Automatic Speech Recognition
Transformers
PyTorch
TensorFlow
Safetensors
English
wav2vec2
audio
hf-asr-leaderboard
Eval Results (legacy)
Eval Results
Instructions to use facebook/wav2vec2-base-960h with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use facebook/wav2vec2-base-960h with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-base-960h")# Load model directly from transformers import AutoProcessor, AutoModelForCTC processor = AutoProcessor.from_pretrained("facebook/wav2vec2-base-960h") model = AutoModelForCTC.from_pretrained("facebook/wav2vec2-base-960h") - Notebooks
- Google Colab
- Kaggle
weights were not initialized from the checkpoint and are newly initialized:
#16
by barathm11111 - opened
I'm encountering an issue when loading the Wav2Vec2Model from the checkpoint facebook/wav2vec2-base-960h. The following weights were not initialized from the checkpoint and are newly initialized: ['wav2vec2.encoder.pos_conv_embed.conv.parametrizations.weight.original0', 'wav2vec2.encoder.pos_conv_embed.conv.parametrizations.weight.original1', 'wav2vec2.masked_spec_embed']. How can I resolve this?
I suggest running the model and see if the results are as expected. This issue has been around for a long time and is caused by migration of parameter names between different versions of the transformers library.