 1.4354 | 600 | 0.2354 | 0.1951 |
0.447 | 1.6746 | 700 | 0.2343 | 0.1942 |
0.4513 | 1.9139 | 800 | 0.2311 | 0.1955 |
0.474 | 2.1531 | 900 | 0.2302 | 0.1940 |
0.451 | 2.3923 | 1000 | 0.2307 | 0.1924 |
0.4616 | 2.6316 | 1100 | 0.2279 | 0.1918 |
0.4489 | 2.8708 | 1200 | 0.2275 | 0.1927 |
0.4452 | 3.1100 | 1300 | 0.2297 | 0.1927 |
0.4023 | 3.3493 | 1400 | 0.2287 | 0.1934 |
0.4471 | 3.5885 | 1500 | 0.2271 | 0.1906 |
0.4471 | 3.8278 | 1600 | 0.2274 | 0.1893 |
0.4386 | 4.0670 | 1700 | 0.2293 | 0.1925 |
0.4281 | 4.3062 | 1800 | 0.2269 | 0.1930 |
0.4058 | 4.5455 | 1900 | 0.2279 | 0.1922 |
0.4246 | 4.7847 | 2000 | 0.2285 | 0.1939 |
0.4457 | 5.0239 | 2100 | 0.2246 | 0.1900 |
0.424 | 5.2632 | 2200 | 0.2258 | 0.1953 |
0.4463 | 5.5024 | 2300 | 0.2237 | 0.1921 |
0.4095 | 5.7416 | 2400 | 0.2279 | 0.1896 |
0.3959 | 5.9809 | 2500 | 0.2265 | 0.1884 |
0.3916 | 6.2201 | 2600 | 0.2275 | 0.1912 |
0.435 | 6.4593 | 2700 | 0.2260 | 0.1896 |
Framework versions
- Transformers 4.47.1
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0
- Downloads last month
- 8
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
Model tree for csikasote/mms-1b-swagen-combined-5hrs-model
Base model
facebook/mms-1b-all