]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/shortlog
pkg/ggml/sources/llama.cpp
2023-08-22 KawrakowQuantization imrovements for k_quants (#2707)
2023-08-22 slarenembedding : evaluate prompt in batches (#2713)
2023-08-22 slarenggml-cuda : use graph allocator (#2684)
2023-08-22 Georgi Gerganovggml : sync latest (SAM + SD operators, CUDA alibi...
2023-08-22 slarenllama-bench : minor fixes (#2695)
2023-08-22 Kylinggml : support CUDA's half type for aarch64(#1455)...
2023-08-22 Shouzheng Liumetal : add missing barriers for mul-mat (#2699)
2023-08-22 Jhen-Jie Hongserver : fallback to default if client param is null...
2023-08-22 KerfuffleFix convert-llama-ggmlv3-to-gguf.py vocab conversion...
2023-08-21 Georgi Gerganovpy : remove obsolete script
2023-08-21 Georgi Gerganovgguf : new file format with flexible meta data (beta...
2023-08-21 Shouzheng Liumetal : fix synchronization in new matrix multiplicatio...
2023-08-21 KawrakowHellaSwag: split token evaluation into batches if neede...
2023-08-20 slarenggml : move all type info to ggml_type_traits (#2663)
2023-08-20 KawrakowMore efficient Hellaswag implementation (#2677)
2023-08-18 Georgi Gerganovserver : better default prompt (#2646)
2023-08-18 Jhen-Jie Hongserver : update xxd usage for older versions compatibil...
2023-08-18 AdrianAdd link to clojure bindings to Readme. (#2659)
2023-08-18 Georgi Gerganovreadme : incoming BREAKING CHANGE
2023-08-18 slarenllama : add benchmark example (#2626)
2023-08-18 mdrokzreadme : add link to Rust bindings (#2656)
2023-08-18 Georgi Gerganovperplexity : more meaningful ETA number - 2 decimal...
2023-08-17 Evan JonesFix unicode in grammars (fixes #2501) (#2553)
2023-08-17 staviqserver : support for saving templates in browser LocalS...
2023-08-17 Johannes GäßlerREADME: fix LLAMA_CUDA_MMV_Y documentation (#2647)
2023-08-17 Henri Vasserman[Zig] Fixing Zig build and improvements (#2554)
2023-08-17 KerfuffleAdd --cfg-negative-prompt-file option for examples...
2023-08-17 Georgi Gerganovllama : replace (permute + reshape + view_1d) with...
2023-08-17 drbhtests : adds simple llama grammar tests (#2618)
2023-08-17 Shouzheng Liuggml-alloc : fix discrepency between measure&eval ...
2023-08-16 Kolen Cheungcmake : install ggml-meta.metal if LLAMA_METAL (#2449)
2023-08-16 Jhen-Jie Hongmetal : print error of load pipeline state (#2564)
2023-08-16 Shouzheng Liumetal : enable ggml-alloc (#2627)
2023-08-16 Shouzheng Liumetal : matrix-matrix multiplication kernel (#2615)
2023-08-15 Georgi Gerganovscripts : add helper script to get wikitext
2023-08-14 Jhen-Jie Hongserver : add missing /json-schema-to-grammar.mjs (...
2023-08-14 Jhen-Jie Hongmetal : return null instead of exit(1) (#2573)
2023-08-14 Cheng Shaoserver : add --numa support (#2524)
2023-08-14 Kamil Tomšíkllama : add missing enum keyword in function signatures...
2023-08-14 Johannes GäßlerCUDA: launch_bounds, small q4_K, q5_K mmq refactor...
2023-08-14 Jhen-Jie Hongserver : fix default grammar by use empty string in...
2023-08-14 Jhen-Jie Hongserver : implement json-schema-to-grammar.mjs & add...
2023-08-14 vxiiduuEnhance Windows 7 and below compatibility. (#2592)
2023-08-13 drbhtest : add simple grammar parsing tests (#2594)
2023-08-12 Johannes GäßlerCUDA: Fixed OpenLLaMA 3b mmq, reduced compile time...
2023-08-11 byte-6174Adding support for llama2.c models (#2559)
2023-08-11 Equimserver: fixed wrong variable name in timing json (...
2023-08-10 DannyDaemonicHandle `ENABLE_VIRTUAL_TERMINAL_PROCESSING` more gracef...
2023-08-10 Christian DemsarAdd --n-predict -2 for stopping generation on full...
2023-08-10 Martin KrasserFix grammar-based sampling issue in server (#2566)
2023-08-09 Sam Spilsburyggml-alloc: Don't try to re-use buffers of external...
