Skip to content
This repository was archived by the owner on Sep 10, 2025. It is now read-only.

Conversation

@malfet
Copy link
Contributor

@malfet malfet commented Apr 18, 2024

$10^{-5}$ is too cold for fdtype, whose max is 65504.0 which is less than 100000

Fixes #250

1e-5 is too cold for dtype whose max is 65504.0 which is less than 1e5

Fixes #250
@malfet malfet requested a review from mikekgfb April 18, 2024 04:11
@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Meta Open Source bot. label Apr 18, 2024
Copy link
Contributor

@mikekgfb mikekgfb left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you!

@malfet malfet merged commit e41c395 into main Apr 18, 2024
@malfet malfet deleted the malfet-patch-1 branch April 18, 2024 04:21
malfet added a commit that referenced this pull request Jul 17, 2024
$10^{-5}$ is too cold for fdtype, whose max is 65504.0 which is less than 100000

Fixes #250
malfet added a commit that referenced this pull request Jul 17, 2024
$10^{-5}$ is too cold for fdtype, whose max is 65504.0 which is less than 100000

Fixes #250
malfet added a commit that referenced this pull request Jul 17, 2024
$10^{-5}$ is too cold for fdtype, whose max is 65504.0 which is less than 100000

Fixes #250
malfet added a commit that referenced this pull request Jul 17, 2024
$10^{-5}$ is too cold for fdtype, whose max is 65504.0 which is less than 100000

Fixes #250
malfet added a commit that referenced this pull request Jul 17, 2024
$10^{-5}$ is too cold for fdtype, whose max is 65504.0 which is less than 100000

Fixes #250
malfet added a commit that referenced this pull request Jul 17, 2024
$10^{-5}$ is too cold for fdtype, whose max is 65504.0 which is less than 100000

Fixes #250
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

CLA Signed This label is managed by the Meta Open Source bot.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

float16 model w/ various quant options will generate garbage output

4 participants