Default Branch

b2cb4512c5 · Create parameters overview (#1269) · Updated 2026-02-20 07:20:56 +01:00

Branches

60043847c2 · Add missing gguf-py constants · Updated 2025-05-25 08:53:53 +02:00    git

4211
3713

ceed25bf8a · Remove GGML_IQK_MUL_MAT option · Updated 2025-05-25 07:18:42 +02:00    git

4211
3711

dad5464e34 · Merge remote-tracking branch 'origin/main' into s6/fp8_native · Updated 2025-05-24 11:45:19 +02:00    git

4211
3715

3fe6c0a6e1 · Very slightly faster iq4_kt TG · Updated 2025-05-24 07:08:32 +02:00    git

4211
3714

858f2a55a5 · Arghhh · Updated 2025-05-23 15:26:03 +02:00    git

4211
3711

193a15b465 · Fix bug in MMVQ kernel · Updated 2025-05-23 11:04:53 +02:00    git

4211
3709

745973f294 · Fix typo in non-AVX2 code branch · Updated 2025-05-23 11:00:23 +02:00    git

4211
3708

b79be8a191 · Fix bug in MMVQ kernel · Updated 2025-05-23 07:02:47 +02:00    git

4211
3722

1375c21deb · Slighty faster iq2_kt · Updated 2025-05-21 16:16:51 +02:00    git

4211
3772

82871cc2a3 · Another attempt to fix the illegal memory access bug · Updated 2025-05-20 16:04:26 +02:00    git

4211
3704

5252e95552 · Clearing padding · Updated 2025-05-20 15:59:31 +02:00    git

4211
3704

0943331ec9 · Fix q6_0 K cache · Updated 2025-05-20 09:46:36 +02:00    git

4211
3739

97ce7edb62 · Disable multi-add for now · Updated 2025-05-18 07:36:12 +02:00    git

4211
3702

8c56fb3a72 · Option to enable disable the IQK CPU FA kernels · Updated 2025-05-17 10:03:10 +02:00    git

4211
3701

d7ebb3eae4 · Zen4: faster PP for iq2_ks · Updated 2025-05-17 09:22:38 +02:00    git

4211
3701

9dd452d4cb · Fix iq5_ks on NEON · Updated 2025-05-16 16:26:07 +02:00    git

4211
3703

177dd173d6 · Fix IQ6_K on AVX2 · Updated 2025-05-16 15:49:26 +02:00    git

4211
3701

349a697654 · Adding forgotten template instance for iq5_ks · Updated 2025-05-15 15:48:20 +02:00    git

4211
3697

a7ceba3dc6 · iq5_ks: Metal dot product · Updated 2025-05-15 14:48:29 +02:00    git

4211
3705

ab6077718f · Fix standard attention on the CPU · Updated 2025-05-15 07:40:47 +02:00    git

4211
3695