-
Do full accuracy and performance runs for all the scenarios with below commands:
**cmr "generate-run-cmds inference _submission _all-scenarios" --model=rnnt \
--device=cuda --implementation=nvidia…
-
I am looking at the beam search decoding script for RNN-T models in espnet1 https://github.com/espnet/espnet/blob/c4aba12f9de93cb021e869a2ecce61da18b0c484/espnet/nets/beam_search_transducer.py#L247. I…
-
hello
I would like to ask you a question that may be somewhat trivial.
The shape of logits of RNN T loss is Batch, max_seq_len, max_target_len+1, class.
Why is max_target_len+1 here?
Shouldn't t…
-
Please help me with this issue.
I build targets like the installation instructions. I modified the env via $CUDA_HOME before build and used
cmake -DCUDA_TOOLKIT_ROOT_DIR=$CUDA_HOME ..
make
Wh…
-
I got the error which is shown as follows. python=3.8 torch verison=1.10.2 cudatoolkit=10.2.89 CUDA version=10.2. The GCC version is 5.4.0
![1647595370(1)](https://user-images.githubusercontent.com/…
-
Hi team,
I installed successfully with `pip install fast_rnnt`.
But can't import fast_rnnt (as attached img).
Here's my environment information.
- Win11 / Python 3.11
- CUDA 11.8 / torch 2.0…
-
Hey
I am trying to train the model but the loss is constant = '0'. I found that the loss output from the PyTorch binding of WarpRNNT is a zero tensor. I am getting an error of 'src/binding.cpp:151:…
-
Hi @HawkAaron
Thanks for this implementation. I found it very usefull.
Do you have any plans to implement this using Lingvo framework?
-
안녕하세요, 좋은 자료 공유해주셔서 감사드립니다.
```
git clone https://github.com/HawkAaron/warp-transducer
cd warp-transducer
mkdir build; cd build
cmake ..
make
```
위의 RNNT Loss를 설치하는 과정에서, cmake .. 명령어까지 한 …
-
D:\Anaconda3\envs\conformer\lib\site-packages\torch\include\ATen/core/DeprecatedTypeProperties.h(95): error C2664: 'caffe2::TypeMeta c10::TensorOptions::dtype(void) noexcept const': cannot convert ar…