Shigeki Karita

Thank you for your kindness! It really helps me. I'm trying to go back dcompute soon. > Could you please point me to an example of loading multiple module so...

Yes. I will push it in a few days (but maybe obsolete for the latest pytorch). Thanks for your request. 2020年2月10日(月) 17:40 xmuspeech : > — > You are receiving...

Hi, sorry for late. I have uploaded my working branch as-is (not organized yet). https://github.com/ShigekiKarita/espnet-semi-supervised/blob/karita-asrtts/egs/librispeech/asrtts1/note/results.org

Thanks for your request. I've never read the paper yet but just check out pytorch implentation https://github.com/timomernick/pytorch-capsule There are several missing ops in grain - binary ops: max - reductions...

Thanks for report. Did you even get `dub.selection.json`? Could you upload it here?

I just have made Travis CI build examples. It seems to success. These `dub.selections.json` might help you - ldc1.12.0 https://travis-ci.org/ShigekiKarita/grain/jobs/459560385#L745 - dmd2.083.0 https://travis-ci.org/ShigekiKarita/grain/jobs/459560386#L1990

Sorry, not yet. However, you can also check this PR https://github.com/espnet/espnet/pull/818 2019年12月23日(月) 10:47 Takeo Shibata : > Thank you for sharing this valuable info. > I read http://www.merl.com/publications/docs/TR2019-100.pdf > I...

I did codecov https://codecov.io/gh/ShigekiKarita/grain but I also found DMD on CUDA is not working. maybe `__gshared` does dangerous things? The `cublasHandle` and `cudnnHandle` should be in the safe singleton.

Thank you for report. Which pytorch version are you using? Please consider using 0.3.1. I have no plan to upgrade pytorch in this repository. Instead, I will upload new code...

Can you add tests comparing `recognize_batch` vs for-loop `recognize` results like this? https://github.com/espnet/espnet/blob/b6f7e53ade7206c33bc9beac62cdc191a0c761be/test/test_batch_beam_search.py#L89