Activity for Troy
-
Troy posted a comment on discussion Help
Hi, I received an error when running nnet-train-simple: ERROR (nnet-train-simple:operator...
-
Hello, Does Kaldi DNN training support sparse feature input? I mean may I use features...
-
Thank you. I am using GPU for the nnet2 recipe, does the randomness still occur ...
-
Hi all, I am wondering if there is any randomness in the DNN training recipe such...
-
Hi Dan, Thank you very much! It fixed my problem.
-
Here's my version info: Path: . URL: https://svn.code.sf.net/p/kaldi/code/trunk Repository...
-
Hi Dan, The acoustic likelihood and real time factor of the two systems are rather...
-
I checked the features I used for nnet-get-egs for training, and nnet-latgen-faster...
-
Here're the nnet-am-info outputs: I add another scale component but it doesn't do...
-
Thank you Dan for your time. Sadly the problem persists: I double checked my features...
-
I checked the decoding results on the training set. It's only 30% WER which is much...
-
I checked the training frames, they're close as well. So it is the decoding that...
-
Here are the logs on the validation examples: This is the original DNN (with Splice...
-
Thanks Dan, Do I need to change other parameters as I changed the network structure?...
-
Hi, I am trying to explore alternative feature representation for DNN training. It...
-
Hi, I am trying to explore alternative feature representation for DNN training. It...
-
Hi, I am trying to explore another feature representation for DNN training. It looks...
1