Default Branch

292767afb4 · CI: fix win arm build (#12502) · Updated 2025-10-04 20:46:45 +02:00

Branches

2c2f4deaa9 · openai: refactor to split compat layer and middleware · Updated 2025-10-05 23:18:56 +02:00

4
1

b91c1f6749 · update tests · Updated 2025-10-03 23:49:49 +02:00

17
4

c3fa8b2f54 · clean up the renderer · Updated 2025-10-03 21:13:13 +02:00

7
1

198e7a02d6 · llm: Allow overriding flash attention setting · Updated 2025-10-02 00:28:20 +02:00

62
2

2047dd2b38 · add tests for new code paths · Updated 2025-10-01 19:06:41 +02:00

6
6

ff1b9bb2f3 · build: call find_package to instantiate library paths · Updated 2025-09-30 21:58:37 +02:00

7
1

f944382424 · lint · Updated 2025-09-30 05:10:38 +02:00

7
3

abc6a300de · model: tweak renderer for qwen3coder · Updated 2025-09-29 00:41:13 +02:00

0
1

76cc9135ad · ggml: Preallocate CUDA pool memory · Updated 2025-09-27 00:08:41 +02:00

8
3

c5cd7fbead · works for 3.1, but regression in 3??? · Updated 2025-09-26 23:35:06 +02:00

6
4

6fd37e573f · start of deepseek v3.1 stuff · Updated 2025-09-25 19:55:21 +02:00

6
1

909232168d · deepseek tests · Updated 2025-09-23 23:08:17 +02:00

12
1

ffaf2e7916 · update tests · Updated 2025-09-22 23:25:51 +02:00

22
3

4ef2b2852d · server: serve original error for remote models · Updated 2025-09-21 01:46:32 +02:00

19
1

220a0da37e · simplify expand path · Updated 2025-09-19 22:12:23 +02:00

22
1

b47b9d9063 · s/From*Slice/From*s/ · Updated 2025-09-16 18:50:59 +02:00

40
1

c10a40db99 · parser: tidy up parameter/message parsing · Updated 2025-09-16 03:09:05 +02:00

42
1

92f77a32fc · gemma3: make embedding non-causal · Updated 2025-09-16 00:25:23 +02:00

44
1

7eb0ff7dca · set_rows · Updated 2025-09-15 22:01:18 +02:00

45
1

e8c1f7a54d · add pre:, suf: to tags · Updated 2025-09-12 23:01:25 +02:00

49
1