Alex Cheema comments

Results 404 comments of


                                            Alex Cheema

If BF16 isn't supported, it should autocast to float

Watching this. I think some users might have run into this issue using exo: https://github.com/exo-explore/exo/issues/152

[#253] Removed extra </link> tag from index.html

Thanks for this! This should be a one line change. Please fix conflicts and keep formatting the same as it was.

[#253] Removed extra </link> tag from index.html

lgtm after conflicts resolved.

Add Dutch translation

Not a fan of adding a dependency on watchdog for this. We don't need instant config application. Periodically checking every 5 secs for config changes is fine.

Closing in favour of https://github.com/exo-explore/exo/pull/383 Please contact me for a $100 bounty with your ethereum address as I'd like to reward good experimentation: [[email protected]](mailto:[email protected])

Unable to format index.html

good catch. assigned. go for it!

(Bug?) `ValueError: size mismatched, can't reshape self.shape=(1, 25, 128256, 4096) -> new_shape=(1, 25, 32, 128)`

I don't see anything obviously wrong with your setup. It looks all correct. The logs suggest perhaps some networking issues. The fact that it generates some tokens then stops also...

System requirements

Just updated the README since exo only requires that you have enough memory across all your devices and is designed to run on anything that has a Python runtime.

System requirements

Was there anything else you had in mind here @shad-ct

Question about Exo

Hey, thanks for reporting this. The reason here is likely: ollama is using a 4-bit quantized model. However exo is using the fp16 unquantized model. I have created an issue...