]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/shortlog
pkg/ggml/sources/llama.cpp
2023-10-05 shibe2CLBlast: Fix handling of on-device tensor data
2023-10-05 Jhen-Jie Hongserver : fix incorrect num_tokens_predicted (#3480)
2023-10-05 Jhen-Jie Hongswift : disable ACCELERATE_NEW_LAPACK (#3481)
2023-10-05 Jhen-Jie Hongci : add swift build via xcodebuild (#3482)
2023-10-04 Kerfuffleconvert : fix Baichuan2 models by using vocab size...
2023-10-04 Georgi Gerganovreadme : add project status link
2023-10-04 Georgi Gerganovggml : fix build after #3329
2023-10-04 ds5t5llm : add Refact model (#3329)
2023-10-04 Georgi Gerganovsync : ggml (conv 1d + 2d updates, UB fixes) (#3468)
2023-10-04 Merrick Christensenfinetune : readme fix typo (#3465)
2023-10-03 Tameemggml : add RISC-V Vector Support for K-Quants and impro...
2023-10-03 h-h-h-hmain : consistent prefix/suffix coloring (#3425)
2023-10-03 Georgi Gerganovllama : fix session saving/loading (#3400)
2023-10-03 Alex Klinkhamerllama : expose model's rope_freq_scale in the API ...
2023-10-03 Jiahao Limetal : alibi for arbitrary number of heads (#3426)
2023-10-03 Evecmake : make LLAMA_NATIVE flag actually use the instruc...
2023-10-03 goerchWork on the BPE tokenizer (#3252)
2023-10-02 cebtenzzreconvert : fix vocab size when not defined in hparams...
2023-10-02 cebtenzzrecmake : increase minimum version for add_link_options...
2023-10-02 shibe2CLBlast: Add broadcast support for matrix multiplicatio...
2023-10-02 cebtenzzregguf : add BERT, MPT, and GPT-J arch info (#3408)
2023-10-02 cebtenzzregguf : general usability improvements (#3409) gguf-v0.4.0
2023-10-02 cebtenzzrecmake : make CUDA flags more similar to the Makefile...
2023-10-02 xaedesfinetune : fix #3404 (#3437)
2023-10-02 Adrianmetal : set log callback before initializing (#3427)
2023-10-02 bandoticmake : fix transient definitions in find pkg (#3411)
2023-10-02 Kevin Jidocker : ignore Git files (#3314)
2023-10-02 vvhg1infill : add new example + extend server API (#3296)
2023-09-30 slarenggml-cuda : perform cublas mat mul of quantized types...
2023-09-29 slarenllama.cpp : add documentation about rope_freq_base...
2023-09-29 Georgi Gerganovtrain : fix KQ_pos allocation (#3392)
2023-09-29 Cebtenzzrellama : quantize up to 31% faster on Linux and Windows...
2023-09-29 BarfingLemursreadme : update hot topics + model links (#3399)
2023-09-29 Andrew Duffyreadme : add link to grammars app (#3388)
2023-09-29 Jhen-Jie Hongswift : fix build on xcode 15 (#3387)
2023-09-28 Cebtenzzrebuild : enable more non-default compiler warnings ...
2023-09-28 Hua Jiangggml_tensor: update the structure comments. (#3283)
2023-09-28 Qu Zongfuggml : release the requested thread pool resource ...
2023-09-28 slarenllama.cpp : split llama_context_params into model and...
2023-09-28 Eveci : multithreaded builds (#3311)
2023-09-28 xaedestrain : finetune LORA (#2632)
2023-09-28 Cebtenzzregguf : basic type checking in gguf_get_* (#3346)
2023-09-28 Cebtenzzregguf : make token scores and types optional (#3347)
2023-09-28 Georgi Gerganovci : disable freeBSD builds due to lack of VMs (#3381)
2023-09-28 Georgi Gerganovllama : custom attention mask + parallel decoding ...
2023-09-28 Kevin Jidocs : mark code as Bash (#3375)
2023-09-28 Pierre Alexandre... readme : add Mistral AI release 0.1 (#3362)
2023-09-28 slarenggml-cuda : perform cublas fp16 matrix multiplication...
2023-09-27 Zhang Peiyuanconvert : remove bug in convert.py permute function...
2023-09-27 Richard Robersonmake-ggml.py : compatibility with more models and GGUF...
