# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForCTC
processor = AutoProcessor.from_pretrained("amin-oj/wav2vec2-base-960h-finetuned-asr-PolyAI_minds14-en-US")
model = AutoModelForCTC.from_pretrained("amin-oj/wav2vec2-base-960h-finetuned-asr-PolyAI_minds14-en-US", device_map="auto")Quick Links
wav2vec2-base-960h-finetuned-asr-PolyAI_minds14-en-US
This model is a fine-tuned version of facebook/wav2vec2-base-960h on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.6485
- Wer: 0.3319
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 125
- training_steps: 500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.5518 | 12.4 | 62 | 1.5976 | 0.4723 |
| 1.0477 | 24.8 | 124 | 1.1343 | 0.3617 |
| 0.7908 | 37.2 | 186 | 1.5092 | 0.3447 |
| 0.6269 | 49.6 | 248 | 1.5146 | 0.3617 |
| 0.5075 | 62.0 | 310 | 1.4677 | 0.3574 |
| 0.4534 | 74.4 | 372 | 1.5825 | 0.3191 |
| 0.4621 | 86.8 | 434 | 1.6253 | 0.3191 |
| 0.4314 | 99.2 | 496 | 1.6485 | 0.3319 |
Framework versions
- Transformers 4.57.1
- Pytorch 2.8.0+cu126
- Datasets 4.4.2
- Tokenizers 0.22.1
- Downloads last month
- 9
Model tree for amin-oj/wav2vec2-base-960h-finetuned-asr-PolyAI_minds14-en-US
Base model
facebook/wav2vec2-base-960h
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="amin-oj/wav2vec2-base-960h-finetuned-asr-PolyAI_minds14-en-US")