facebookresearch/seamless_communication issues and pull requests

#421 - Error with GPU 40GB !!!!!!!!!

Issue - State: closed - Opened by developeranalyser 6 months ago - 6 comments

#421 - Error with GPU 40GB !!!!!!!!!

Issue - State: closed - Opened by developeranalyser 6 months ago - 6 comments

#418 - Fixed finetuning trainer and enable freezing layers

Pull Request - State: closed - Opened by zrthxn 6 months ago
Labels: CLA Signed

#418 - Fixed finetuning trainer and enable freezing layers

Pull Request - State: closed - Opened by zrthxn 6 months ago
Labels: CLA Signed

#409 - Why always Downloading the tokenizer of seamlessM4T_v2_large

Issue - State: open - Opened by Longleaves 6 months ago - 7 comments

#406 - Segment audio with Silero VAD and pipeline with Transcriber

Pull Request - State: closed - Opened by am831 6 months ago
Labels: CLA Signed

#406 - Segment audio with Silero VAD and pipeline with Transcriber

Pull Request - State: closed - Opened by am831 6 months ago
Labels: CLA Signed

#403 - unknown result in t2tt

Issue - State: open - Opened by asulada 6 months ago - 3 comments

#396 - finetune results

Issue - State: open - Opened by FHT-hub 6 months ago - 5 comments

#387 - Is there a fine-tuning script for Seamless Streaming?

Issue - State: open - Opened by shushanxingzhe 7 months ago - 3 comments

#383 - How to train the vocoder?

Issue - State: open - Opened by kzmaker 7 months ago - 1 comment

#372 - Difference between "cmn" and "cmn_Hant"?

Issue - State: open - Opened by GrayGriffin 7 months ago - 1 comment

#372 - Difference between "cmn" and "cmn_Hant"?

Issue - State: open - Opened by GrayGriffin 7 months ago - 1 comment

#370 - How to reproduce T2TT result using HF?

Issue - State: open - Opened by aliencaocao 7 months ago - 1 comment

#369 - Seamless-M4T-v2 Catastrophic transcription error on clear audio (german), but file works fine in whisper v2

Issue - State: open - Opened by asusdisciple 7 months ago

#368 - What is the format of the input data .tsv in speech recognition? Is there any example file?

Issue - State: open - Opened by wangsang123 7 months ago

#367 - Initializing the model on an M3 Mac fails on a fresh conda environment

Issue - State: closed - Opened by d0minYk 7 months ago

#366 - LM Rescoring for Seamless text decoder

Issue - State: open - Opened by Sameep-c 7 months ago - 1 comment

#365 - Facebook information

Issue - State: open - Opened by mt64423 7 months ago

#364 - fairseq2.assets.metadata_provider.AssetNotFoundError: An asset with the name '/Models/seamlessM4T_v2_large.pt' cannot be found.

Issue - State: open - Opened by dawit3228 7 months ago

#363 - Some languages do not support speech synthesis

Issue - State: open - Opened by kzmaker 8 months ago - 1 comment

#362 - Wrong result for traditional Chinese

Issue - State: closed - Opened by fumin 8 months ago - 5 comments

#360 - Fix the url of the Mutox dataset

Pull Request - State: closed - Opened by avidale 8 months ago
Labels: CLA Signed

#359 - MuTox dataset not accessible

Issue - State: open - Opened by BhashaBluff 8 months ago - 3 comments

#358 - Confidence scores for the predictions generated?

Issue - State: open - Opened by Awaisn25 8 months ago

#355 - Update README.md

Pull Request - State: closed - Opened by yaya-sy 8 months ago - 1 comment
Labels: CLA Signed

#354 - seamlessM4T_v2_large finetuning on speech translation task

Issue - State: open - Opened by laleye 8 months ago

#353 - Analysis of Audio Frame Alignment Discrepancies in Metadata Retrieval Process

Issue - State: open - Opened by nassergharbi 8 months ago

#352 - fix cmn ASR eval in SeamlessExpressive

Pull Request - State: open - Opened by hygong-fb 8 months ago
Labels: CLA Signed

#351 - finetune.run failed on assert batch.text_to_units.prev_output_tokens is not None

Issue - State: open - Opened by laleye 8 months ago - 4 comments

#350 - [Finetune] Error(s) in loading state_dict for UnitYModel

Issue - State: open - Opened by muratozbek23 8 months ago

#349 - Is it possible to run SeamlessStreaming on an Apple M1 Pro?

