Skip to content

Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

Commits on main

fixes for maverick
Michael Yang committed 1 year ago
7ba9fa9
chunked attention
Michael Yang committed 1 year ago
8bf11b8
connect vision to text
Michael Yang committed 1 year ago
470af8a
image processing
Michael Yang committed 1 year ago
178761a
llama4
Michael Yang committed 1 year ago
f0c66e6
fix test
Michael Yang committed 1 year ago
54055a6
explicitly decode maxarraysize 1024
Michael Yang committed 1 year ago
340448d
fix parameter count
Michael Yang committed 1 year ago
ced7d0e
default slice values
Michael Yang committed 1 year ago
a0dba0f
update comment
Michael Yang committed 1 year ago
5e20b17
fix token type
Michael Yang committed 1 year ago
d26c18e
zero means zero
Michael Yang committed 1 year ago
8d376ac
convert: use -1 for read all
Michael Yang committed 1 year ago
dc1e81f
generic ggml.array
Michael Yang committed 1 year ago
5d02791
fix superfluous call to WriteHeader
Michael Yang committed 1 year ago
214a767
convert: change to colmajor
Michael Yang committed 1 year ago
4892872
ci: silence deprecated gpu targets warning
Michael Yang committed 1 year ago
0b9198b
llama: update to commit 2016f07b (#10352)
Jeffrey Morgan committed 1 year ago
e9e5f61
server: improve spacing for JSON grammar (#10131)
Parth Sareen committed 1 year ago
11dde41
llama: remove model loading for grammar (#10096)
Parth Sareen committed 1 year ago
a53d744
increase default context length to 4096 (#10364)
Devon Rifkin committed 1 year ago
424f648
2eb1fb3
cmd: add support for escaping ~ in filepath (#10339)
greengrass821 committed 1 year ago
0806521
create tempdir in models directory
Michael Yang committed 1 year ago
88738b3
arange
Michael Yang committed 1 year ago
40b8fdb
09bb2e3