NVIDIA-NeMo/NeMo

Speed up RNNT model inference using TRT

Open

#14.531 aperta il 20 ago 2025

Vedi su GitHub
 (1 commento) (0 reazioni) (1 assegnatario)Python (3421 fork)github user discovery
ASRcommunity-requesthelp wantedwaiting-on-customer

Metriche repository

Star
 (17.298 star)
Metriche merge PR
 (Merge medio 12g) (49 PR mergiate in 30 g)

Descrizione

Hi,

I previously trained an RNNT model and now want to accelerate it by converting it to TensorRT. I’ve exported the model to ONNX and have encoder.onnx and decoder.onnx.

I’m using the TensorRT 25.03 Docker image and trtexec to convert the models. The decoder works fine with --fp16, but when I use --fp16 for the encoder, some outputs return NaN and the results are incorrect.

Has anyone encountered this issue or knows how to fix it?

Are there any methods to accelerate RNNT model inference?

Guida contributor