brain dead

2023-03-07 05:43:26 +07:00 · 2023-03-07 05:43:26 +07:00 · df5ba634c0
parent 2726d98ee1
commit df5ba634c0
1 changed files with 4 additions and 0 deletions
--- a/src/utils.py
+++ b/src/utils.py
@ -1201,6 +1201,10 @@ def prepare_dataset( files, outdir, language=None, skip_existings=False, progres
 	unload_whisper()

 	joined = "\n".join(transcription)
+	if not skip_existings:
+		with open(f'{outdir}/train.txt', 'w', encoding="utf-8") as f:
+			f.write(joined)
+
 	return f"Processed dataset to: {outdir}\n{joined}"

 def calc_iterations( epochs, lines, batch_size ):