Refactor GGUF to speed-up inference #47779
gguf non quantized
783f1ef3
add support for q_k
a7fbe8f5
Merge branch 'main' into poc-gguf
9e57369f
kernels and config
e34177d7
Merge remote-tracking branch 'origin/poc-gguf' into poc-gguf
ad4e7c2c
better integration with kernels
5f48600a
small update
0c556c4e
Merge remote-tracking branch 'origin/poc-gguf' into poc-gguf
de3bf700
fix
1a101a1d
fix
1a3d470d
support for ggml-attn
fa45d579
update
a822ee02
Merge remote-tracking branch 'upstream/main' into poc-gguf
20e97611
fix
e946e7c2
update
8413aa69
fix
2abc18bc
kernels gated delta
b963ea10
Revert "kernels gated delta"
bf7d97c7
kernels gated delta agin
6a7de5d3
input permutation
39104425
Merge remote-tracking branch 'upstream/main' into poc-gguf
5d95dcfc
style
4e3ebbfc
update integration
31cec70a
up
ef373059
docs and bunch of fixes
3cd2049e
tokenizer for qwen3.5
8320a1ac
fix ci and doc
965029bb
cleaner integration
fac572b1
up
74dd50fb
fix
4d2c2e08
update
e6c7b422
fix
6a5d9a1a
up
b7267db9
serve fix
5eb6df28
doc
b155b5ea
leftover stuff
73179955
Merge remote-tracking branch 'origin/poc-gguf' into poc-gguf
0bb36abb
Merge remote-tracking branch 'upstream/poc-gguf' into poc-gguf
8bd1b773
Merge remote-tracking branch 'origin/poc-gguf' into poc-gguf
3e1ef578
minor fix
70033acb
Merge remote-tracking branch 'upstream/poc-gguf' into poc-gguf
f2f9370c
norm ! 11% speed increase
fa74f127
up
5bf39863
Merge remote-tracking branch 'upstream/main' into poc-gguf
728c5080
correct path !
e234ff4c
update
abdb5345
less verbose
f3cf4e72
SunMarc
enabled auto-merge 20 days ago
SunMarc
merged
f62dc9bf
into main 20 days ago
SunMarc
deleted the poc-gguf branch 20 days ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub