Default Branch

17313a4226 · Fix cuda memory error for Qwen3 non-quantized (#2987) · Updated 2025-06-07 14:02:58 +00:00

Branches

cdbdb4af9c · Update yew-agent requirement from 0.2.0 to 0.3.0 · Updated 2024-01-10 14:14:03 +00:00    huggingface

712
1

c2261d0222 · Merge. · Updated 2024-01-07 19:27:33 +00:00    huggingface

715
4

9cd0cc1f65 · Ignore rotary for mistral. · Updated 2024-01-05 20:55:13 +00:00    huggingface

722
6

289c57d600 · Removing metal fences. Increases performance substantially on m1 pro. · Updated 2023-12-28 16:31:07 +00:00    huggingface

751
1

5edb07a5b1 · mps matmul · Updated 2023-12-20 01:53:18 +00:00    huggingface

820
1

03641293ee · Clippy pass. · Updated 2023-12-18 14:22:43 +00:00    huggingface

803
0
Included

cf27868b57 · More cleanup. · Updated 2023-12-15 00:44:22 +00:00    huggingface

820
0
Included

1f23cea90c · MFA · Updated 2023-12-13 15:09:20 +00:00    huggingface

833
3

a9d0657432 · Better version ? · Updated 2023-12-13 11:09:20 +00:00    huggingface

828
0
Included

a0282751d5 · Tmp. · Updated 2023-12-11 18:51:46 +00:00    huggingface

830
1

ce0783d9ff · Stash for debugging · Updated 2023-12-10 12:11:53 +00:00    huggingface

833
2

03ad494fcd · Tweak the basic example to show how to implement sort. · Updated 2023-11-30 08:01:42 +00:00    huggingface

835
1

c93a17694b · Speeding up copies using blit. · Updated 2023-11-19 22:00:10 +00:00    huggingface

877
32

c65f68e988 · Tmp gemm. · Updated 2023-11-19 19:43:59 +00:00    huggingface

877
29

7e49e0af96 · Tmp for allocator. · Updated 2023-11-16 11:50:41 +00:00    huggingface

877
27

7e49e0af96 · Tmp for allocator. · Updated 2023-11-16 11:50:41 +00:00    huggingface

877
27

e8c1c31245 · Tmp commit for the heap experiment (heap is indeed decreasing). · Updated 2023-11-14 16:04:23 +00:00    huggingface

877
26

d9c1f7e201 · Fixed matmul (display still broken without casting back to CPU first? ) · Updated 2023-11-10 19:09:25 +00:00    huggingface

886
8

eb24875856 · Reworked affine and it works ? No idea how it's different. · Updated 2023-11-08 01:37:20 +00:00    huggingface

934
28

9a27f11c3f · Adding tons of profiling and removing the metal allocation (still slow). · Updated 2023-11-02 16:48:07 +00:00    huggingface

934
8