Skip to content

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

Commits on main

fix(integration): check truncated length (#12337)
Michael Yang committed 1 year ago
ceac416
convert: convert bf16 vision weights to fp16 (#12324)
Patrick Devine committed 1 year ago
2717dce
engine: add remote proxy (#12307)
Patrick Devine committed 1 year ago
8b89493
fix: multi-cuda version skew (#12318)
Daniel Hiltgen committed 1 year ago
9c5bf34
fix(llama): other llama flavours (#12308)
Michael Yang committed 1 year ago
564b558
prefer ollama engine for qwen3 (#12310)
Michael Yang committed 1 year ago
a417ac9
logutil: fix source field (#12279)
Michael Yang committed 1 year ago
b225508
use split activations when possible (#12293)
Michael Yang committed 1 year ago
ad95d5b
embed: cleanup (#12299)
Michael Yang committed 1 year ago
c253433
fix: fix CUDA detection for older GPUs (#12300)
Beshoy Girgis committed 1 year ago
a1cff89
doc: show how to clear the cgo cache (#12298)
Daniel Hiltgen committed 1 year ago
93c64ea
model: implement bert in ollama engine (#9080)
Michael Yang committed 1 year ago
3f6642f
batch: use tensors for outputs (#12185)
Michael Yang committed 1 year ago
6f71171
address comments
Devon Rifkin committed 1 year ago
472feec
add qwen3-coder tool support
Devon Rifkin committed 1 year ago
4799194
92b96d5
tests: tighten up a few flaky tests (#12271)
Daniel Hiltgen committed 1 year ago
44a6792
e4ce683
2621412
CI: fix windows cuda build (#12246)
Daniel Hiltgen committed 1 year ago
61fb912
llm: Enable new memory estimates by default
Jesse Gross committed 1 year ago
eb10390
feb18cd
cmd: use slices.Contains to simplify code (#12249)
fengyuchuanshen committed 1 year ago
8a7e205