Default Branch

c0bc8591e8 · hexagon: fix Windows crash when op_poll is enabled (#26029) · Updated 2026-07-23 11:08:10 -05:00

Branches

128118fdbe · convert : use F32 for dequant of pack-quantized tensors · Updated 2025-11-06 20:59:32 -06:00    upstream-archive

3137
6

23b70f4f70 · Initial plan · Updated 2025-11-04 05:00:12 -06:00    upstream-archive

3165
1

d441c31b19 · metal : remove stray return · Updated 2025-11-02 10:24:00 -06:00    upstream-archive

3189
9

d7f794eadb · convert : avoid dequantizing mxfp4 for GPT-OSS · Updated 2025-10-24 06:56:26 -05:00    upstream-archive

3276
1

93fbd407f3 · Merge branch 'master' into compilade/convert-prequant · Updated 2025-10-23 13:23:12 -05:00    upstream-archive

3279
6

f0076dc5a0 · metal : adjust .get_alloc_size to be alloc friendly · Updated 2025-10-19 09:20:54 -05:00    upstream-archive

3309
1

96f9f391c7 · ggml : fix unaligned access in AMX code · Updated 2025-09-29 02:37:15 -05:00    upstream-archive

3489
1

a8b0089a5b · ggml : remove SVE paths · Updated 2025-09-28 12:26:03 -05:00    upstream-archive

3489
1

837b1b4563 · ggml : remove KQ mask padding · Updated 2025-09-28 10:10:17 -05:00    upstream-archive

3492
6

17ca6ed540 · Implement llama-pull tool · Updated 2025-09-20 11:25:21 -05:00    upstream-archive

3580
1

e83ef74733 · one less magic number · Updated 2025-09-20 00:58:36 -05:00    upstream-archive

3599
6

652d303b32 · metal : fuse add + rms · Updated 2025-09-18 08:29:25 -05:00    upstream-archive

3597
1

64c6dcbe6d · metal : make the NSG a function constant in mul_mv kernels · Updated 2025-09-18 03:31:59 -05:00    upstream-archive

3602
2

6045c5a263 · cont : put all buffers in the same virtual address space · Updated 2025-09-14 07:46:57 -05:00    upstream-archive

3638
2

3f62ee8bee · metal : back to a single queue per device · Updated 2025-09-09 09:06:46 -05:00    upstream-archive

3680
9

3f62ee8bee · metal : back to a single queue per device · Updated 2025-09-09 09:06:46 -05:00    upstream-archive

3680
9

7b717fb4b2 · Rewrite llama-run to use llama-server · Updated 2025-09-05 11:22:36 -05:00    upstream-archive

3717
1

9f2636b7dc · wip · Updated 2025-09-01 03:17:56 -05:00    upstream-archive

3766
1

4317d5abf5 · wip · Updated 2025-08-28 05:55:21 -05:00    upstream-archive

3800
1

dc2187d48d · ggml : fix SSM_SCAN for n_groups > 1 · Updated 2025-08-27 16:37:04 -05:00    upstream-archive

3805
1