-
**Core Concept/Theory**
A clear and concise description of what the concept is. Ex. It would be cool if [...]
### Idea
Utilize SeamlessM4T for building out language functionality for ATLAS/HQPl…
-
seamless_communication/assets/cards/seamlessM4T_medium.yaml Modified below:
name: seamlessM4T_medium
base: unity_nllb-200
model_arch: medium
checkpoint: "/home/app/models/multitask_unity_medium.…
-
Hey, I was looking to try out SeamlessM4T, but I'm not a python guy, and don't love installing project specific dependencies directly on my machine. I found your docker image, but I don't know how to …
-
Hi fairseq2 team!
README mentions that w2v-bert is avaliable as a pre-trained model, but I cannot find the link to download the weights anywhere - only the classes. I can see the links for S2T und…
-
I want to just t2t service (text to text translation)
do i need to load vocoder_name_or_card ??
i want to reduce resource of gpu and improve translation time
when i load no vocoder_name_or_ca…
-
Hi, is there anything I can use in the API to download/retrieve the model from a specific folder in my filesystem?
I am looking for something similar to e.g., Whisper or HuggingFace classes allow at…
g8a9 updated
9 months ago
-
translator = Translator("seamlessM4T_large", vocoder_name_or_card="vocoder_36langs", device=torch.device("cuda:0"), dtype = "float16") this line is throwing TypeError: empty() received an invalid com…
-
I am running a Debian system, with python 3.10.8.
When I try to load the model,
```
translator = Translator(
SEAMLESSM4T_MODEL_CARD_NAME,
"vocoder_36langs",
torch.device(device),…
-
Looks to be a great system. However, cannot seem to find the list of supported languages for s2st. Please advise.
-
In `src/fairseq2/data/audio.py` `AudioDecoderOutput`, `WaveformToFbankInput` and `WaveformToFbankOutput`, the `sample_rate` is defined as float and not as an integer.
I think this might be an error…