fairseq

fairseq copied to clipboard

## ❓ Questions and Help ### What have you tried? In fairseq/examples/wav2vec, I completed the "Prepare training data manifest" step, aka. ``` python examples/wav2vec/wav2vec_manifest.py /path/to/waves --dest /manifest/path --ext $ext --valid-percent...

Haotian-Fu

Currently, there is no information about lexicons for wav2vec 2.0 decoding in the README. I added both a download link for KenLM lexicon (taken from #2502) and the corresponding argument...

AndreasPlt

Changed metrics import location on speech_dlm_criterion

1

Fixes issue #5366. Fixes deprecated `from fairseq import metrics` changing the import with the new location `from fairseq.logging import metrics` on the `speech_dlm_criterion` used in the `fairseq-train` run on the...

Efrainq07

## 🐛 Bug Hello, I've been attempting to collect code coverage data by running the [CI process](https://github.com/facebookresearch/fairseq/actions) in your repository. However, I encountered a problem where the `build.yml` workflow is...

yunbolyu

#### What is your question? Let's say I have a base config called "roberta_test.yaml": ``` common: fp16: true log_format: json log_interval: 200 checkpoint: no_epoch_checkpoints: true #only keep best and last...

henrycharlesworth

## 🐛 Bug ype "help", "copyright", "credits" or "license" for more information. >>> from fairseq.modules import LayerNorm, MultiheadAttention Traceback (most recent call last): File "", line 1, in File "/home/jiahui/workspace/nmt/bert_nmt/fairseq/fairseq/modules/__init__.py",...

jiaohuix

Can NLLB deal with formatting such as Italics, Bold, underline, new sentence tokens etc? Ie. "I am asleep" where the word am will be bold in the translation or surrounded...

sj681

usage: fairseq-hydra-train [--help] [--hydra-help] [--version] [--cfg {job,hydra,all}] [--package PACKAGE] [--run] [--multirun] [--shell-completion] [--config-path CONFIG_PATH] [--config-name CONFIG_NAME] [--config-dir CONFIG_DIR] [--info] [overrides [overrides ...]] fairseq-hydra-train: error: unrecognized arguments: --layer-type conformer --attn-type espnet...

Minspeech

## 🐛 Bug Following the [SimpleLSTM tutorial](https://fairseq.readthedocs.io/en/latest/tutorial_simple_lstm.html#training-the-model), everything works fine until `fairseq-train` is used as follows: `fairseq-train data-bin/iwslt14.tokenized.de-en \ --arch tutorial_simple_lstm \ --encoder-dropout 0.2 --decoder-dropout 0.2 \ --optimizer adam --lr...

Efrainq07

## ❓ Questions and Help ### Before asking: 1. search the issues. 2. search the docs. #### What is your question? fairseq-hydra-train: error: unrecognized arguments: --user-dir --arch linformer_roberta_base #### Code...

GoGolden