seamless_communication
seamless_communication copied to clipboard
Foundational Models for State-of-the-Art Speech and Text Translation
PR request for Pytorch blog post. Summary: This post is the fourth part of a multi-series blog focused on how to accelerate generative AI models with pure, native PyTorch. In...

Excuse me, there is a buzzing sound in the translated voice, how to deal with it?
Thanks for releasing w2v-bert 2.0! i have two questions 1. is released checkpoint fine-tuned on seamless? or checkpoint trained on only self-supervised objective? 2. is there any data statistics to...
import torch from seamless_communication.inference import Translator translator = Translator( "seamlessM4T_v2_large", "vocoder_v2", torch.device("cuda:0"), torch.float16, ) translated_text, _, _ = translator.predict( "./input_s2tt.wav", "s2tt", tgt_lang="eng" ) print(translated_text) ----------------------------------------------------- python s2tt1.py Using the cached...
having the following error while running v2 model ` translator = Translator( File "/opt/conda/lib/python3.10/site-packages/seamless_communication/inference/translator.py", line 145, in __init__ pad_value=self.text_tokenizer.vocab_info.pad_idx or 0, pad_to_multiple=2 AttributeError: 'torch.dtype' object has no attribute 'vocab_info' `
Guess how much I love you. Oh, yes, yes, it was just like that. One autumn morning, little nut-brown hair, big nut-brown hair, and little field mouse sat bathed in...
Text to text is not available for some reason for task. The instructions in the read me don't work. m4t_predict "Hello how are you doing?" --task T2TT --tgt_lang arz --src_lang...
Dear Developers, As a title says, I would like to know whether you can share scripts for funetuning seamless streaming. As I inspected it is far from primitive to figure...