seamless_communication

seamless_communication copied to clipboard

The following error creeps up, while: m4t_predict /workspace/english/data/lib_light/100_sea_fairies_0812_librivox_64kb_mp3_01_baum_sea_fairies_64kb_0.wav --task asr --tgt_lang "eng" --model_name seamlessM4T_v2_large warnings.warn("torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.") 2024-01-05 16:53:03,472 INFO -- seamless_communication.cli.m4t.predict.predict: text_generation_opts=SequenceGeneratorOptions(beam_size=5, soft_max_seq_len=(1, 200), hard_max_seq_len=1024,...

YKoustubhRao

i have fine-tuned seamless model. i tried to pickle the model and i found out it's not possible so i was forced to save the stat_dict and every time i...

ARman-AT

T2TT not works

1

➜ seamless_communication git:(main) ✗ m4t_predict hello --task T2TT --tgt_lang eng --src_lang cmn usage: m4t_predict [-h] [--task {ASR,S2ST,S2TT}] [--tgt_lang TGT_LANG] [--src_lang SRC_LANG] [--output_path OUTPUT_PATH] [--model_name MODEL_NAME] [--vocoder_name VOCODER_NAME] [--text_generation_beam_size TEXT_GENERATION_BEAM_SIZE] [--text_generation_max_len_a...

zsq2010

Update README.md

1

adding name option to m4t_prepare_dataset to m4t-evaluation example.

alihejazi97

packaging conflict

1

when i runthe code ,"python setup.py install",i get the error: packaging 24.1 is installed but packaging~=23.1 is required by {'fairseq2'},i try to update fairseq2 version to 0.2.1, but the error...

Albert337

seamless_streaming_unity相关权重自行下载，怎么设置路径

4

在seamless_streaming_unity.yaml配置文件中，修改了char_tokenizer: 和checkpoint:参数，改成了我下载好的权重路径，为什么推理运行还要下载权重呢？

sunyclj

Is it possible to fine-tune the model using a fine-tuning method such as LORA?

2

Is it possible to fine-tune the model using a fine-tuning method such as LORA?

FHT-hub

Fix broken m4t_prepare_dataset guides

2

- Fix broken m4t_prepare_dataset guides cause from https://github.com/facebookresearch/seamless_communication/pull/443

kujyp

hey there i have checked out your fine-tuning code. my query is i have custom w2vBERT/whisper models fine-tuned for wider spectrum of Indian languages. as well as there exists MT...

StephennFernandes

about tsv file

1

command: streaming_evaluate --task s2st --data-file ../audio-tsv/test.tsv --audio-root-dir ../test_data/audio/ --output ../ --tgt-lang eng --no-scoring 2024-07-11 05:44:48,463 INFO -- seamless_communication.streaming.agents.unity_pipeline: Loading the UnitY model: seamless_streaming_unity on device=cuda:0, dtype=torch.float16 2024-07-11 05:44:56,398 INFO --...

sunyclj