为了安全，强烈建议开启2FA双因子认证：User Settings -> Account -> Enable two-factor authentication！！！

Tags

Tags give the ability to mark specific points in history as being important

b4854

7c7f3b7f · ggml : skip intermediate .air file when compiling .metallib (#12247) · Mar 07, 2025
b4853

102ac189 · sync : ggml · Mar 07, 2025
b4851

68d0027f · ggml-cpu: faster AVX2 variant for IQ1_M (#12216) · Mar 07, 2025
b4849

8fad3c7a · server : Log original chat template parsing error (#12233) · Mar 07, 2025
b4848

7cf64f6b · sync: minja - support QwQ-32B (#12235) · Mar 07, 2025
b4847

5e2d57b2 · metal : simplify kernel arguments using a struct (#3229) (#12194) · Mar 07, 2025
b4846

f1648e91 · HIP: fix rocWMMA build flags under Windows (#12230) · Mar 07, 2025
b4837

e721c05c · HIP/CUDA: set the paramerter value in maintain_cuda_graph instead of replaceing it. (#12209) · Mar 06, 2025
b4836

57b6abf8 · android : fix KV cache log message condition (#12212) · Mar 06, 2025
b4835

94bb63e4 · opencl : fix buffer alignment (#12197) · Mar 06, 2025
b4834

f7924399 · opencl : fix `ulong` kernel args were set from `int` variables (#12174) · Mar 06, 2025
b4833

ed4ce0dd · opencl : fix profile-related errors (#12095) · Mar 06, 2025
b4832

07d15723 · ggml-cpu: Faster IQ1 mul_mat_vec on AVX2 using BMI2 instructions (#12154) · Mar 06, 2025
b4831

5e43f104 · SYCL: Disable f16 Unary OPs as not supported by the kernels (#12201) · Mar 05, 2025
b4830

16e4b22c · ggml : fix GGMLMetalClass ODR (#12200) · Mar 05, 2025
b4829

074c4fd3 · ci : add fetch-depth to xcframework upload (#12195) · Mar 05, 2025
b4827

fa31c438 · ci : fix xcframework artifact tag (#12191) · Mar 05, 2025
b4826

3ccbfe5a · ci : remove xframework upload (#12190) · Mar 05, 2025
b4824

a057897a · llama : add xcframework build script (#11996) · Mar 05, 2025
b4823

5bbe6a9f · ggml : portability fixes for VS 2017 (#12150) · Mar 04, 2025

1
…
16
17
18
19
20
21
22
23
24
…
178