Issue - State: open - Opened by snmishra311 8 months ago

#347 - OOM with 20GB GPU on SeamlessStreaming evaluate

Issue - State: open - Opened by Mega4alik 8 months ago

#346 - make T2TT in blingual model auto-concatenated

Pull Request - State: open - Opened by antoine-tran 8 months ago
Labels: CLA Signed

#345 - Incorrect layer index between offline discrete unit extraction and UnitY2 forced alignment

Issue - State: open - Opened by julianzaidi 8 months ago

#344 - Add back unity lib files

Pull Request - State: closed - Opened by cndn 8 months ago
Labels: CLA Signed

#343 - Why only one previous word is used as input when predict the current word?

Issue - State: closed - Opened by KeepSilentNow 8 months ago - 2 comments

#342 - scripts to reproduce w2v-BERT 2.0 pretraining ?

Issue - State: open - Opened by StephennFernandes 8 months ago - 2 comments

#341 - finetuning required GPU

Issue - State: open - Opened by zouberou-sayibou 8 months ago - 2 comments

#340 - Support Pip Wheel / PyPi Package

Issue - State: open - Opened by moinnadeem 8 months ago

#339 - FineTune T2S

Issue - State: closed - Opened by developeranalyser 8 months ago - 1 comment

#338 - Trouble Extracting Monolingual Datasets from SeamlessAlign

Issue - State: open - Opened by nassergharbi 8 months ago

#337 - Question about Maltese Dataset Consistency - Extension of the previous S2S release (November 30, 2023)

Issue - State: open - Opened by nassergharbi 8 months ago

#336 - sos!!! fairseq2.assets.download_manager.AssetDownloadError: The download of the tokenizer of mintox has failed. See nested exception for details.

Issue - State: open - Opened by caiji2019-cai 8 months ago - 6 comments

#335 - fix typo

Pull Request - State: open - Opened by gri11 9 months ago - 1 comment
Labels: CLA Signed

#334 - Interactive only unity.cpp & Fix warnings

Pull Request - State: closed - Opened by cndn 9 months ago - 2 comments
Labels: CLA Signed

#333 - Updated README with nllb converted model path

Pull Request - State: closed - Opened by cndn 9 months ago
Labels: CLA Signed

#332 - Mutox Classifier Model

Pull Request - State: closed - Opened by Mortimerp9 9 months ago
Labels: CLA Signed

#331 - local model path can not be recognized and always downloading model again

Issue - State: closed - Opened by weiyshay 9 months ago - 2 comments

#329 - unity.cpp sync

Pull Request - State: closed - Opened by cndn 9 months ago
Labels: CLA Signed

#328 - torch.compile + CUDA Graph optimization for bs=1

Pull Request - State: open - Opened by YJYJLee 9 months ago - 1 comment
Labels: CLA Signed

#327 - ASR fine-tuning

Issue - State: open - Opened by h9-tect 9 months ago - 4 comments

#326 - Adjust M4T example finetuning script to align with November changes

Pull Request - State: closed - Opened by mavlyutovr 9 months ago - 2 comments
Labels: CLA Signed

#325 - M4T-v2 Language Identification for Audio / Spoken Language

Issue - State: open - Opened by asusdisciple 9 months ago - 5 comments

#324 - t2st

Issue - State: open - Opened by xialixx 9 months ago

#323 - Missing key(s) and size mismatch error for SeamlessM4T_medium

Issue - State: open - Opened by YJYJLee 9 months ago - 9 comments

#322 - question about w2v-bert 2.0

Issue - State: open - Opened by seastar105 9 months ago

#321 - Finetuning for ASR and the dataset preparation

Issue - State: open - Opened by Gxhappiness 9 months ago - 3 comments

#320 - GGML Allow loading fp16 models, add --mem flag

Pull Request - State: closed - Opened by gwenzek 9 months ago
Labels: CLA Signed

#319 - AttributeError: 'torch.dtype' object has no attribute 'vocab_info'

Issue - State: open - Opened by bk111 9 months ago - 3 comments

#318 - after T2TT, the English period became the double question mark in Chinese!

Issue - State: open - Opened by bk111 9 months ago - 1 comment

#317 - seamless v2 finetunning assert model.t2u_model.target_vocab_info == unit_tokenizer.vocab_info

Issue - State: open - Opened by adnankarimjs 9 months ago - 1 comment

#316 - ASR inference

Issue - State: open - Opened by YKoustubhRao 9 months ago

#315 - How to finetune large v2 model?

Issue - State: open - Opened by adnankarimjs 9 months ago - 4 comments

#314 - model inference

Issue - State: open - Opened by haiderasad 9 months ago - 1 comment

#313 - What can I do if I want to use gpu to inference with a cuda less than 11.4?

