diff options
| author | Pherkel | 2023-09-18 18:29:37 +0200 |
|---|---|---|
| committer | Pherkel | 2023-09-18 18:29:37 +0200 |
| commit | f6e6377d90aa4771bdae7c73364d7fe3ae6bf507 (patch) | |
| tree | 4907486cd22ea9f2edc02b7f550f5906d26d800a | |
| parent | f94506764bde3e4d41dc593e9d11aa7330c00e30 (diff) | |
small lm fixes
| -rw-r--r-- | swr2_asr/utils/data.py | 2 | ||||
| -rw-r--r-- | swr2_asr/utils/decoder.py | 8 |
2 files changed, 6 insertions, 4 deletions
diff --git a/swr2_asr/utils/data.py b/swr2_asr/utils/data.py index 74cd572..bb3233e 100644 --- a/swr2_asr/utils/data.py +++ b/swr2_asr/utils/data.py @@ -360,4 +360,4 @@ def create_lexicon(vocab_counts_path, lexicon_path): file.write(f"{word} ") for char in word: file.write(char + " ") - file.write("<SPACE>") + file.write("<SPACE>\n") diff --git a/swr2_asr/utils/decoder.py b/swr2_asr/utils/decoder.py index 1fd002a..38f218f 100644 --- a/swr2_asr/utils/decoder.py +++ b/swr2_asr/utils/decoder.py @@ -52,9 +52,11 @@ def get_beam_search_decoder( ) if not os.path.isdir(os.path.join(lang_model_path, f"mls_lm_{language}")): - url = f"https://dl.fbaipublicfiles.com/mls/mls_lm_{language}.tar.gz" - torch.hub.download_url_to_file(url, f"data/mls_lm_{language}.tar.gz") - _extract_tar("data/mls_lm_{language}.tar.gz", overwrite=True) + # check if zip file exists + if not os.path.isfile(f"data/mls_lm_{language}.tar.gz"): + url = f"https://dl.fbaipublicfiles.com/mls/mls_lm_{language}.tar.gz" + torch.hub.download_url_to_file(url, f"data/mls_lm_{language}.tar.gz") + _extract_tar(f"data/mls_lm_{language}.tar.gz", overwrite=True) tokens_path = os.path.join(lang_model_path, f"mls_lm_{language}", "tokens.txt") if not os.path.isfile(tokens_path): |