Skip to content

Fix conversion of unnormalized BF16->BF16 weights #7843

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 10 commits into from
Aug 2, 2024
Merged

Fix conversion of unnormalized BF16->BF16 weights #7843

Changes from 1 commit
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
6a52bfe
add truncate_bf16
CISC Jun 10, 2024
46054d1
truncate intermediate fp32 if converting bf16 to bf16
CISC Jun 10, 2024
069369f
fix masking in __compute_fp32_to_bf16
CISC Jun 14, 2024
225ec48
np.int16 no longer used
CISC Jun 14, 2024
e8e2b7e
missing cast and additional numpy 2.x fix
CISC Jun 27, 2024
5b67a6c
ggml-impl : do not flush bf16 subnormals to zero
compilade Jun 28, 2024
675a741
Merge branch 'master' into convert-bf16-fix
compilade Jun 28, 2024
2b74648
Merge branch 'master' of github.com:ggerganov/llama.cpp into convert-…
CISC Aug 1, 2024
dc05154
missed prototype update in merge
CISC Aug 1, 2024
3a3a752
merge cleanup
CISC Aug 1, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Loading
You are viewing a condensed version of this merge commit. You can view the full changes here.