Downloads · 30 days
2.8K
27% of all-time downloads
cstr/wav2vec2-large-xlsr-53-english-GGUF
wav2vec2-large-xlsr-53-english-GGUF is a automatic speech recognition model from cstr. Use it when you need speech turned into text. It is set up for ggml. The card lists the license as apache-2.0.
GGUF conversion of jonatasgrosman/wav2vec2-large-xlsr-53-english for English speech recognition.
Downloads · 30 days
2.8K
27% of all-time downloads
All-time downloads
10.7K
Public
Repo size
1.5 GB
Likes
0
Public
Click a slice to open those files.
.gguf1.5 GB · 100%
From the Hugging Face model README
GGUF conversion of jonatasgrosman/wav2vec2-large-xlsr-53-english for English speech recognition.
| Property | Value |
|---|---|
| Architecture | Wav2Vec2 (CNN feature extractor + 24 Transformer layers) |
| Hidden size | 1024 |
| Attention heads | 16 |
| CTC vocabulary | 33 tokens |
| Format | GGUF (F16 weights) |
| Size | 627 MB |
The model was pre-trained on 53 languages (XLSR-53) and fine-tuned on English Common Voice data with a CTC head. It accepts 16 kHz mono audio and outputs character-level transcriptions.
crispasr \
--backend wav2vec2 \
-m wav2vec2-xlsr-en.gguf \
audio.wav
Weights were converted from the original Hugging Face PyTorch checkpoint into GGUF format with F16 precision for all transformer and feature-extractor parameters.
Apache-2.0 — same as the original model.