Skip to content

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

Commits on main

log: trace logging for scheduler (#12961)
Daniel Hiltgen committed 1 year ago
408c2f9
Add Tool Call ID (#12956)
Grace committed 1 year ago
809b9c6
log: instrument CPU discovery timing (#12960)
Daniel Hiltgen committed 1 year ago
ba8c035
discovery: only retry AMD GPUs (#12894)
Daniel Hiltgen committed 1 year ago
27f1fde
d3b4b99
vulkan: enable flash attention (#12937)
Daniel Hiltgen committed 1 year ago
a477010
ggml: Increase maximum graph size
Jesse Gross committed 1 year ago
ef549d5
d2158ca
chore(gptoss): cleanup dead code (#12932)
Michael Yang committed 1 year ago
ce3eb0a
9a50fd5
ggml: Avoid cudaMemsetAsync during memory fitting
Jesse Gross committed 1 year ago
392a270
cpu: always ensure LibOllamaPath included (#12890)
Daniel Hiltgen committed 1 year ago
3bee3af
logs: catch rocm errors (#12888)
Daniel Hiltgen committed 1 year ago
8353799
7dd4862
win: avoid ID mixups on refresh (#12869)
Daniel Hiltgen committed 1 year ago
db973c8
26465fb
win: use copy for subprocess logs (#12864)
Daniel Hiltgen committed 1 year ago
88236bc
testing: test more models with tool calling (#12867)
Patrick Devine committed 1 year ago
76eb7d0
interleaved mrope (#12807)
Michael Yang committed 1 year ago
f67a6df
qwen3vl: enable flash attention by default (#12862)
Michael Yang committed 1 year ago
75e75d9
fix(cmd): unload model before removal (#12832)
Michael Yang committed 1 year ago
ed78e12
fix: qwen2.5vl, qwen3vl composite image (#12841)
Michael Yang committed 1 year ago
d432ade
06b3422
Update README.md (#12822)
Athiban Sharon committed 1 year ago
cbe1cf0
int: harden server lifecycle (#12835)
Daniel Hiltgen committed 1 year ago
c886471