How to inference on other <midi, text> files using the opencpop phone_set.json?

[SkyAndCloud]

Hi, thanks for the great work! I want to inference on my own <midi, text> files. I generated the corresponding meta.json and tried to binarize it. But the binarizer can only generate phone_set.json based on the input files, which will be incompatible with the pretrained checkpoint (since it was pretrained on opencpop with opencpop phone_set.json). A solution is to preprocess the customized testset with the original opencpop meta.json to obtain the full phone_set.json (i.e., the same phoneme dict). However, it is too inconvenient to do that thing.

[MrZixi]

Check this closed issue and my reply. We will upload this phone_set file to the repo later. https://github.com/MoonInTheRiver/DiffSinger/issues/15#issuecomment-1041207925

[SkyAndCloud]

How to binarize the new testset using the existing phone_set.json rather than generating a new phone_set?

[MrZixi]

How to binarize the new testset using the existing phone_set.json rather than generating a new phone_set?

See this line in binarize.py https://github.com/MoonInTheRiver/DiffSinger/blob/69ccf417e82834b6f4fc474e046f3bbaa79b3827/data_gen/singing/binarize.py#L104 Change the "reset_phone_dict" in your config file to "false" and place the existing phone_set.json in the "binary_data_dir". I think this will do.

[Wayne-wonderai]

Hello, I have issue as I try to use another english dataset. And I'm wondering why Inference from packed test set can work (CUDA_VISIBLE_DEVICES=0 python tasks/run.py --config usr/configs/midi/e2e/opencpop/ds100_adj_rel.yaml --exp_name $MY_DS_EXP_NAME --reset --infer) but inference model from raw input (python inference/svs/ds_e2e.py --config usr/configs/midi/e2e/opencpop/ds100_adj_rel.yaml --exp_name $MY_DS_EXP_NAME) needs same phoneme set size?