Default Branch

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

Branches

ed353eb76d · revert some changes · Updated 2025-05-17 03:46:18 +00:00

8
2

5ed764213d · Add dtype size to benchmark throughput calculation · Updated 2025-05-06 08:40:16 +00:00

63
3

3b24f8f302 · Add metal precompilation via build.rs · Updated 2025-04-17 13:56:52 +00:00

39
1

6381023982 · Adding cuda feature for easier integration with extensions. · Updated 2025-04-15 14:28:51 +00:00

73
4

8e62723b2d · Set the algo. · Updated 2025-04-13 18:58:18 +00:00

47
1

83bbbc6265 · Deploy f3a73f80d1 to gh-pages · Updated 2025-04-13 14:47:49 +00:00

2385
1

543b5b5898 · Update for the latest cudarc. · Updated 2025-04-11 12:02:41 +00:00

57
5

5341bf4cd5 · Fixes for clippy 1.86. · Updated 2025-04-03 17:30:20 +00:00

65
9

ec6d7ca773 · Cudarc static-linking enabled. · Updated 2025-03-29 08:27:53 +00:00    huggingface

73
3

2c0f6b008e · Fixing order. · Updated 2025-03-28 10:43:33 +00:00    huggingface

73
2

2e273ddf31 · Fixing the mkl dependency hell. · Updated 2025-03-27 17:01:21 +00:00    huggingface

73
1

777ad954eb · Avoid some clippy lints on 1.85. · Updated 2025-02-21 09:39:55 +00:00    huggingface

104
4

10b2e693ff · Add the SmolLM2 models. · Updated 2024-11-03 15:42:02 +00:00    huggingface

168
4

ab12425bff · Another tweak. · Updated 2024-09-26 08:14:53 +00:00    huggingface

220
3

5221146cfa · Cuda quantization padding fix. · Updated 2024-09-25 21:35:16 +00:00    huggingface

220
6

42c702a023 · Update cudarc to 0.12.1. · Updated 2024-09-22 18:16:57 +00:00    huggingface

223
11

7ec4f64d38 · Attempt at fixing M1/M2 metal async copy bug · Updated 2024-09-06 13:59:35 +00:00    huggingface

271
10

9105aa4390 · batched gemm work · Updated 2024-07-26 16:53:58 +00:00    huggingface

304
5

56a1b7d97e · Apply rustfmt. · Updated 2024-06-04 20:47:20 +00:00    huggingface

324
18

84cd5158ad · Update gemm requirement from 0.17.0 to 0.18.0 · Updated 2024-06-01 06:19:34 +00:00    huggingface

330
1