Issue - State: open - Opened by huangcanan 9 months ago

#312 - Efficient Monotonic Multihead Attention - Pchoose layers not used to calculated weight attention of key and query in encoder-decoder attention.

Issue - State: open - Opened by v-tuenv 9 months ago

#311 - Where can I find the source code for Watermarking?

Issue - State: open - Opened by jishengpeng 9 months ago

#310 - Real-Time Implementation of Seamless Expressive

Issue - State: open - Opened by HardikJain02 9 months ago - 1 comment

#309 - Where can I find the source code for SeamlessWM?

Issue - State: closed - Opened by systemdevart 9 months ago - 7 comments

#306 - multi-gpu

Issue - State: open - Opened by tkone2018 9 months ago

#305 - Do ASR must specify the parameter “tgt_lang” ? (ASR 必须要指定tgt_lang这个参数吗)

Issue - State: open - Opened by lilongwei5054 9 months ago

#304 - Update README.md

Pull Request - State: closed - Opened by nguyenvulong 9 months ago - 2 comments
Labels: CLA Signed

#303 - Poor output for Japanese S2TT tasks, and `mps` device

Issue - State: open - Opened by bcherny 9 months ago - 1 comment

#297 - Are SeamlessM4T v2, SeamlessExpressive, and SeamlessStreaming under an MIT license and therefore free for commercial use?

Issue - State: closed - Opened by troublesprouter 10 months ago - 1 comment

#296 - Is it possible to add another language?

Issue - State: open - Opened by abrakadabrar 10 months ago - 1 comment

#295 - Fix HF model sampling_rate in M4T README.md

Pull Request - State: closed - Opened by elbayadm 10 months ago
Labels: CLA Signed

#294 - Decrease mem usage and add `--mem` to allow more memory

Pull Request - State: closed - Opened by gwenzek 10 months ago
Labels: CLA Signed

#292 - Refactoring expressivity/predict into ExpressiveTranslator.

Pull Request - State: open - Opened by kauterry 10 months ago - 11 comments
Labels: CLA Signed

#291 - We tried Seamless for VideoDubber.ai

Issue - State: open - Opened by Souvic 10 months ago

#289 - Chckpoint error during finetuning

Issue - State: open - Opened by zouberou-sayibou 10 months ago - 3 comments

#288 - Add HF Hub Model card + checkpoint. - Conformer shaw

Pull Request - State: closed - Opened by Vaibhavs10 10 months ago
Labels: CLA Signed

#287 - Silent Parts of the Audio

Issue - State: open - Opened by m-pektas 10 months ago

#286 - Open-sourcing W2v-BERT 2.0 speech encoder.

Pull Request - State: closed - Opened by kauterry 10 months ago
Labels: CLA Signed

#285 - got "killed" message when I run m4t_predict a mp3 file.

Issue - State: open - Opened by martinzhangdidi 10 months ago - 3 comments

#281 - Seamless Expressive cli doesnt refer to the same checkpoint name as in expressives repo

Issue - State: open - Opened by helloworld53 10 months ago - 1 comment

#279 - Updated Readme.md

Pull Request - State: closed - Opened by Oheed911 10 months ago
Labels: CLA Signed

#277 - save model as fp16

Pull Request - State: closed - Opened by gwenzek 10 months ago
Labels: CLA Signed

#276 - Remove T2ST task from audio resampling and add ASR and S2TT

Pull Request - State: closed - Opened by yeyinthtoon 10 months ago
Labels: CLA Signed

#270 - Translation

Issue - State: closed - Opened by drhate7000 10 months ago

#269 - SeamlessM4T v2: Incomplete Text Synthesis and Audio Quality Issues (example with French Text)

Issue - State: open - Opened by nassergharbi 10 months ago - 3 comments

#266 - Microphone is not working well for Large V2 model demo!

Issue - State: open - Opened by arash-aut 10 months ago - 2 comments

#262 - Update README.md

Pull Request - State: closed - Opened by OvidijusParsiunas 10 months ago - 1 comment
Labels: CLA Signed

#257 - How to utilize "expressive function" in "LANGUAGE_CODE_TO_NAME" to translate from English to other languages?

Issue - State: closed - Opened by jiroh1 10 months ago - 1 comment

#256 - allocr for encoder

Pull Request - State: closed - Opened by gwenzek 10 months ago - 1 comment
Labels: CLA Signed

GitHub / facebookresearch/seamless_communication issues and pull requests