Tags

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

b1299

f5ef5cfb · ggml-cuda : perform cublas mat mul of quantized types as f16 (#3412) · Sep 30, 2023
b1298

40e07a60 · llama.cpp : add documentation about rope_freq_base and scale values (#3401) · Sep 29, 2023
b1297

bc34dd4f · train : fix KQ_pos allocation (#3392) · Sep 29, 2023
b1296

2777a84b · llama : quantize up to 31% faster on Linux and Windows with mmap (#3206) · Sep 29, 2023
b1292

bc39553c · build : enable more non-default compiler warnings (#3200) · Sep 28, 2023
b1291

0ccfc62a · ggml_tensor: update the structure comments. (#3283) · Sep 28, 2023
b1290

7f1a0fe7 · ggml : release the requested thread pool resource (#3292) · Sep 28, 2023
b1289

16bc66d9 · llama.cpp : split llama_context_params into model and context params (#3301) · Sep 28, 2023
b1288

0512d666 · ci : multithreaded builds (#3311) · Sep 28, 2023
b1287

0e76a899 · train : finetune LORA (#2632) · Sep 28, 2023
b1286

2db94d98 · gguf : basic type checking in gguf_get_* (#3346) · Sep 28, 2023
b1285

ecf90b1a · gguf : make token scores and types optional (#3347) · Sep 28, 2023
b1284

2619109a · ci : disable freeBSD builds due to lack of VMs (#3381) · Sep 28, 2023
b1283

ec893798 · llama : custom attention mask + parallel decoding + no context swaps (#3228) · Sep 28, 2023
b1280

da040034 · ggml-cuda : perform cublas fp16 matrix multiplication as fp16 (#3370) · Sep 28, 2023
b1277

20c7e1e8 · gguf : fix a few general keys (#3341) · Sep 27, 2023
b1276

dc689740 · metal : reusing llama.cpp logging (#3152) · Sep 27, 2023
b1275

527e57cf · build : add ACCELERATE_NEW_LAPACK to fix warning on macOS Sonoma (#3342) · Sep 27, 2023
master-99c5c9a

99c5c9a0 · Upload immediately to device. · Sep 26, 2023
b1273

99115f3f · cmake : fix build-info.h on MSVC (#3309) · Sep 25, 2023

Prev
1
…
54
55
56
57
58
59
60
61
62
…
99
Next