metadata
language:
- eng
license: apache-2.0
tags:
- '[finetuned_model, lj_speech11]'
- generated_from_trainer
base_model: facebook/wav2vec2-base-960h
datasets:
- FYP/LJ-SpeechLJ
model-index:
- name: SpeechT5 STT Wav2Vec2
results: []
SpeechT5 STT Wav2Vec2
This model is a fine-tuned version of facebook/wav2vec2-base-960h on the Lj-Speech dataset. It achieves the following results on the evaluation set:
- Loss: 491.2500
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: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- 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: 100
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 494.9709 | 0.3795 | 50 | 486.0701 |
| 534.2352 | 0.7590 | 100 | 488.7172 |
| 816.0739 | 1.1385 | 150 | 490.4418 |
| 566.1295 | 1.5180 | 200 | 504.5211 |
| 586.0909 | 1.8975 | 250 | 489.5141 |
| 601.5043 | 2.2770 | 300 | 486.6875 |
| 487.8737 | 2.6565 | 350 | 489.5807 |
| 1145.4591 | 3.0361 | 400 | 511.4276 |
| 686.6008 | 3.4156 | 450 | 496.0722 |
| 664.612 | 3.7951 | 500 | 486.9992 |
| 630.4309 | 4.1746 | 550 | 500.0555 |
| 513.7977 | 4.5541 | 600 | 488.6891 |
| 494.3428 | 4.9336 | 650 | 491.2500 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1