fairseq

fairseq copied to clipboard

How to convert mms model to hf model?

1

## ❓ How to convert mms model to hf model? #### What is your question? #### I tried using the conversion script in Huggingface Transformers ```python convert_wav2vec2_original_pytorch_checkpoint_to_pytorch.py --pytorch_dump_folder_path=/hy-tmp/asr-hf --checkpoint_path=/hy-tmp/asr/model.pt --dict_path=/hy-tmp/asr/uig-script_arabic.txt...

glogskun12303

For MMS TTS, is it possible to add pauses, emotion, inflection, ect?

1

## ❓ Questions and Help #### What is your question? I am playing with and learning about the MMS TTS. I have it running and am curious if it is...

JWesorick

How can i run your tts locally on cpu, 16gb ram, Windows 11?

3

JohnClaw

lbfcxz

Romanian vocab has a different ț

1

## 🐛 Bug With the release of [MMS_TTS_Inference_Colab.ipynb] I wanted to test the TTS feature. ### To Reproduce Steps to reproduce the behavior (**always include the command you ran**): LANG:...

ainub

## ❓ Questions and Help ### Before asking: 1. search the issues. 2. search the docs. #### What is your question? I followed **examples/roberta/README.pretraining.md** and wanted to continue pretraining in...

Once2gain

mBART model code for translation encounters errors.

1

## 🐛 Bug The code for the mBART50 model seems to contain some bugs. The file 'hub_utils.py' contains a 'translate' function. It encounters an error: TypeError: TranslationMultiSimpleEpochTask.build_generator() got an unexpected...

HenningBuhl

## ❓ Questions and Help #### What is your question? Are there any plans to release MMS ASR models with shared multilingual vocabulary (as described in the first paragraph of...

akreal

# Before submitting - [x] Was this discussed/approved via a Github issue? (no need for typos, doc improvements) - [x] Did you read the [contributor guideline](https://github.com/pytorch/fairseq/blob/main/CONTRIBUTING.md)? - [x] Did you...

chevalierNoir

fairseq-train: error: unrecognized arguments: --mask-multiple-length 10 --mask-stdev 10

6

## 🐛 Bug ### To Reproduce Steps to reproduce the behavior (**always include the command you ran**): 1. Run cmd ``` fairseq-train --fp16 $RESULT/quantized/fairseq-bin-data \ --task masked_lm --criterion masked_lm \...

cpark-dev