-
Thank you for the great RNN-T library :)
I got a "segmentation fault" message while running test_gpu.
```
(py3-tf-gpu) sephiroce@Sephiroce-BiKE:~/warp-transducer/build$ ./test_gpu
Running gpu te…
-
I used post-training quantization on a neural network architecture with LSTMs and I have a quantizer.model that has the scale-factors and activations for each layer. However, I want the actual quantiz…
-
Hello, I used RNNT training on the Chinese speech recognition library of more than 300 hours (the encoder did pretrain, but the decoder is a random initialization parameter). After training dozens of …
-
Hi. First thank you for your code.
I have a problem GPU version warp-transducer.
I applied gpu version warp-transducer to your E2E-ASR code. (with minor adjust)
At this case, loss became "nan" af…
-
I adopt this model with timit datasets, but the rnnt model does not coverage, the loss is about 500+, and I want to know the reason which cause it?
-
Hi everyone!
Many thanks for an amazing repo!
So far I could achieve stable / parallelized training on a modified version of the code in this repo:
- stft instead of MFCC;
- librosa / numpy / …
-
-
getting test.py output as :
[ 9.53674316e-07]
CPU Tests passed!
GPU execution requested, but not compiled with GPU support
Traceback (most recent call last):
File "/home/tanish/multi_gpu_ds3…
-
My code was ok when using MxNet v1.1.0. When I upgrade it to v1.4.0, I got the current error when using lstm:
```
Traceback (most recent call last):
File "train3.py", line 271, in
train(…
-
@HawkAaron I tried to train the transducer on my own dataset. When I increase the hidden_size from 250 to 512 and increase the num_layers from 3 to 5, the performance became worse, why? The bigger mod…