Tags

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

b1477

629f917c · cuda : add ROCM aliases for CUDA pool stuff (#3918) · Nov 02, 2023
b1476

51b2fc11 · cmake : fix relative path to git submodule index (#3915) · Nov 02, 2023
b1474

c7743fe1 · cuda : fix const ptrs warning causing ROCm build issues (#3913) · Nov 02, 2023
b1473

d6069051 · cuda : use CUDA memory pool with async memory allocation/deallocation when available (#3903) · Nov 02, 2023
b1472

4ff1046d · gguf : print error for GGUFv1 files (#3908) · Nov 02, 2023
b1471

21958bb3 · cmake : disable LLAMA_NATIVE by default (#3906) · Nov 02, 2023
b1470

2756c4fb · gguf : remove special-case code for GGUFv1 (#3901) · Nov 02, 2023
b1469

1efae9b7 · llm : prevent from 1-D tensors being GPU split (#3697) · Nov 02, 2023
b1468

b12fa0d1 · build : link against build info instead of compiling against it (#3879) · Nov 02, 2023
b1467

4d719a6d · cuda : check if this fixes Pascal card regression (#3882) · Nov 02, 2023
b1466

183b3fac · metal : fix build errors and kernel sig after #2268 (#3898) · Nov 02, 2023
b1465

2fffa0d6 · cuda : fix RoPE after #2268 (#3897) · Nov 02, 2023
b1464

0eb332a1 · llama : fix llama_context_default_params after #2268 (#3893) · Nov 01, 2023
b1463

d02e98cd · ggml-cuda : compute ptrs for cublasGemmBatchedEx in a kernel (#3891) · Nov 01, 2023
b1462

898aeca9 · llama : implement YaRN RoPE scaling (#2268) · Nov 01, 2023
b1461

c43c2da8 · llm : fix llm_build_kqv taking unused tensor (benign, #3837) · Nov 01, 2023
b1460

523e49b1 · llm : fix falcon norm after refactoring (#3837) · Nov 01, 2023
b1459

e16b9fa4 · metal : multi-simd softmax (#3710) · Nov 01, 2023
b1458

ff8f9a88 · common : minor (#3715) · Nov 01, 2023
b1457

50337961 · llm : add llm_build_context (#3881) · Nov 01, 2023

Prev
1
…
47
48
49
50
51
52
53
54
55
…
99
Next