#onnx #machine-learning #r-ten

rten-generate

Utilities to simplify running auto-regressive models with RTen

11 releases (6 breaking)

new 0.17.0 Apr 9, 2025
0.16.0 Feb 8, 2025
0.15.1 Jan 6, 2025
0.15.0 Dec 28, 2024
0.12.0 Jul 30, 2024

#814 in Machine learning

Download history 109/week @ 2024-12-27 115/week @ 2025-01-03 16/week @ 2025-01-10 1/week @ 2025-01-31 124/week @ 2025-02-07 9/week @ 2025-02-14 3/week @ 2025-02-21 4/week @ 2025-02-28 83/week @ 2025-04-04

83 downloads per month

MIT/Apache

2.5MB
54K SLoC

Utilities to simplify running auto-regressive RTen models such as transformer decoders.

For working examples, see the examples in the rten-examples crate which import rten_generate.


rten-generate is a layer on top of RTen which handles the generation loop for auto-regressive transformer models (aka. "transformer decoders" or "generative AI"). This includes managing the KV cache, sampling and post-processing logits etc.

Dependencies

~1.7–3MB
~63K SLoC