search

qute012 / Real-Time-SR

Real time speech recognition model based on pytorch on android application
3 stars 2 forks source link

about the code of speech recognition based on pytorch #1

Open lvyilan23 opened 3 years ago

lvyilan23 commented 3 years ago

Hi, can you give me the code of speech recognition based on pytorch?

qute012 commented 3 years ago

Hi, you means python code for training?

2020년 10월 15일 (목) 오후 2:48, lvyilan23 notifications@github.com님이 작성:

Hi, can you give me the code of speech recognition based on pytorch?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/qute012/Real-Time-SR/issues/1, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIDIU3FPOB26HN5RC63N2MDSK2EJLANCNFSM4SRQIC3A .

lvyilan23 commented 3 years ago

yes. Because I want to know how to set the input and output in android according to the python model.

qute012 commented 3 years ago

Thank you for attention. that app returns bad recognition, but model works well not in application for test set. I refered https://github.com/tugstugi/pytorch-speech-commands. My code is similar to that, if you wanna get mine, tell me :D.

2020년 10월 15일 (목) 오후 9:26, lvyilan23 notifications@github.com님이 작성:

yes. Because I want to know how to set the input and output in android according to the python model.

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/qute012/Real-Time-SR/issues/1#issuecomment-709280735, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIDIU3FJHEHH5GMVMUVV7OLSK3TBDANCNFSM4SRQIC3A .

lvyilan23 commented 3 years ago

Hi, thank u for responding. I still wants to get your code. In fact, I just want to know how the input and output shape correspond in pytorch and android. through your code. The performance is not important. Thank u very much!!

lvyilan23 commented 3 years ago

Hi, can u give me a copy of your paper? I used my own trained model in Android, and met some mistakes. I want to know how you train it.

qute012 commented 3 years ago

Hello, i’m sorry for that i can’t find my code. It’s too old T.T. I can help you. You may meet error when resnet implements skip connection, right? It solves not to use any functions for residual.

2020년 10월 19일 (월) 오후 9:29, lvyilan23 notifications@github.com님이 작성:

Hi, can u give me a copy of your paper? I used my own trained model in Android, and met some mistakes. I want to know how you train it.

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/qute012/Real-Time-SR/issues/1#issuecomment-712123023, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIDIU3CLR5QADHNHZQOYPRLSLQWKNANCNFSM4SRQIC3A .

qute012 commented 3 years ago

And when i used tf-lite l, it’s good works~ so you need to try to make tf-lite speech command models.

2020년 10월 20일 (화) 오전 11:54, 서쩰루리 ejrwls012@gmail.com님이 작성:

Hello, i’m sorry for that i can’t find my code. It’s too old T.T. I can help you. You may meet error when resnet implements skip connection, right? It solves not to use any functions for residual.

2020년 10월 19일 (월) 오후 9:29, lvyilan23 notifications@github.com님이 작성:

Hi, can u give me a copy of your paper? I used my own trained model in Android, and met some mistakes. I want to know how you train it.

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/qute012/Real-Time-SR/issues/1#issuecomment-712123023, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIDIU3CLR5QADHNHZQOYPRLSLQWKNANCNFSM4SRQIC3A .