| 2023-05-01 |
slaren | cuBLAS: fall back to pageable memory if pinned alloc... |
commit | commitdiff | tree |
| 2023-05-01 |
Alex Klinkhamer | llama : let context be const when accessing const data... |
commit | commitdiff | tree |
| 2023-04-30 |
Georgi Gerganov | ggml : fix UB (int << 31) |
commit | commitdiff | tree |
| 2023-04-30 |
Pavol Rusnak | build: add armv{6,7,8} support to cmake (#1251) |
commit | commitdiff | tree |
| 2023-04-30 |
jon-chuang | common : better default number of threads (#934) |
commit | commitdiff | tree |
| 2023-04-30 |
0cc4m | ggml : add CLBlast q5_0, q5_1, q8_0 dequant kernels... |
commit | commitdiff | tree |
| 2023-04-30 |
Georgi Gerganov | ggml : add Q5 WASM SIMD + GGML_FTYPE |
commit | commitdiff | tree |
| 2023-04-30 |
Stephan Walter | Various fixes to mat_mul benchmark (#1253) |
commit | commitdiff | tree |
| 2023-04-30 |
Georgi Gerganov | ggml : fix labels for GGML_OP_ALIBI |
commit | commitdiff | tree |
| 2023-04-29 |
Georgi Gerganov | ggml : fix 32-bit ARM NEON |
commit | commitdiff | tree |
| 2023-04-29 |
Georgi Gerganov | ggml : use vzip instead of vuzp for consistency |
commit | commitdiff | tree |
| 2023-04-29 |
Georgi Gerganov | ggml : fix visibility and unused warnings |
commit | commitdiff | tree |
| 2023-04-29 |
Georgi Gerganov | ggml : fix #if for f32_f32 mul_mat (CLBlast) (#1229) |
commit | commitdiff | tree |
| 2023-04-29 |
Georgi Gerganov | ggml : adjust mul_mat_f16 work memory (#1226) |
commit | commitdiff | tree |
| 2023-04-29 |
Georgi Gerganov | build : fix reference to old llama_util.h |
commit | commitdiff | tree |
| 2023-04-29 |
Georgi Gerganov | examples : fix save-load-state + rename llama-util.h |
commit | commitdiff | tree |
| 2023-04-29 |
Georgi Gerganov | common : change default parameters to pre-#1126 (#1223) |
commit | commitdiff | tree |
| 2023-04-29 |
Ivan Stepanov | llama : new sampling algorithms (#1126) |
commit | commitdiff | tree |
| 2023-04-29 |
slaren | cuBLAS: use host pinned memory and dequantize while... |
commit | commitdiff | tree |
| 2023-04-28 |
Henri Vasserman | cuBLAS: non-contiguous tensor support (#1215) |
commit | commitdiff | tree |
| 2023-04-28 |
Stephan Walter | Remove Q4_3 which is no better than Q5 (#1218) |
commit | commitdiff | tree |
| 2023-04-28 |
Georgi Gerganov | readme : update hot topics |
commit | commitdiff | tree |
| 2023-04-28 |
Georgi Gerganov | ggml : sync ggml (ggml_alibi) |
commit | commitdiff | tree |
| 2023-04-28 |
CRD716 | examples : add Jeopardy example (#1168) |
commit | commitdiff | tree |
| 2023-04-28 |
Evan Jones | llama : add session file format and saved sessions... |
commit | commitdiff | tree |
| 2023-04-28 |
Georgi Gerganov | ggml : add helper debug printf in soft_max |
commit | commitdiff | tree |
| 2023-04-28 |
0cc4m | ggml : add CLBlast support (#1164) |
commit | commitdiff | tree |
| 2023-04-28 |
Folko-Ven | Correcting link to w64devkit (#1214) |
commit | commitdiff | tree |
| 2023-04-28 |
Johannes Gäßler | Add Manjaro CUDA include and lib dirs to Makefile ... |
commit | commitdiff | tree |
| 2023-04-28 |
Yann Follet | add avx2 for dot_q8_0_q8_0, 2x faster than scalar ... |
commit | commitdiff | tree |
| 2023-04-26 |
Stephan Walter | ggml : slightly faster AVX2 implementation for Q5 ... |
commit | commitdiff | tree |
| 2023-04-26 |
Georgi Gerganov | readme : add quantization info |
commit | commitdiff | tree |
| 2023-04-26 |
Georgi Gerganov | ggml : add Q5_0 and Q5_1 quantization (#1187) |
commit | commitdiff | tree |
| 2023-04-26 |
Ásgeir Bjarni... | Allow setting the rng seed after initialization. (... |
commit | commitdiff | tree |
| 2023-04-26 |
