|
--- |
|
license: apache-2.0 |
|
base_model: facebook/wav2vec2-xls-r-300m |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- common_voice_17_0 |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: xls-r-300m-hbs-ru-unfrozen-batch16 |
|
results: |
|
- task: |
|
name: Automatic Speech Recognition |
|
type: automatic-speech-recognition |
|
dataset: |
|
name: common_voice_17_0 |
|
type: common_voice_17_0 |
|
config: hsb |
|
split: test |
|
args: hsb |
|
metrics: |
|
- name: Wer |
|
type: wer |
|
value: 0.37207122774133083 |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/badr-nlp/xlsr-continual-finetuning/runs/dgxuea1c) |
|
# xls-r-300m-hbs-ru-unfrozen-batch16 |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice_17_0 dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.6191 |
|
- Wer: 0.3721 |
|
- Cer: 0.0853 |
|
|
|
## 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: 16 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 32 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 500 |
|
- num_epochs: 100 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer | |
|
|:-------------:|:-------:|:----:|:---------------:|:------:|:------:| |
|
| 3.3829 | 3.2258 | 100 | 3.3113 | 1.0 | 1.0 | |
|
| 3.0722 | 6.4516 | 200 | 3.0062 | 1.0 | 0.9991 | |
|
| 0.5001 | 9.6774 | 300 | 0.6462 | 0.6396 | 0.1553 | |
|
| 0.2668 | 12.9032 | 400 | 0.5761 | 0.5567 | 0.1386 | |
|
| 0.1468 | 16.1290 | 500 | 0.5573 | 0.4986 | 0.1192 | |
|
| 0.1351 | 19.3548 | 600 | 0.5716 | 0.4862 | 0.1139 | |
|
| 0.1263 | 22.5806 | 700 | 0.5959 | 0.4841 | 0.1178 | |
|
| 0.094 | 25.8065 | 800 | 0.5752 | 0.4391 | 0.1024 | |
|
| 0.0473 | 29.0323 | 900 | 0.6015 | 0.4445 | 0.1059 | |
|
| 0.0442 | 32.2581 | 1000 | 0.6266 | 0.4616 | 0.1127 | |
|
| 0.0727 | 35.4839 | 1100 | 0.6193 | 0.4442 | 0.1069 | |
|
| 0.0494 | 38.7097 | 1200 | 0.6244 | 0.4349 | 0.1023 | |
|
| 0.027 | 41.9355 | 1300 | 0.6457 | 0.4391 | 0.1038 | |
|
| 0.0277 | 45.1613 | 1400 | 0.6470 | 0.4351 | 0.1045 | |
|
| 0.0326 | 48.3871 | 1500 | 0.6137 | 0.4093 | 0.0986 | |
|
| 0.0511 | 51.6129 | 1600 | 0.6152 | 0.4182 | 0.0975 | |
|
| 0.0431 | 54.8387 | 1700 | 0.5967 | 0.4210 | 0.1011 | |
|
| 0.0749 | 58.0645 | 1800 | 0.6173 | 0.4276 | 0.1034 | |
|
| 0.032 | 61.2903 | 1900 | 0.6318 | 0.4201 | 0.0990 | |
|
| 0.0504 | 64.5161 | 2000 | 0.6174 | 0.4227 | 0.0999 | |
|
| 0.0308 | 67.7419 | 2100 | 0.6174 | 0.4007 | 0.0937 | |
|
| 0.0301 | 70.9677 | 2200 | 0.6148 | 0.3962 | 0.0923 | |
|
| 0.0178 | 74.1935 | 2300 | 0.6038 | 0.4044 | 0.0945 | |
|
| 0.018 | 77.4194 | 2400 | 0.5975 | 0.3878 | 0.0912 | |
|
| 0.0112 | 80.6452 | 2500 | 0.6183 | 0.3913 | 0.0927 | |
|
| 0.0432 | 83.8710 | 2600 | 0.6346 | 0.3845 | 0.0905 | |
|
| 0.0327 | 87.0968 | 2700 | 0.6327 | 0.3793 | 0.0877 | |
|
| 0.0254 | 90.3226 | 2800 | 0.6270 | 0.3770 | 0.0882 | |
|
| 0.0199 | 93.5484 | 2900 | 0.6250 | 0.3751 | 0.0868 | |
|
| 0.0147 | 96.7742 | 3000 | 0.6222 | 0.3709 | 0.0855 | |
|
| 0.0025 | 100.0 | 3100 | 0.6191 | 0.3721 | 0.0853 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.42.0.dev0 |
|
- Pytorch 2.3.1+cu121 |
|
- Datasets 2.19.2 |
|
- Tokenizers 0.19.1 |
|
|