-
Hi,
I am working on an offline speaker recognition task and I need to enroll people by vector embeddings without training the network again.
I know that the AM-MobileNet1D model is not for speak…
-
Dataloader name: `nie_spoken_sg_eng/nie_spoken_sg_eng.py`
DataCatalogue: http://seacrowd.github.io/seacrowd-catalogue/card.html?nie_spoken_sg_eng
| Dataset| nie_spoken_sg_eng |
|-------------|---…
-
Hello i am having a problem with the lybrary when I execute the code this error 'Exception has occurred: ModuleNotFoundError
No module named 'neuralintents'
File "/Users/sebastianocampopena/Deskto…
-
hello @nicholas-leonard
I am very interested in **Sequence to One RNN**, woud it be possible to apply RNN on **voiceprint recognition/speaker recognition**, and what kind structure of RNN can be use…
-
Hi sir,
when I used "IBM Speech to Text" to make some speech recognition,here are some results I got:
![default](https://user-images.githubusercontent.com/15179223/35223474-710d6648-ffbc-11e7-8…
-
||link|
|----|---|
|paper| [CoMPM: Context Modeling with Speaker's Pre-trained Memory Tracking for Emotion Recognition in Conversation](https://arxiv.org/abs/2108.11626) |
|code| [paperswithcode](h…
-
I'm developing chatbot with Ionic.
We need a property in this plugin which automatically stops recognition after speech end like Android. And a solution for realtime text displays of recognition on…
-
Hi!!!
I have the following error training the model :
```
File ".\ubm.py", line 202, in
ubm.train()
File ".\ubm.py", line 50, in train
iterations=(1, 2, 2, 4, 4, 4, 4, 8, 8, 8, 8, 8, 8)
File "…
-
I am trying to reproduce the results on timit dataset.
I am having some doubt related to inference:
1. I have not seen any documentation for inferring a pre-trained model/checkpoint on sample au…
-
```python
from speechbox import ASRDiarizationPipeline
from pyannote.audio.pipelines.utils.hook import ProgressHook
pipe = ASRDiarizationPipeline.from_pretrained(asr_model="openai/whisper-base", …