Default Branch

33ca0dcb9d · ggml-hip : add -fno-finite-math-only alongside -ffast-math (#25373) · Updated 2026-07-07 11:27:50 +00:00

Branches

66a66a05a8 · readme : add notice about new file format · Updated 2023-08-21 19:42:14 +00:00    kanshan

8895
253

6a9e6375b5 · gguf.py : indentation · Updated 2023-08-17 18:53:15 +00:00    kanshan

8910
205

28046d1e52 · Merge and update · Updated 2023-08-08 21:36:11 +00:00    kanshan

8963
12

511055722e · undo formatting · Updated 2023-07-28 06:09:14 +00:00    kanshan

8992
26

af1c9966c8 · gguf : start write tensor info · Updated 2023-07-27 07:32:31 +00:00    kanshan

8992
15

d273bfd2c9 · allocator: cleanup, more comments · Updated 2023-07-22 13:05:24 +00:00    kanshan

9063
21

d45c1631bc · metal : rewrite to fit new backend interface correctly (WIP) · Updated 2023-07-20 19:51:19 +00:00    kanshan

9063
18

0492363137 · mpi : fix after master merge · Updated 2023-07-09 19:23:04 +00:00    kanshan

9094
21

26cc1bd7a2 · llama : uniform variable names + struct init · Updated 2023-07-05 20:22:17 +00:00    kanshan

9111
4

ff6e39f138 · use javascript generators as much cleaner API · Updated 2023-07-05 19:03:01 +00:00    kanshan

9124
20

f46db27ea0 · ci : disable FMA on Mac OS · Updated 2023-07-05 15:29:08 +00:00    kanshan

9121
5

5cc672a9a5 · metal : try to utilize more of the shared memory using smaller views · Updated 2023-06-26 19:23:04 +00:00    kanshan

9158
1

78fafcaf10 · ggml : do not use _GNU_SOURCE gratuitously · Updated 2023-06-25 14:21:02 +00:00    kanshan

9166
1

20054a38c1 · Fix directory name · Updated 2023-05-26 23:00:08 +00:00    kanshan

9316
1

a1cdd29cd2 · ggml : rms_norm in chunks · Updated 2023-05-20 07:15:54 +00:00    kanshan

9337
2

95dc4d7270 · Merge 'origin/master' into steering · Updated 2023-05-19 20:19:57 +00:00    kanshan

9339
9

40ec4882c4 · ggml : use F16C conversion when available · Updated 2023-05-17 17:05:51 +00:00    kanshan

9348
1

a3e6d62283 · cuda : alternative q4_q8 kernel · Updated 2023-05-12 14:02:39 +00:00    kanshan

9382
8

e116eb638c · ggml : speed-up Q5_0 + Q5_1 at 4 threads · Updated 2023-05-11 15:51:56 +00:00    kanshan

9384
20

4baa85633a · Fix build · Updated 2023-05-07 01:44:07 +00:00    kanshan

9392
5