--- license: cc-by-4.0 base_model: - nvidia/parakeet-tdt-0.6b-v3 pipeline_tag: automatic-speech-recognition library_name: onnx tags: - onnx - rostrum - automatic-speech-recognition - sherpa-onnx --- # Parakeet TDT 0.6B v3 (int8 ONNX, sherpa-onnx), mirrored for Rostrum NVIDIA's Parakeet TDT 0.6B v3 speech recognition model (English and 24 other European languages, with punctuation, capitals and word times), as the int8 ONNX export made by the k2-fsa sherpa-onnx project. This repository is a pinned, public copy of the exact files the [Rostrum](https://huggingface.co/rostrumstudios) apps download. Rostrum fetches each file at a fixed commit of this repository and checks its SHA-256 before use. ## What Rostrum uses it for Sound's transcripts and Compose's captions: it turns speech into text with each word's start and end time. Rostrum uses it for **analysis only, never generation**: it reads what is in a recording or a picture and makes no new sound, voice or image. ## Files | File | Bytes | SHA-256 | |---|---:|---| | `encoder.int8.onnx` | 652,184,281 | `acfc2b4456377e15d04f0243af540b7fe7c992f8d898d751cf134c3a55fd2247` | | `decoder.int8.onnx` | 11,845,275 | `179e50c43d1a9de79c8a24149a2f9bac6eb5981823f2a2ed88d655b24248db4e` | | `joiner.int8.onnx` | 6,355,277 | `3164c13fc2821009440d20fcb5fdc78bff28b4db2f8d0f0b329101719c0948b3` | | `tokens.txt` | 93,939 | `d58544679ea4bc6ac563d1f545eb7d474bd6cfa467f0a6e2c1dc1c7d37e3c35d` | In Rostrum's models folder they go under `parakeet-tdt-0.6b-v3/`. ## Source | What | Where | Commit / version | |---|---|---| | Exact files (unchanged copy) | https://huggingface.co/csukuangfj/sherpa-onnx-nemo-parakeet-tdt-0.6b-v3-int8 | 2bda32ec70b097a55adaa07d9a7173915b43cc78 | | Original model, NVIDIA | https://huggingface.co/nvidia/parakeet-tdt-0.6b-v3 | 541d1f99c6b0c3cd0b11a95167540bb8edefd82b (main when mirrored) | | Export code, k2-fsa sherpa-onnx | https://github.com/k2-fsa/sherpa-onnx | 5d2bd37de78b223a2e8e00e25483d12c5b164299 (master when mirrored) | ## Authors NVIDIA (model and weights); Fangjun Kuang and the k2-fsa / Next-gen Kaldi team (int8 ONNX export). ## Licence The weights are licensed **CC-BY-4.0** by NVIDIA (full text in `LICENSE`). The sherpa-onnx export code is Apache-2.0 (`LICENSE-sherpa-onnx`). **Attribution (CC-BY-4.0):** "Parakeet TDT 0.6B v3" by NVIDIA, https://huggingface.co/nvidia/parakeet-tdt-0.6b-v3, licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Converted to int8 ONNX by the k2-fsa sherpa-onnx project. This repository redistributes those files unchanged; no other changes were made.