Tags

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

b1643

88ae8952 · server : add optional API Key Authentication example (#4441) · Dec 15, 2023
b1642

ee4725a6 · ggml : group mul_mat_id rows by matrix (cpu only) (#4480) · Dec 15, 2023
b1641

6744dbe9 · ggml : use ggml_row_size where possible (#4472) · Dec 14, 2023
b1640

cafcd4f8 · ggml : remove n_dims from ggml_tensor (#4469) · Dec 14, 2023
b1638

20a68a70 · ggml : add ggml_row_size() (fixes llama out of space) (#4461) · Dec 14, 2023
b1637

55e87c37 · ggml : fix OpenCL broadcast requirement for ggml_mul (close #4453) · Dec 14, 2023
b1634

948ff137 · server : fix handling of characters that span multiple tokens when streaming (#4446) · Dec 13, 2023
b1633

4d98d9a6 · sync : ggml (SD ops, tests, kernels) (#4444) · Dec 13, 2023
b1632

70f806b8 · build : detect host compiler and cuda compiler separately (#4414) · Dec 13, 2023
b1631

9fb13f95 · common : add `--version` option to show build info in CLI (#4433) · Dec 13, 2023
b1629

799a1cb1 · llama : add Mixtral support (#4406) · Dec 13, 2023
b1627

9494d7c4 · english : use `typos` to fix comments and logs (#4354) · Dec 12, 2023
b1626

6138963f · build : target Windows 8 for standard mingw-w64 (#4405) · Dec 12, 2023
b1625

6391817c · llama : document logits_all deprecation (#4418) · Dec 12, 2023
b1624

d9d4cfef · server : fix local model name in server (#4420) · Dec 12, 2023
b1623

41a11aaf · ggml : increased GGML_MAX_PARAMS to allow finetuning of 70b models (#4424) · Dec 12, 2023
b1621

e18f7345 · grammar : revert the replacement of llama_token_to_piece with id_to_token (#4396) · Dec 09, 2023
b1620

fe680e3d · sync : ggml (new ops, tests, backend, etc.) (#4359) · Dec 07, 2023
b1619

bcc0eb45 · llama : per-layer KV cache + quantum K cache (#4309) · Dec 07, 2023
b1618

81bc9214 · train : fix #4227 (double free in... · Dec 07, 2023

Prev
1
…
41
42
43
44
45
46
47
48
49
…
99
Next