llama.cpp
Add qwen2moe
#6074
Merged
Overview
Commits
11
Changes
View On GitHub

Add qwen2moe #6074

ggerganov merged 11 commits into ggml-org:master from simonJJJ:add_qwen2moe
simonJJJ
simonJJJ support qwen2moe
56a38c46
simonJJJ resolve-conflicts
43c2c136
ggerganov
ggerganov commented on 2024-03-15
simonJJJ
sorasoras
andy-zhangtao
ggerganov
jpmottin
foldl
slaren
ggerganov
simonJJJ merge qwen2moe
edd8e2e5
simonJJJ
simonJJJ simonJJJ requested a review from ggerganov ggerganov 2 years ago
ggerganov
ggerganov commented on 2024-04-15
phymbert
phymbert commented on 2024-04-15
simonJJJ fix-review
4256fe6a
simonJJJ simonJJJ requested a review from ggerganov ggerganov 2 years ago
ggerganov metal : support unary ops for nelements % 4 != 0
70482f58
ggerganov
ggerganov approved these changes on 2024-04-15
ggerganov ggerganov requested a review from slaren slaren 2 years ago
ggerganov metal : require contiguousness for float4 unary kernels
7c1ab981
ggerganov metal : require contiguousness for float4 unary kernels (cont)
00102800
slaren
slaren commented on 2024-04-15
slaren
slaren commented on 2024-04-15
simonJJJ fix-review
7355ca84
github-actions
ggerganov names : for brevity "SHARED_EXP" -> "SHEXP"
f88e6844
ggerganov
simonJJJ
simonJJJ
ggerganov llama : reuse build_moe_ffn()
de355511
ggerganov
slaren
slaren commented on 2024-04-16
slaren
slaren approved these changes on 2024-04-16
ggerganov llama : add model type name
245565fc
ggerganov ggerganov merged f4dea7da into master 2 years ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Terms Of Use Privacy Policy Refund Policy Impressum