ModelPublicadoExterno

Medocr Asr Bn

INT8 (MatMul-only dynamic quantization) ONNX export of the encoder of ai4bharat/indicconformersttbnhybridctcrnntlarge (MIT license, AI4Bharat), plus its CTC head, vocabulary, and mel filterbank. MatMul-only...

Fuentehuggingface.co/MrJAJA/medocr-asr-bn/tree/4d66c7de1a0860f79686e18a317c4b7564e9fa0e

TipoOtherVersiónAsr ctc bnCcalidad

Fuente del modelo

Tipo: Other
Tamaño del archivo: 80,3 KB–177 MB
Versión: vAsr ctc bn
Actualizado: 26 oct 2026
Fuente: Hugging Face

Descripción de la fuente

medocr-asr-bn — on-device Bangla ASR (ONNX) for the MedOCR pharmacy app

INT8 (MatMul-only dynamic quantization) ONNX export of the encoder of ai4bharat/indicconformer_stt_bn_hybrid_ctc_rnnt_large (MIT license, AI4Bharat), plus its CTC head, vocabulary, and mel filterbank. MatMul-only quantization is used because onnxruntime-android has no ConvInteger kernel.

Used by the MedOCR Android app for offline voice stocking (the pharmacist speaks price / expiry / quantity). The app downloads asr_encoder_bn.onnx once on first use; the small files ship inside the APK.

file	bytes
asr_encoder_bn.onnx	185341627
asr_ctc_bn.onnx	11559655
asr_vocab_bn.json	57466
asr_mel_filters.bin	82240

Contract: encoder audio_signal [1,80,T] float32 + length [1] int64 → [1,512,T']; ctc → logprobs [1,T',5633]; aggregate SentencePiece vocab, blank id 5632; 16 kHz mono input.

Fuentes

1 fuente

Hugging FacePrincipalVersión 4d66c7de1a0860f79686e18a317c4b7564e9fa0e

huggingface.co/MrJAJA/medocr-asr-bn/tree/4d66c7de1a0860f79686e18a317c4b7564e9fa0e

Verificado 26 oct

Download files

Artefactos del modelo

3 artefactos

asr_ctc_bn.onnx

onnx · 11,0 MB · SHA-256 385ad959364b…03bb · Hugging Face

Download file

asr_encoder_bn.onnx

onnx · 177 MB · SHA-256 ee53f6ee92e8…13d0 · Hugging Face

Download file

asr_mel_filters.bin

bin · 80,3 KB · SHA-256 2676995a34af…4cec · Hugging Face

Download file

Extractos de fuentes

2 extractos

Readme7 agohuggingface.co/MrJAJA/medocr-asr-bn

--- license: mit language: - bn tags: - automatic-speech-recognition - onnx - bangla --- # medocr-asr-bn — on-device Bangla ASR (ONNX) for the MedOCR pharmacy app INT8 (MatMul-only dynamic quantization) ONNX export of the encoder of [ai4bharat/indicconformer_stt_bn_hybrid_ctc_rnnt_large](https://huggingface.co/ai4bharat/indicconformer_stt_bn_hybrid_ctc_rnnt_large) (MIT license, AI4Bharat), plus its CTC head, vocabulary, and mel filterbank. MatMul-only quantization is used because onnxruntime-android has no ConvInteger kernel. Used by the MedOCR Android app for offline voice stocking (the pharmacist speaks price / expiry / quantity). The app downloads `asr_encoder_bn.onnx` once on first use; the small files ship inside the APK. | file | bytes | |---|---| | asr_encoder_bn.onnx | 185341627 | | asr_ctc_bn.onnx | 11559655 | | asr_vocab_bn.json | 57466 | | asr_mel_filters.bin | 82240 | Contract: encoder `audio_signal [1,80,T] float32` + `length [1] int64` → `[1,512,T']`; ctc → logprobs `[1,T',5633]`; aggregate SentencePiece vocab, blank id 5632; 16 kHz mono input.

Note7 agohuggingface.co/MrJAJA/medocr-asr-bn

Source context: 0 downloads · 0 likes · Pipeline automatic-speech-recognition · Repo MrJAJA/medocr-asr-bn