liodon-ai/wav2vec2-base-960h-GGUF

wav2vec2-base-960h

GGUF quantizations of facebook/wav2vec2-base-960h for use with wav2vec2.cpp.

Files

Filename	Quant	Size	Precision
model_f16.gguf	F16	189 MB	~fp16 (near-lossless)
model_q8_0.gguf	Q8_0	127 MB	8-bit quantized linear layers
model_q4_0.gguf	Q4_0	84 MB	4-bit quantized linear layers

Parity vs HuggingFace Reference

Tested on 10 audio samples from LibriSpeech.

Quant	Mean CER (vs HF)	Mean WER (vs HF)	Status
F16	0.00%	0.00%	PASS
Q8_0	0.32%	1.97%	PASS
Q4_0	2.03%	10.27%	PASS

CER/WER measured against HF model output (not ground truth) to isolate quantization error.

Usage

C++ CLI

# Clone and build
git clone --recursive https://github.com/liodon-ai/wav2vec2.cpp
cd wav2vec2.cpp && mkdir build && cd build && cmake .. && make -j

# Download and run
huggingface-cli download liodon-ai/wav2vec2-base-960h-GGUF model_q8_0.gguf --local-dir .
./wav2vec2-cli -m model_q8_0.gguf -f audio.wav

# With word timestamps
./wav2vec2-cli -m model_q8_0.gguf -f audio.wav -w

# SRT subtitle output
./wav2vec2-cli -m model_q8_0.gguf -f audio.wav --format srt > output.srt

Python (via subprocess)

import subprocess, json
result = subprocess.run(
    ['./wav2vec2-cli', '-m', 'model_q8_0.gguf', '-f', 'audio.wav', '--format', 'json'],
    capture_output=True, text=True
)
data = json.loads(result.stdout)
print(data['transcript'])

Source Model

• Original: facebook/wav2vec2-base-960h
• Architecture: base-960h
• Language: en
• Task: Automatic Speech Recognition

Conversion

Converted using wav2vec2.cpp scripts/convert_to_gguf.py.

python scripts/convert_to_gguf.py facebook/wav2vec2-base-960h model_f16.gguf --dtype f16
python scripts/convert_to_gguf.py facebook/wav2vec2-base-960h model_q8_0.gguf --dtype q8_0
python scripts/convert_to_gguf.py facebook/wav2vec2-base-960h model_q4_0.gguf --dtype q4_0

License

Follows the license of the source model.