zakuro-ai / asr Goto Github PK

Hi Jean,

Thank you for updating the new code.

Please update the new package gnutools-python.
The old package on PyPI does not contain gnutools.test.

ModuleNotFoundError: No module named 'gnutools.test'

Hello. Thank you for making me a good model.

I'm using AutoSub (https://github.com/abhirooptalasila/AutoSub).

I'm trying to create a Japanese caption file using .

For some reason, I can't learn on my computer.

Could you please upload the pretrained Japanese model file and scorer file?

You pushed to the master branch on the previous issue..., but I cannot find files..

I don't know where it is.

Although you said in April that you would upload the pretrained model, I did not find it. Is there any difficulty? May I ask that is the pretrained model does not work well. I hope to get your reply

Hi, is there any pretrained model for Japanese language available? I would like to try how it works

Thanks in advance

I train the deep speech on Mozilla Persian Dataset (https://voice.mozilla.org/en/datasets). In the training procedure, always the value of the loss becomes negative then the parameters don't update. I read the document of the CTCLoss of pytorch that the output of the model should be logarithmized probabilities. Then, I add log_softmax function and the loss function becomes positive and this problem is solved, but another problem appears. The loss of the model stuck in a specific value, also the WRE and CRE don't change. What is the problem? (I can share my codes!)

Hello, I sent you an email for jp_labels.json, but I didn't get a reply. So here to consult you, can you provide the japanese preprocessed files and jp_labels.json?I want to train Japanese ASR model, these files are very helpful for my work, thank you very much.

The make script show the following error when execute.

PS D:\ws\asr> make pull
makefile:9: *** Please create a '.env' file first, for example by copying example_conf.env. No '.env' found.  Stop.

But there is no example_conf.env file in the repository to create a '.env' file from.

Hi, thanks for your repo, very clean. Can you provide some inference in this? Thanks very much!