-
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…
-
## 🚀 Feature
While the current container tests are run against bare python, there are many important libraries that require conda to be available to properly install dependencies. Therefore conda s…
-
Hi Hiro!
First, thank you for the repo. I've been following for a while and I saw you implement a big number of dl architectures.
So far I was only watching the repo from time to time, but now I…
-
**Describe the bug**
After following installation instructions for Linux there is a weird bug about torch tensor not being on CPU when converting to numpy.
```
Transcribing: 0%| …
-
I followed the [document](https://docs.mlcommons.org/inference/benchmarks/image_classification/resnet50) to inference ResNet50, using MLCommons-Python -> edge -> Tensorflow -> CUDA -> Native
The comm…
-
I have installed k2 from source in cpu .It has created new directory where there are cc files corresponding to cu files.If i make some changes in cc files .How to reflect these changes.
I have tried …
-
In the `fast_beam_search()` method, the lattice is eventually generated at: https://github.com/k2-fsa/icefall/blob/ffe816e2a8314318a4ef6d5eaba34b62b842ba3f/egs/librispeech/ASR/pruned_transducer_statel…
-
这要训练多少轮效果才好,感觉损失都好高,效果很差
-
I'd like to have a way to penalize symbol delay in RNN-T computations, and to measure it during training.
(This is relevant for systems that use time masking, to avoid the time-masking encouraging th…
-
**Is your feature request related to a problem? Please describe.**
is there a way for the canary and parakeet models (https://huggingface.co/nvidia/canary-1b) to steer transcription like whisper in…