Seren Autonome — on-device models
ONNX model bundle downloaded on first launch by the Seren Autonome Android app (embedded hub, on-device inference). Layout mirrors the app's external models dir.
| Path | Size | Purpose | Upstream |
|---|---|---|---|
parakeet/encoder-model.int8.onnx |
652 MB | Parakeet TDT 0.6B v3 encoder (STT) | NVIDIA NeMo parakeet-tdt-0.6b-v3 (CC-BY-4.0) |
parakeet/decoder_joint-model.int8.onnx |
18 MB | Parakeet TDT decoder/joint | idem |
parakeet/nemo128.onnx |
— | mel preprocessor | idem |
parakeet/vocab.txt |
— | tokenizer vocab | idem |
embeddings/model.onnx |
539 MB | sentence embeddings (memory recall) | BGE-family |
embeddings/vocab.txt |
— | tokenizer vocab | idem |
speaker/3dspeaker_eres2net_base_16k.onnx |
40 MB | speaker embedding | 3D-Speaker ERes2Net (Apache-2.0) |
speaker/segmentation_pyannote_3.onnx |
6 MB | speaker segmentation (diarization) | pyannote segmentation-3.0 (MIT) |
Weights are redistributed from their respective upstream projects under their original licenses; see each upstream for attribution and terms.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support