DaniAndTheWeb | Updating build instructions to include BLAS support... |
commit | commitdiff | tree |
| 2023-04-26 |
Pavol Rusnak | quantize : use `map` to assign quantization type from... |
commit | commitdiff | tree |
| 2023-04-25 |
Stephan Walter | Update SHA256SUMS after quantization change (#1181) |
commit | commitdiff | tree |
| 2023-04-25 |
ostix360 | py : cast lora_alpha to int in convert-lora-to-ggml... |
commit | commitdiff | tree |
| 2023-04-25 |
Pavol Rusnak | nix: use convert.py instead of legacy wrapper convert... |
commit | commitdiff | tree |
| 2023-04-25 |
Georgi Gerganov | ggml : add Q8_0 quantization format (rename the old... |
commit | commitdiff | tree |
| 2023-04-25 |
unbounded | ggml : use full range for Q4_0 and Q4_2 quantization... |
commit | commitdiff | tree |
| 2023-04-24 |
xaedes | ggml : fix bug in ggml_compute_forward_sum_f32 (#1162) |
commit | commitdiff | tree |
| 2023-04-24 |
Georgi Gerganov | ggml : export symbols (#1155) |
commit | commitdiff | tree |
| 2023-04-24 |
xaedes | examples : add save_load_state example (#1150) |
commit | commitdiff | tree |
| 2023-04-24 |
Georgi Gerganov | llama : increase scratch buffer size for 65B (ref ... |
commit | commitdiff | tree |
| 2023-04-24 |
mgroeber9110 | examples/main README improvements and some light refact... |
commit | commitdiff | tree |
| 2023-04-24 |
Stephan Walter | Fix build for gcc 8 and test in CI (#1154) |
commit | commitdiff | tree |
| 2023-04-24 |
slaren | Fix cuda compilation (#1128) |
commit | commitdiff | tree |
| 2023-04-24 |
Georgi Gerganov | llama : refactor get / set state + remove redundant... |
commit | commitdiff | tree |
| 2023-04-23 |
slaren | Fix LoRA acronym (#1145) |
commit | commitdiff | tree |
| 2023-04-23 |
Georgi Gerganov | scripts : add helper scripts to synch ggml repo |
commit | commitdiff | tree |
| 2023-04-23 |
DannyDaemonic | Added README.md for main with examples and explanations... |
commit | commitdiff | tree |
| 2023-04-23 |
Georgi Gerganov | ggml : do not print perf ops that have not been used... |
commit | commitdiff | tree |
| 2023-04-23 |
Georgi Gerganov | ggml : better PERF prints + support "LLAMA_PERF=1 make" |
commit | commitdiff | tree |
| 2023-04-23 |
Stephan Walter | Improve AVX2 for vec_dot_q4_3_q8_0 (#1138) |
commit | commitdiff | tree |
| 2023-04-23 |
Pavol Rusnak | readme : update gpt4all instructions (#980) |
commit | commitdiff | tree |
| 2023-04-23 |
Yishuo Wang | A better `packNibbles` and `mul_sum_i8_pairs_float... |
commit | commitdiff | tree |
| 2023-04-22 |
Georgi Gerganov | ggml : fix Q4_3 cuBLAS |
commit | commitdiff | tree |
| 2023-04-22 |
Stephan Walter | ci : trigger CI for drafts, but not most PR actions... |
commit | commitdiff | tree |
| 2023-04-22 |
Stephan Walter | Fix CI: ARM NEON, quantization unit tests, editorconfig... |
commit | commitdiff | tree |
| 2023-04-22 |
unbounded | ggml : unit test for quantization functions (#953) |
commit | commitdiff | tree |
| 2023-04-22 |
wbpxre150 | llama : print timings on ctrl+c exit (#1021) |
commit | commitdiff | tree |
| 2023-04-22 |
eiery | llama : have n_batch default to 512 (#1091) |
commit | commitdiff | tree |
| 2023-04-22 |
Howard Su | cmake : fix build under Windows when enable BUILD_SHARE... |
commit | commitdiff | tree |
| 2023-04-22 |
Georgi Gerganov | ggml : fix AVX build + update to new Q8_0 format |
commit | commitdiff | tree |
| 2023-04-22 |
Georgi Gerganov | ggml : alternative Q4_3 implementation using modified... |
commit | commitdiff | tree |
| 2023-04-22 |
Stephan Walter | ggml : AVX2 optimization for vec_dot_q4_3_q8_0 and... |
commit | commitdiff | tree |
| 2023-04-22 |