2023-09-27 Cebtenzzregguf : fix a few general keys (#3341)
2023-09-27 Rickard Hallerbäckmetal : reusing llama.cpp logging (#3152)
2023-09-27 Jag Chadhabuild : add ACCELERATE_NEW_LAPACK to fix warning on...
2023-09-27 BarfingLemursreadme : add some recent perplexity and bpw measurement...
2023-09-25 DAN™cmake : fix build-info.h on MSVC (#3309)
2023-09-25 2f38b454docs: Fix typo CLBlast_DIR var. (#3330)
2023-09-25 Erik Scholznix : add cuda, use a symlinked toolkit for cmake ...
2023-09-23 slarenllama-bench : add README (#3317)
2023-09-23 Cebtenzzreexamples : fix RoPE defaults to match PR #3240 (#3315)
2023-09-23 Kevin Jiscripts : use `/usr/bin/env` in shebang (#3313)
2023-09-21 Lee DrakeUpdate README.md (#3289)
2023-09-21 shibe2ggml-opencl.cpp: Make private functions static (#3300)
2023-09-21 Edward Taylorzig : fix for updated c lib (#3259)
2023-09-21 yuisekiembedding : update README.md (#3224)
2023-09-21 Johannes GäßlerCUDA: use only 1 thread if fully offloaded (#2915)
2023-09-20 Georgi Gerganovreadme : update hot topics
2023-09-20 Cebtenzzrellama : allow gguf RoPE keys to be overridden with...
2023-09-20 Cebtenzzrebenchmark-matmult : do not use integer abs() on a float...
2023-09-20 kangflake : Restore default package's buildInputs (#3262)
2023-09-20 AlonCI: FreeBSD fix (#3258)
2023-09-20 Georgi Gerganovexamples : fix benchmark-matmult (#1554)
2023-09-18 Cebtenzzremake : restore build-info.h dependency for several...
2023-09-18 Erik Scholzci : switch cudatoolkit install on windows to networked...
2023-09-17 Johannes GäßlerCUDA: fix peer access logic (#3231)
2023-09-17 Johannes GäßlerCUDA: enable peer access between devices (#2470)
2023-09-17 slarenllama.cpp : show model size and BPW on load (#3223)
2023-09-17 Johannes GäßlerCUDA: fix scratch malloced on non-main device (#3220)
2023-09-16 IsaacDynamoEnable BUILD_SHARED_LIBS=ON on all Windows builds ...
2023-09-16 VladEnable build with CUDA 11.0 (make) (#3132)
2023-09-16 goerchFixing the last deviations from sentencepiece indicated...
2023-09-15 Cebtenzzreexamples : add compiler version and target to build...
2023-09-15 Cebtenzzrecheck C++ code with -Wmissing-declarations (#3184)
2023-09-15 Cebtenzzrefix build numbers by setting fetch-depth=0 (#3197)
2023-09-15 Meng Zhangllama : add support for StarCoder model architectures...
2023-09-15 Cebtenzzrecommon : do not use GNU zero-length __VA_ARGS__ extensi...
2023-09-15 Georgi Gerganovmetal : fix bug in soft_max kernels (out-of-bounds...
2023-09-15 Cebtenzzreconvert : make ftype optional in simple scripts (#3185)
2023-09-15 Georgi Gerganovsync : ggml (Metal F32 support + reduce ggml-alloc...
2023-09-15 Engininja2cmake : fix building shared libs for clang (rocm) on...
2023-09-15 Evgeny Kurnevskyflake : use pkg-config instead of pkgconfig (#3188)
2023-09-15 Georgi Gerganovmetal : relax conditions on fast matrix multiplication...
2023-09-15 Andreicmake : fix llama.h location when built outside of...
2023-09-15 Ali Tariqci : Cloud-V for RISC-V builds (#3160)
2023-09-15 Rolandllama : remove mtest (#3177)
2023-09-15 Cebtenzzrellama : make quantize example up to 2.7x faster (#3115)
2023-09-14 jneemflake : allow $out/include to already exist (#3175)
2023-09-14 Andreicmake : compile ggml-rocm with -fpic when building...
2023-09-14 Asbjørn Ollingflake : include llama.h in nix output (#3159)
2023-09-14 Cebtenzzremake : fix clang++ detection, move some definitions...
2023-09-14 AlonCI: add FreeBSD & simplify CUDA windows (#3053)
next