merging with ttsds/datasets
Hi, it looks like we have worked on something similar in parallel: https://github.com/ttsds/datasets
Do you think there is scope to merge the two? I have only focused on voice cloning systems at the moment, since they are easier to evaluate.
You seem to have some information my version is missing, while I have a few details that aren't tracked here (e.g. number of parameters, hours of training data)
Good work there! This tracker was started at the beginning of the year on GitHub by @reach-vb .
Your table could be merged in here as another dataset, since you do not link embedded links. Unlike the table in README of this repo.
Could you create a legend for your last 8 columns that would describe each of them?
Also I chose not to use checkbox emojis as it became confusing at what capability cell you are looking at.
Your table could be merged in here as another dataset, since you do not link embedded links. Unlike the table in README of this repo.
Sorry, what do you mean by embedded links? The links to Repo, Hub, etc?
Could you create a legend for your last 8 columns that would describe each of them?
Will do and report back here when that's finished!
Also I chose not to use checkbox emojis as it became confusing at what capability cell you are looking at.
Yeah I guess I was a bit biased since it's basically taken from a paper I'm writing, in which checkboxes make more sense. Will consider changing it for huggingface/github.
I converted the markdown table to data.csv. You can switch to it in the dataset viewer.
https://huggingface.co/datasets/Pendrokar/open_tts_tracker/viewer/train_data
Used the following tool to convert Markdown table to CSV:
https://www.tablesgenerator.com/markdown_tables#
Sorry, what do you mean by embedded links? The links to Repo, Hub, etc?
Yeah those. Just tried, the Markdown format does not work as links for dataset viewer.
Yeah I guess I was a bit biased since it's basically taken from a paper I'm writing, in which checkboxes make more sense. Will consider changing it for huggingface/github.
You know what, nevermind, this was only an issue with a lot of rows and the first row not freezing in markdown table format.
I see you've created a legend. I've added it to the README of this dataset.