Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
Generated from Trainer
Eval Results (legacy)
# Load model directly
from transformers import AutoProcessor, AutoModelForCTC
processor = AutoProcessor.from_pretrained("misiker/trainer_output")
model = AutoModelForCTC.from_pretrained("misiker/trainer_output", device_map="auto")Quick Links
misiker/trainer_output
This model is a fine-tuned version of facebook/wav2vec2-base on the PolyAI/minds14 dataset. It achieves the following results on the evaluation set:
- Loss: 18.8318
- Wer: 0.9748
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: 1e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- 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: 40
- training_steps: 80
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| No log | 0.8 | 20 | 37.9601 | 1.6562 |
| 41.3656 | 1.6 | 40 | 20.2900 | 0.9755 |
| 18.9017 | 2.4 | 60 | 10.7917 | 0.9734 |
| 18.9017 | 3.2 | 80 | 11.6330 | 0.9734 |
Framework versions
- Transformers 4.52.4
- Pytorch 2.7.1+cpu
- Datasets 3.6.0
- Tokenizers 0.21.1
- Downloads last month
- 1
Model tree for misiker/trainer_output
Base model
facebook/wav2vec2-baseEvaluation results
- Wer on PolyAI/minds14self-reported0.975
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="misiker/trainer_output")