KaiTao Song

> 请问一下，songmass的READ.md文件，第四部分 Evaluation中提到的这四个文件是处理得到的吗？还是需要我自己添加呢？ 谢谢解惑 `LYRIC=lyric.gt # The lyric file of ground truth` `MELODY=melody.gt # The melody file of ground truth` `HYPOS=hypo.txt # The generated result in fairseq format` `SONG_ID=song_id_test.txt #...

> 首先非常感谢各位能开源这么好的工具！ 我尝试了一下songmass的旋律生成歌词的功能.没有找到可以直接给一段旋律，然后生成歌词的方案。我现在的操作流程是. 制作旋律的文件和歌词文件，然后把他们当做验证集，使用验证集的方式来出结果. 具体流程如下: ` fairseq-preprocess \ --user-dir mass \ --task xmasked_seq2seq \ --source-lang lyric --target-lang melody \ --trainpref $para_data_dir/train --validpref $para_data_dir/t3 \ --destdir $save_dir \ --srcdict $para_data_dir/dict.lyric.txt \...

@GragonWing we have provide the data processing that includes the split of training/valid/test. You just need to consider valid(test).lyric(melody) as the ground truth.

Our data processing pipeline includes the split of training/valid/test. Please use valid or test data as ground truth.

Based on the error descriptions, you may need to install nvidia driver. Can you detect GPU card by running `nvidia-smi`?

Can you use WSL on the Windows platform?

Thanks. The current version of JARVIS is to connect AI models from HugigngFace, rather than the Internet.

Hi, we are still working on this project and release a [langchain version](https://github.com/langchain-ai/langchain/tree/master/libs/experimental/langchain_experimental/autonomous_agents/hugginggpt) recently. We will take your suggestions into the consideration and please be patient.

@ssdasgupta We are currently working with our labeling teams to iteratively improve the quality of this dataset and our legal team to ensure compliance of the dataset release. We will...