The repetition issue is better, it isn't fixed completely but now it's much better and won't repeat like before, also, the model has been improved based on the feedback you sent
I'm working on the PAI-002 which has much better quality than the first version of PAI, the early testers have already tested it, and in the next update you can expect a new model with a very high quality (maybe the same as or better than gpt-3.5-turbo) You can send your Feedback to 🙏𝐅𝐞𝐞𝐝𝐁𝐚𝐜𝐤, i will work on them in the next version of PAI
The only thing you have to change is replacing "beta" with "rp" for example, if you are using (https://api.pawan.krd/pai-001-beta/v1) change it to (https://api.pawan.krd/pai-001-rp/v1) or if you are using (https://api.pawan.krd/pai-001-light-beta/v1) change it to (https://api.pawan.krd/pai-001-light-rp/v1)
Developers can use the normal model names like pai-001 and pai-001-light, which means only removing "beta" from the model name in your implementations
you can use (https://api.pawan.krd/pai-001/v1/chat/completions) or (https://api.pawan.krd/pai-001-light/v1/chat/completions) to force override the model name
Also, I like to Thank all the @Supporters who donated to this project, this project is running on your donations ❤️🔥
If you like this project please consider donating, Patreon: https://patreon.com/pawanosman ❤️