Nos_ASR-wav2vec2-xls-r-300m-gl-onnx

Nos_ASR-wav2vec2-xls-r-300m-gl (ONNX)

ONNX export of proxectonos/Nos_ASR-wav2vec2-xls-r-300m-gl, a Galician wav2vec2 XLS-R 300M CTC ASR model by Proxecto Nós, for use with onnx-asr (wav2vec2-ctc model type).

Per-utterance zero-mean/unit-variance normalization (do_normalize=True in the original feature extractor) is baked into the ONNX graph, masked by input_lengths for correct behavior with padded/batched input, so the model works with onnx-asr's plain identity preprocessor (raw 16kHz waveform in).

Usage

import onnx_asr

model = onnx_asr.load_model("wav2vec2-ctc", "path/to/downloaded/model")
# or, once registered upstream with a friendly name:
# model = onnx_asr.load_model("OpenVoiceOS/Nos_ASR-wav2vec2-xls-r-300m-gl-onnx")
print(model.recognize("test.wav"))

Files

• model.onnx / model.onnx.data — fp32 ONNX graph (inputs: input_values (batch, samples) float32, input_lengths (batch,) int64; output: logprobs (batch, frames, vocab) float32 log-softmax).
• vocab.txt — CTC vocabulary in onnx-asr's token id format (word-delimiter → ▁, pad token → <blk>).
• config.json — {"model_type": "wav2vec2-ctc", "subsampling_factor": 320}.

No int8 quantized variant is included yet — onnxruntime.quantization does not currently support the torch.onnx dynamo-exported graph for this architecture.

License

Apache 2.0, inherited from the base model.