Alexander Borzunov

@PicoCreator The error caused by the `bitsandbytes` library that doesn't support macOS at the moment. This library is optional for hivemind and now Petals (once #180 is merged), so you...

@ineiti The message you see is a part of the p2pd's `-help` outputs. It is shown when the daemon encounters some unknown arguments. Thus, I suppose there is some kind...

Hi @PicoCreator @ineiti @vrosca @LuciferianInk, We've shipped native macOS support in #477 - both macOS clients and servers (including ones using Apple M1/M2 GPU) now work out of the box....

Hi @bionicles, A small addition to the @mryab's response - while Petals does not support training from scratch, both Petals and SWARM are based on [hivemind](https://github.com/learning-at-home/hivemind), our library for training...

Hi @allenyllee! We'd like to start with a centralized incentive system since it is easier to develop and maintain (so we have more time to focus on the ML-side of...

Hi @lonnietc, Please take a look at our publications that contain experiments for different algorithms implemented in hivemind: https://github.com/learning-at-home/hivemind#citation (take a look at the newest papers in "Additional publications" too)....

Hi @chavinlo! Thanks for the feature request! We're busy with [another project](https://petals.ml/) built on top of hivemind at the moment, but will consider adding this feature when we get back...

Hi @ikmckenz! Thank you for this PR! We'll merge it once we decide that the next version will be 1.2.0. I've rerun the test, it should be all good now.

Hi @TomLBZ @redcap3000 @edugamerplay1228, This may be an issue with DNS/IPv6 addresses present among the default [bootstrap peers](https://github.com/bigscience-workshop/petals/blob/main/src/petals/constants.py#L3). Can you please try again with this option (this uses IPv4 addresses...

We discussed that we may revive this feature for loading NF4-pre-quantized weights for Llama 2 and Stable Beluga 2.