Divyesh Rajpura

Sure. The decoding method is `ctc-decoding` and below is the command which I am using export CUDA_VISIBLE_DEVICES=0; python3 zipformer/ctc_decode.py --epoch 10 --avg 1 --exp-dir exp/dnn/zipformer_ctc --use-transducer 0 --use-ctc 1 --max-duration...

Thanks for your suggestion @csukuangfj. > A larger vocab size does not necessarily imply a better performance. Will give it a try with reduced vocab size also. Thanks, Divyesh Rajpura

Changing `modified=True`, I am successfully able to run decoding on GPU with larger `--max-duration`. By reducing vocab to 500, the GPU memory usage is reduced to ~10GB. Further, I observed...

When I changed the device from GPU to CPU, its working fine. I am using cuda 12.3 and as described in [k2-1204](https://github.com/k2-fsa/k2/issues/1204), `top_sort.cu` is not working for cuda>12. Is this...