Open lk-geimfari opened 2 weeks ago
I'm currently working on this, but help is welcome.
Hi @lk-geimfari , If you are not working on this I could create this plugin from your example code.
Hi @lk-geimfari , If you are not working on this I could create this plugin from your example code.
Sure!
Hi @lk-geimfari , couple of thoughts. For the accessibility should use free LLM to generate the results like random text, image, and audio based on the prompt. It will unlock lot of users who can test us without any subscription to chatgpt plus or to even chatgpt.
Hi @lk-geimfari , couple of thoughts. For the accessibility should use free LLM to generate the results like random text, image, and audio based on the prompt. It will unlock lot of users who can test us without any subscription to chatgpt plus or to even chatgpt.
I'd prefer to avoid a lot of optional deps, but overall the idea is a good one.
@lk-geimfari check this draft PR. https://github.com/Rangeeshar/mimesis/pull/1 Let me know if you need any other change on top of this. I can make model dynamic and other params dynamic which could be passed to hugging face by our users.
After this I can create an Image based one and audio based one. Once we push this out. I could create a plugin which you can attach any AI in future.
I like it! Looks good. A good idea would be to use try/except when importing libs that will be optional. See the example in plugins/pytest.py. Also, do not foget include optional deps to pyproject.toml so we will be able to install all optional AI-related deps like this:
> pip install mimesis[ai]
Also, don't forget the tests. Tests are good, we love them.
Hi man @lk-geimfari , added a draft PR. Please add some comments If you think something needs a change. https://github.com/lk-geimfari/mimesis/pull/1571
If something the user needs is not supported by Mimesis, they can generate it using OpenAI, like this:
This is just an example; the API is not finalized yet.