-
Hello!
I'm trying to launch QuartzNet on dataset, where jasper does just fine, we've also checked wavs / features correctness. After several epochs training in mixed precision diverges, cause nans…
-
Hi all,
i am workin on a quartznet model with 5-gram beam search LM for decode the CTC matrix.
I was wondering if it could be possible to implement a BERT-based LM on top of my asr model.
If i si…
-
**Description**
I used ensemble model, which consists of three models: 2 trt + 1pt + 1 custom backend.
However the output is quite strange:
```
perf_analyzer -m quartznet-ensemble -b 1 --concurren…
Slyne updated
3 years ago
-
**Describe your question**: As Nemo is a Pytorch lightening module (as wav2letter) is there an easy to implement Wav2letter with Nemo ?
A clear and concise description of your question.
Describe w…
-
hi, i think there's a mistake here,
https://github.com/stefanpantic/asr/blob/86e13ee4d76779318caf66a81d6342c096e5fb35/models/quartznet.py#L30,
should be: filters: [512, 1024], kernel_size: [87,…
-
**Describe the bug**
The ` quartznet/src/Quartz/Xml/job_scheduling_data_2_0.xsd ` file contains this code for job definition:
```xsd
Define a JobDetail
…
-
Hello,
I have a question about one parameter used in the config file of quartznet model. Can someone tell me what is the purpose of stft_conv parameter? in the config file it's set to False.
![imag…
-
**Describe the bug**
Due to `torch` 1.8.0 just was released a few days ago, `pytorch-lightning` is still incompatible with the new torch.
Since NeMo setup pulls the latest stable torch release, im…
-
안녕하세요?
좋은 모듈을 공개해 주셔서 감사합니다.
몇가지 궁금한게 있어서 이슈를 남기게 되었습니다.
AIHUB 음성인식 성능을 향상시키기 위하여 위키피디아 데이터를 활용하여 KenLM(4-gram) 모델을 구축하고 Decoding 과정에 활용하였다고 Read.me 파일에 기술해주셨는데
1. 위키피디아 데이터를 공유 받을 수 있을까요?
…
-
I'm deploying a speech to text model on Triton Inference Server using Quartznet and n-gram language model for the Vietnamese language. I used Python Backend for this model.
I have a problem with Unic…