deep-clustering
deep-clustering copied to clipboard
zhr1201
Folder 'dcdata'
Hi there, In the file train_net.py, you are trying to access 1.pkl, 2.pkl,...12.pkl from a folder dcdata But we haven't generated any such files. Ofc we have generated val.pkl using datagenerator.py
Please help me understand what these are for and how to generate them.
Sorry for the misunderstanding, I meant train.py should be used to generate 1.pkl, 2.pkl...
Hello, I have the same question. There is no train.py provided in you code. What do you mean by that?
@zhr1201 Thank you for replying. Please guide me through this issue.
Did you mean datagenerator.py ? And what exactly are 1.pkl, 2.pkl, ...., how to generate them?
I guess this folder contains features of owner's train audio. Maybe you should extract your train audio's features and write down these features as *.pkl file, then put their path into datagenerator class.
@vzxxbacq there will be one pkl file for one audio or there will be a single pkl file containing the features of the whole data set?
Hello @janhavi028 , I think we can write them in any number of files with any number of features. In train_net.py author pass 12 files to the model and check function re_init() in datagenerator.py we know author extract features and put them into one pkl file.
@vzxxbacq Hi! Do you mean I should use datagenerator.py to get 12 pkl files for 12 single persons and one pkl file for the whole 12 persons?
@pramodhrachuri Hey! How did you generate the 12 pkl files? Do they belong to 12 different speakers? I am confused with that,thank you!
@tangchenjun Not necessarily, I generated each file using the same scheme and raw data with random sampling for 12 times.
请问意思是，用同样的数据，利用datagenerator.py产生12个pkl文件？我想问如果采用timit语音库，每个说话人大概只有5语音？怎么使用啊
