From f2f877762a3b560cada2a02f730f1230a0af7923 Mon Sep 17 00:00:00 2001 From: AK Date: Tue, 11 Jul 2023 17:45:27 +0800 Subject: [PATCH 1/4] Update LOCAL.md --- LOCAL.md | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) diff --git a/LOCAL.md b/LOCAL.md index 69faa79..a399596 100644 --- a/LOCAL.md +++ b/LOCAL.md @@ -91,18 +91,21 @@ ``` Replace `"{PRETRAINED_MODEL}"` with one of `{CJ, CJE, C}` according to your previous model choice. Make sure you have a minimum GPU memory of 12GB. If not, change the argument `whisper_size` to `medium` or `small`. -11. Process all text data. + +10. Process all text data. If you choose to add auxiliary data, run `python preprocess_v2.py --add_auxiliary_data True --languages "{PRETRAINED_MODEL}"` If not, run `python3.8 preprocess_v2.py --languages "{PRETRAINED_MODEL}"` Do replace `"{PRETRAINED_MODEL}"` with one of `{CJ, CJE, C}` according to your previous model choice. -12. Start Training. + +11. Start Training. Run `python finetune_speaker_v2.py -m ./OUTPUT_MODEL --max_epochs "{Maximum_epochs}" --drop_speaker_embed True` Do replace `{Maximum_epochs}` with your desired number of epochs. Empirically, 100 or more is recommended. To continue training on previous checkpoint, change the training command to: `python finetune_speaker_v2.py -m ./OUTPUT_MODEL --max_epochs "{Maximum_epochs}" --drop_speaker_embed False --cont True`. Before you do this, make sure you have previous `G_latest.pth` and `D_latest.pth` under `./OUTPUT_MODEL/` directory. To view training progress, open a new terminal and `cd` to the project root directory, run `tensorboard --logdir=./OUTPUT_MODEL`, then visit `localhost:6006` with your web browser. -13. After training is completed, you can use your model by running: + +12. After training is completed, you can use your model by running: `python VC_inference.py --model_dir ./OUTPUT_MODEL/G_latest.pth --share True` -14. To clear all audio data, run: +13. To clear all audio data, run: ``` rm -rf ./custom_character_voice/* ./video_data/* ./raw_audio/* ./denoised_audio/* ./segmented_character_voice/* long_character_anno.txt short_character_anno.txt ``` From 1ba65b1b551fa4647949df8c5f8a7c9a8a81d302 Mon Sep 17 00:00:00 2001 From: AK Date: Tue, 11 Jul 2023 18:02:37 +0800 Subject: [PATCH 2/4] Update LOCAL.md Add Windows version delete training data command --- LOCAL.md | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/LOCAL.md b/LOCAL.md index a399596..61d5b33 100644 --- a/LOCAL.md +++ b/LOCAL.md @@ -106,6 +106,13 @@ 12. After training is completed, you can use your model by running: `python VC_inference.py --model_dir ./OUTPUT_MODEL/G_latest.pth --share True` 13. To clear all audio data, run: + ### Linux ``` rm -rf ./custom_character_voice/* ./video_data/* ./raw_audio/* ./denoised_audio/* ./segmented_character_voice/* long_character_anno.txt short_character_anno.txt ``` + ### Windos + ``` + del /Q /S .\custom_character_voice\* .\video_data\* .\raw_audio\* .\denoised_audio\* .\segmented_character_voice\* long_character_anno.txt short_character_anno.txt + ``` + + From 0cdb8554b58f49e9bbc5831c03ad607b7aad0cd9 Mon Sep 17 00:00:00 2001 From: AK Date: Tue, 11 Jul 2023 18:15:10 +0800 Subject: [PATCH 3/4] Update LOCAL.md Add remove "separated" folder's training data --- LOCAL.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/LOCAL.md b/LOCAL.md index 61d5b33..3a14d6c 100644 --- a/LOCAL.md +++ b/LOCAL.md @@ -108,11 +108,11 @@ 13. To clear all audio data, run: ### Linux ``` - rm -rf ./custom_character_voice/* ./video_data/* ./raw_audio/* ./denoised_audio/* ./segmented_character_voice/* long_character_anno.txt short_character_anno.txt + rm -rf ./custom_character_voice/* ./video_data/* ./raw_audio/* ./denoised_audio/* ./segmented_character_voice/* ./separated/* long_character_anno.txt short_character_anno.txt ``` ### Windos ``` - del /Q /S .\custom_character_voice\* .\video_data\* .\raw_audio\* .\denoised_audio\* .\segmented_character_voice\* long_character_anno.txt short_character_anno.txt + del /Q /S .\custom_character_voice\* .\video_data\* .\raw_audio\* .\denoised_audio\* .\segmented_character_voice\* .\separated\* long_character_anno.txt short_character_anno.txt ``` From ccaa1db0e360e704ef56ac42dfebeca6ddc4db50 Mon Sep 17 00:00:00 2001 From: AK Date: Tue, 11 Jul 2023 18:23:38 +0800 Subject: [PATCH 4/4] Update LOCAL.md modify the command from "python3.8" -> "python" --- LOCAL.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/LOCAL.md b/LOCAL.md index 3a14d6c..137fd9c 100644 --- a/LOCAL.md +++ b/LOCAL.md @@ -94,7 +94,7 @@ 10. Process all text data. If you choose to add auxiliary data, run `python preprocess_v2.py --add_auxiliary_data True --languages "{PRETRAINED_MODEL}"` - If not, run `python3.8 preprocess_v2.py --languages "{PRETRAINED_MODEL}"` + If not, run `python preprocess_v2.py --languages "{PRETRAINED_MODEL}"` Do replace `"{PRETRAINED_MODEL}"` with one of `{CJ, CJE, C}` according to your previous model choice. 11. Start Training.