Clint Herron | examples : Improve Alpaca Default Repeat Penalty: Bette... |
commit | commitdiff | tree |
| 2023-04-22 |
xaedes | llama : add api for getting/setting the complete state... |
commit | commitdiff | tree |
| 2023-04-21 |
slaren | Improve cuBLAS performance by using a memory pool ... |
commit | commitdiff | tree |
| 2023-04-21 |
apaz | llama : fixed rlimit error message (#888) |
commit | commitdiff | tree |
| 2023-04-21 |
源文雨 | cmake : link threads publicly to ggml (#1042) |
commit | commitdiff | tree |
| 2023-04-21 |
Alex Klinkhamer | main : evaluate tokens in batches after swapping contex... |
commit | commitdiff | tree |
| 2023-04-21 |
xaedes | llama : remember and restore kv cache data pointers... |
commit | commitdiff | tree |
| 2023-04-21 |
Kawrakow | ggml : a faster version for Q4_1 x Q8_0 dot products... |
commit | commitdiff | tree |
| 2023-04-21 |
slaren | Show perplexity ETA in hours and minutes (#1096) |
commit | commitdiff | tree |
| 2023-04-21 |
Georgi Gerganov | llama : fix comment for "output.weight" tensor |
commit | commitdiff | tree |
| 2023-04-20 |
Stephan Walter | Add ggml-model-*.bin checksums for 7B, 13B, 30B, 65B... |
commit | commitdiff | tree |
| 2023-04-20 |
Georgi Gerganov | ggml : sync ggml (add GPT-NeoX RoPE implementation) |
commit | commitdiff | tree |
| 2023-04-20 |
Georgi Gerganov | ggml : fix bug in ggml_compute_forward_dup_f32() |
commit | commitdiff | tree |
| 2023-04-20 |
slaren | Add Q4_3 support to cuBLAS (#1086) |
commit | commitdiff | tree |
| 2023-04-20 |
Georgi Gerganov | ggml : do not break cuBLAS build (Q4_3 is not yet imple... |
commit | commitdiff | tree |
| 2023-04-20 |
Georgi Gerganov | ggml : fix Q4_3 quantization |
commit | commitdiff | tree |
| 2023-04-20 |
Kawrakow | llama : multi-threaded quantization (#1075) |
commit | commitdiff | tree |
| 2023-04-20 |
Georgi Gerganov | ggml : add Q4_3 quantization (#1082) |
commit | commitdiff | tree |
| 2023-04-20 |
Ivan Komarov | ci : remove the LLAMA_ACCELERATE matrix dimension from... |
commit | commitdiff | tree |
| 2023-04-20 |
源文雨 | fix: LLAMA_CUBLAS=1 undefined reference 'shm_open'... |
commit | commitdiff | tree |
| 2023-04-20 |
Stephan Walter | AVX2 optimization for vec_dot_q4_2_q8_0 (#1068) |
commit | commitdiff | tree |
| 2023-04-20 |
slaren | Improve cuBLAS performance by dequantizing on the GPU... |
commit | commitdiff | tree |
| 2023-04-19 |
CRD716 | Minor: Readme fixed grammar, spelling, and misc updates... |
commit | commitdiff | tree |
| 2023-04-19 |
Kawrakow | Q4_2 quantization with rmse-optimized scale and quants... |
commit | commitdiff | tree |
| 2023-04-19 |
Georgi Gerganov | ggml : use 8-bit precision for Q4_1 intermediate result... |
commit | commitdiff | tree |
| 2023-04-19 |
Georgi Gerganov | readme : add warning about Q4_2 and Q4_3 |
commit | commitdiff | tree |
| 2023-04-19 |
Stephan Walter | ggml : Q4 cleanup - remove 4-bit dot product code ... |
commit | commitdiff | tree |
| 2023-04-19 |
slaren | Add NVIDIA cuBLAS support (#1044) |
commit | commitdiff | tree |
| 2023-04-18 |
slaren | Multi-threaded ggml_cpy (#1035) |
commit | commitdiff | tree |
| 2023-04-18 |
Georgi Gerganov | ggml : add new Q4_2 quantization (ARM only) (#1046) |
commit | commitdiff | tree |
| 2023-04-18 |
Georgi Gerganov | ggml : scratch that - vmlaq_n_f32 is always better |
commit | commitdiff | tree |
| 2023-04-18 |
Georgi Gerganov | gitignore : vdot |
commit | commitdiff | tree |
| 2023-04-18 |
Georgi Gerganov | ggml : optimize ggml_vec_dot_q4_0_q8_0() using vectoriz... |
commit | commitdiff | tree |
| next |
overview / pkg / ggml / sources / llama.cpp / shortlog