AI4Bharat IndicConformer 600M multilingual packaged for sherpa-onnx offline recognition on CPU / Android.
The source model has one shared encoder and a 5,633-token CTC head covering 22 languages (256 tokens per language plus a blank). Run unmasked, it mixes scripts, so each file here fixes one language:
-inf to every token outside the chosen language's 256 (blank kept),LogSoftmax output named logprobs, with the NeMo CTC metadata sherpa-onnx expects
(vocab_size=5633, subsampling_factor=8, normalize_type=per_feature).Weights are int8 (from abid1010/indic-conformer-600m-onnx).
Each file is ~659 MB; all languages share indicconformer-tokens.txt.
| File | Language |
|---|---|
indicconformer-as.int8.onnx |
Assamese |
indicconformer-bn.int8.onnx |
Bengali |
indicconformer-brx.int8.onnx |
Bodo |
indicconformer-doi.int8.onnx |
Dogri |
indicconformer-gu.int8.onnx |
Gujarati |
indicconformer-hi.int8.onnx |
Hindi |
indicconformer-kn.int8.onnx |
Kannada |
indicconformer-kok.int8.onnx |
Konkani |
indicconformer-ks.int8.onnx |
Kashmiri |
indicconformer-mai.int8.onnx |
Maithili |
indicconformer-ml.int8.onnx |
Malayalam |
indicconformer-mni.int8.onnx |
Manipuri |
indicconformer-mr.int8.onnx |
Marathi |
indicconformer-ne.int8.onnx |
Nepali |
indicconformer-or.int8.onnx |
Odia |
indicconformer-pa.int8.onnx |
Punjabi |
indicconformer-sa.int8.onnx |
Sanskrit |
indicconformer-sat.int8.onnx |
Santali |
indicconformer-sd.int8.onnx |
Sindhi |
indicconformer-ta.int8.onnx |
Tamil |
indicconformer-te.int8.onnx |
Telugu |
indicconformer-ur.int8.onnx |
Urdu |
indicconformer-tokens.txt |
shared token table |
import sherpa_onnx
rec = sherpa_onnx.OfflineRecognizer.from_nemo_ctc(
model="indicconformer-hi.int8.onnx",
tokens="indicconformer-tokens.txt",
num_threads=4,
decoding_method="greedy_search",
)
stream = rec.create_stream()
stream.accept_waveform(16000, samples) # float32 mono, 16 kHz
rec.decode_stream(stream)
print(stream.result.text)
Input is 16 kHz mono audio. On a Galaxy S25 (4 threads) the Hindi model loads in ~1.4 s and decodes at ~0.1x real time.
Word error rate on synthetic medical consultation dialogues (Svara TTS, ~470 clips per language). Each recording is split into turns with Silero VAD and each turn is decoded once. Punctuation and tags are ignored; spelling variants (e.g. nukta) count as errors.
| Language | WER | Language | WER |
|---|---|---|---|
| Marathi | 11.5% | Malayalam | 21.8% |
| Telugu | 11.5% | Nepali | 24.8% |
| Kannada | 15.4% | Punjabi | 26.4% |
| Hindi | 16.0% | Gujarati | 27.2% |
| Bengali | 16.6% | Maithili | 29.6% |
| Tamil | 18.1% | Dogri | 42.9% |
| Sanskrit | 19.7% | Bodo | 43.6% |
| Assamese | 20.1% |
Odia, Urdu, Konkani, Kashmiri, Manipuri, Santali and Sindhi were not benchmarked.
MIT, same as the source models.
indicconformer-brx.int8.onnx 659 MBindicconformer-doi.int8.onnx 659 MBindicconformer-kok.int8.onnx 659 MBindicconformer-mai.int8.onnx 659 MBindicconformer-mni.int8.onnx 659 MBindicconformer-sat.int8.onnx 659 MBindicconformer-as.int8.onnx 659 MBindicconformer-bn.int8.onnx 659 MBindicconformer-gu.int8.onnx 659 MBindicconformer-hi.int8.onnx 659 MBindicconformer-kn.int8.onnx 659 MBindicconformer-ks.int8.onnx 659 MBindicconformer-ml.int8.onnx 659 MBindicconformer-mr.int8.onnx 659 MBindicconformer-ne.int8.onnx 659 MBindicconformer-or.int8.onnx 659 MBindicconformer-pa.int8.onnx 659 MBindicconformer-sa.int8.onnx 659 MBindicconformer-sd.int8.onnx 659 MBindicconformer-ta.int8.onnx 659 MBindicconformer-te.int8.onnx 659 MBindicconformer-ur.int8.onnx 659 MBHave the files? Submit a torrent.
Submit a torrentHave the complete files and permission to share them? Replace the example folder path. The script checks the files, makes a torrent and submits its magnet and checksums. Keep your torrent client seeding. Get your community key from your settings, and read the script before running it.
curl -fsSL https://pirateface.co/package.sh | bash -s -- --repo Docglasses-AI/indic-conformer-600m-sherpa-onnx-int8 ./model-folderai4bharat/indic-conformer-600m-multilingual