Admin message
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
b4078
2a82891a
·
speculative : fix out-of-bounds access (#10289)
·
Nov 14, 2024
b4077
af148c93
·
vulkan: Optimize binary ops (#10270)
·
Nov 14, 2024
b4076
66798e42
·
vulkan: Use macros to make the mat mul pipeline creation more concise (#10259)
·
Nov 13, 2024
b4075
fb4a0ec0
·
llama : propagate the results of `graph_compute` (#9525)
·
Nov 13, 2024
b4073
1ee9eea0
·
docs : update bindings list (#10261)
·
Nov 13, 2024
b4071
0e712a5a
·
server : fix incorrect res in validate_model_chat_template (#10272)
·
Nov 13, 2024
b4069
2e82ffa4
·
sycl : Fixes to broken builds and test-backend-ops (#10257)
·
Nov 13, 2024
b4068
80dd7ff2
·
vulkan: Optimize contiguous copies (#10254)
·
Nov 13, 2024
b4067
54ef9cfc
·
vulkan: Throttle the number of shader compiles during the build step. (#10222)
·
Nov 11, 2024
b4066
b0cefea5
·
metal : more precise Q*K in FA vec kernel (#10247)
·
Nov 11, 2024
b4065
b141e5f6
·
server : enable KV cache defrag by default (#10233)
·
Nov 11, 2024
b4062
160687b3
·
vulkan: Fix newly added tests for permuted mul_mat and 1D im2col (#10226)
·
Nov 10, 2024
b4061
6423c65a
·
metal : reorder write loop in mul mat kernel + style (#10231)
·
Nov 09, 2024
b4060
39a334a9
·
metal : fix build and some more comments (#10229)
·
Nov 09, 2024
b4059
bb38cdd8
·
metal : fix F32 accumulation in FA vec kernel (#10232)
·
Nov 09, 2024
b4058
f018acba
·
llama : fix Qwen model type strings
·
Nov 09, 2024
b4057
46323fa9
·
metal : hide debug messages from normal log
·
Nov 09, 2024
b4056
5b359bb1
·
ggml: fix zero division in ‘dne’ calculation in CUDA COUNT_EQUAL operator when...
·
Nov 09, 2024
b4055
e8921349
·
ggml : optimize llamafile cpu matrix multiplication for ppc64le (#10156)
·
Nov 09, 2024
b4053
ec450d3b
·
metal : opt-in compile flag for BF16 (#10218)
·
Nov 08, 2024
1
…
41
42
43
44
45
46
47
48
49
…
178