2023-08-09 grahamethadd log_callback to llama_context_params for custom...
2023-08-09 Johannes GäßlerCUDA: tuned mul_mat_q kernels (#2546)
2023-08-08 Martin KrasserAllow passing grammar to completion endpoint (#2532)
2023-08-08 Johannes GäßlerCUDA: tighter VRAM scratch size for 65b/70b (#2551)
2023-08-08 chaihahahallm.vim : multiline autocompletion, get rid of "^@...
2023-08-08 Georgi Gerganovvim : bring back simple llm.vim example
2023-08-08 AustinMrozvim : streaming and more (#2495)
2023-08-07 klosaxAdd --rope-scale parameter (#2544)
2023-08-07 Georgi Gerganovggml : mul mat tweaks (#2372)
2023-08-07 Georgi Gerganovggml : pad result of ggml_nbytes()
2023-08-07 Georgi Gerganovggml : change params pointer (style change) (#2539)
2023-08-07 Georgi Gerganovggml : sync (custom ops) (#2537)
2023-08-07 Johannes GäßlerFixed mmap prefetch for GPU offloading (#2529)
2023-08-07 Georgi Gerganovmetal : fix out-of-bounds access + inc concurrency...
2023-08-07 GiviMAD[Makefile] Move ARM CFLAGS before compilation (#2536)
2023-08-07 Henri Vasserman[Zig] Rewrite build for Zig 0.11 (#2514)
2023-08-06 DannyDaemonicconsole : fix issue related to Windows 11 PowerShell...
2023-08-06 Keiichi Tabataconvert.py : add missing abstract methods for quantized...
2023-08-05 Johannes GäßlerCUDA: faster k-quant mul_mat_q kernels (#2525)
2023-08-04 Jonas Wunderlichfix firefox autoscroll (#2519)
2023-08-04 Cebtenzzreserver: regenerate completion.js.hpp (#2515)
2023-08-04 CebtenzzreCUDA: use min compute capability of GPUs actually used...
2023-08-04 CebtenzzreCUDA: check if event is NULL before cudaStreamWaitEvent...
2023-08-04 DannyDaemonicAdd --simple-io option for subprocesses and break out...
2023-08-04 Stephen NicholsFixing race condition in server and partial stream...
2023-08-04 l3utterflyStream save llama context data to file instead of alloc...
2023-08-04 Borislav Stanimirovbuild : fix several cast and printf warnings (#2499)
2023-08-03 Evan Jonesexamples : generate JSON according to schema (#1887)
2023-08-02 Johannes GäßlerCUDA: faster non k-quant mul_mat_q kernels (#2483)
2023-08-02 Johannes GäßlerCUDA: Fix models with output size != 32000 (#2480)
2023-08-02 ldwangreadme : add Aquila-7B model series to supported models...
2023-08-02 Evetests : Fix compilation warnings (Linux/GCC) (#2451)
2023-08-02 Yiming Cuireadme : Add Chinese LLaMA-2 / Alpaca-2 to supported...
2023-08-01 Bono Lvfix a typo in examples/server/README.md (#2478)
2023-08-01 ebraminioserver : Support dark mode (#2414)
2023-08-01 Matteo Boschinimetal : add gqa8 kernel to allow llama-2-70B on metal...
2023-07-31 Johannes GäßlerCUDA: fixed LLAMA_FAST compilation option (#2473)
2023-07-31 Johannes GäßlerCUDA: fixed cmake F16 option (#2471)
2023-07-31 Johannes GäßlerCUDA: mmq CLI option, fixed mmq build issues (#2453)
2023-07-31 Johannes GäßlerCUDA: Implemented row flattening for non-glm RoPE ...
2023-07-31 Johannes GäßlerCUDA: fewer memory bank conflicts for mul_mat_q (#2458)
2023-07-31 slarenFix Metal backend broken from the allocator changes...
2023-07-30 slarenggml : add graph tensor allocator (#2411)
2023-07-29 Johannes GäßlerCUDA: Quantized matrix matrix multiplication (#2160)
2023-07-29 Johannes GäßlerCUDA: faster multi GPU synchronization (#2448)
2023-07-28 klosaxperplexity : add Hellaswag calculation (#2389)
2023-07-28 Leeggml : workaround for missing _mm256_setr_m128i in...
2023-07-28 eric8607242llama : support more diverse tokenizers? (#2420)
2023-07-28 Georgi Gerganovexamples : fix whitespace
next