Tags

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

b1690

ba661751 · sync : ggml (fix im2col) (#4591) · Dec 22, 2023
b1689

a5587695 · cuda : fix jetson compile error (#4560) · Dec 22, 2023
b1688

6724ef16 · Fix CudaMemcpy direction (#4599) · Dec 22, 2023
b1687

48b7ff19 · llama : fix platforms without mmap (#4578) · Dec 22, 2023
b1686

48b24b17 · ggml : add comment about backward GGML_OP_DIAG_MASK_INF (#4203) · Dec 22, 2023
b1685

28cb35a0 · make : add LLAMA_HIP_UMA option (#4587) · Dec 22, 2023
b1684

f31b9848 · ci : tag docker image with build number (#4584) · Dec 22, 2023
b1682

0137ef88 · ggml : extend `enum ggml_log_level` with `GGML_LOG_LEVEL_DEBUG` (#4579) · Dec 22, 2023
b1681

c7e9701f · llama : add ability to cancel model loading (#4462) · Dec 22, 2023
b1680

afefa319 · ggml : change ggml_scale to take a float instead of tensor (#4573) · Dec 21, 2023
b1678

32259b2d · gguf : simplify example dependencies · Dec 21, 2023
b1677

4a5f9d62 · ci : add `jlumbroso/free-disk-space` to docker workflow (#4150) · Dec 21, 2023
b1676

d232aca5 · llama : initial ggml-backend integration (#4520) · Dec 21, 2023
b1675

31f27758 · llama : allow getting n_batch from llama_context in c api (#4540) · Dec 21, 2023
b1673

0f630fbc · cuda : ROCm AMD Unified Memory Architecture (UMA) handling (#4449) · Dec 21, 2023
b1672

562cf222 · ggml-cuda: Fix HIP build by adding define for __trap (#4569) · Dec 21, 2023
b1671

8fe03ffd · common : remove incorrect --model-draft default (#4568) · Dec 21, 2023
b1670

91544948 · CUDA: mul_mat_id always on GPU for batches >= 32 (#4553) · Dec 21, 2023
b1667

66f35a2f · cuda : better error message for ggml_get_rows (#4561) · Dec 21, 2023
b1666

13988239 · cuda : replace asserts in wrong architecture checks with __trap (#4556) · Dec 21, 2023

Prev
1
…
39
40
41
42
43
44
45
46
47
…
99
Next