| 2024-05-11 |
0cc4m | Vulkan Bugfixes and Improvements (llama/7084) |
commit | commitdiff | tree |
| 2024-05-11 |
Johannes Gäßler | CUDA: generalize FP16 fattn vec kernel (llama/7061) |
commit | commitdiff | tree |
| 2024-05-11 |
Albert Jin | opencl : alignment size converted from bits to bytes... |
commit | commitdiff | tree |
| 2024-05-11 |
agray3 | Introduction of CUDA Graphs to LLama.cpp (llama/6766) |
commit | commitdiff | tree |
| 2024-05-11 |
Gilad S | metal : use `vm_allocate` instead of `posix_memalign... |
commit | commitdiff | tree |
| 2024-05-11 |
Justine Tunney | ggml : introduce bfloat16 support (llama/6412) |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | metal : fix unused warning |
commit | commitdiff | tree |
| 2024-05-11 |
William Tambellini | Add an option to build without CUDA VMM (llama/7067) |
commit | commitdiff | tree |
| 2024-05-11 |
Xuan Son Nguyen | gguf-split: add --no-tensor-first-split (llama/7072) |
commit | commitdiff | tree |
| 2024-05-11 |
Johannes Gäßler | CUDA: CUDART < 11.7 workaround for __hmax, __hmax2... |
commit | commitdiff | tree |
| 2024-05-11 |
Kevin Gibbons | switch to using localizedDescription (llama/7010) |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | metal : remove deprecated error code (llama/7008) |
commit | commitdiff | tree |
| 2024-05-11 |
Kevin Gibbons | metal : log more info on error (llama/6987) |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | ggml : add Flash Attention (llama/5021) |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | ggml : fix __MSC_VER -> _MSC_VER (llama/6977) |
commit | commitdiff | tree |
| 2024-05-11 |
DAN™ | Fix more int overflow during quant (PPL/CUDA). (llama... |
commit | commitdiff | tree |
| 2024-05-11 |
Xuan Son Nguyen | gguf : enforce that tensor names are unique (llama... |
commit | commitdiff | tree |
| 2024-05-11 |
Neo Zhang | add device version in device list (llama/6959) |
commit | commitdiff | tree |
| 2024-05-11 |
agray3 | Reset schedule earlier to allow overlap with ggml graph... |
commit | commitdiff | tree |
| 2024-05-11 |
slaren | add basic tensor data validation function (llama/6884) |
commit | commitdiff | tree |
| 2024-05-11 |
slaren | gguf : fix mismatch between alloc and free functions... |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | Merge pull request from GHSA-p5mv-gjc5-mwqv |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | ggml : fix redefinition of vaddvq_f32 for 32-bit ARM... |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | ggml : fix MIN / MAX macros (llama/6904) |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | ggml : move 32-bit arm compat in ggml-impl.h (llama... |
commit | commitdiff | tree |
| 2024-05-11 |
Justine Tunney | llamafile : improve sgemm.cpp (llama/6796) |
commit | commitdiff | tree |
| 2024-05-11 |
Dave Airlie | ggml : fix calloc argument ordering. (llama/6820) |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | ggml : fix ggml_backend_cpu_supports_op() for CPY ... |
commit | commitdiff | tree |
| 2024-05-11 |
slaren | ggml : group all experts in a single ggml_mul_mat_id... |
commit | commitdiff | tree |
| 2024-05-11 |
Georgi Gerganov | ggml : fix llamafile sgemm wdata offsets (llama/6710) |
commit | commitdiff | tree |
| 2024-05-11 |
Justine Tunney | ggml : add llamafile sgemm (llama/6414) |
commit | commitdiff | tree |
| 2024-05-11 |
Shijie | llama : add qwen2moe (llama/6074) |
commit | commitdiff | tree |
| 2024-05-11 |
Neo Zhang Jianyu | fix mul_mat_id() for new input, make the ut pass (llama... |
commit | commitdiff | tree |
| 2024-05-11 |
Dave | Added support for GGML_OP_CLAMP in Metal (llama/6662) |
commit | commitdiff | tree |
| 2024-05-11 |
Neo Zhang Jianyu | fix memcpy() crash, add missed cmd in guide, fix softma... |
commit | commitdiff | tree |
| 2024-05-11 |
Johannes Gäßler | CUDA: fix matrix multiplication logic for tests (llama... |
commit | commitdiff | tree |
| 2024-05-11 |
slaren | metal : unify mul_mv_id kernels (llama/6556) |
commit | commitdiff | tree |
| 2024-05-11 |
jiez | llama : add gguf_remove_key + remove split meta during... |
commit | commitdiff | tree |
| 2024-05-01 |
Justina Cho | feat: implemented sigmoid function (#806) |
commit | commitdiff | tree |
| 2024-04-25 |
Borislav Stanimirov | build: fix and ignore msvc warnings (#805) |
commit | commitdiff | tree |
| 2024-04-18 |
Daniel Bevenius | spec : fix typo in gguf.md (#798) |
commit | commitdiff | tree |
| 2024-04-09 |
Georgi Gerganov | sync : whisper.cpp |
commit | commitdiff | tree |
| 2024-04-09 |
Brad Murray | whisper : fix DTW memory access (whisper/2012) |
commit | commitdiff | tree |
| 2024-04-09 |
ulatekh | common : fix file-handle leak in read_wav() (whisper... |
commit | commitdiff | tree |
| 2024-04-09 |
Rotem Dan | main : set stdin to binary mode on Windows (whisper... |
commit | commitdiff | tree |
| 2024-04-09 |
ulatekh | main : allow a response-file as the sole parameter... |
commit | commitdiff | tree |
| 2024-04-09 |
ulatekh | whisper : suppress tokens with a regex (whisper/1997) |
commit | commitdiff | tree |
| 2024-04-09 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| 2024-04-09 |
Georgi Gerganov | license : update copyright notice + add AUTHORS |
commit | commitdiff | tree |
| 2024-04-09 |
Carolinabanana | llama : add Command R Plus support (llama/6491) |
commit | commitdiff | tree |
| 2024-04-09 |
Abhilash Majumder | remove row=1 cond (llama/6532) |
commit | commitdiff | tree |
| 2024-04-09 |
Neo Zhang Jianyu | support/fix OPs GGML_TYPE_IQ4_NL, GGML_TYPE_IQ4_XS... |
commit | commitdiff | tree |
| 2024-04-09 |
Georgi Gerganov | scripts : sync LICENSE and gen-authors.sh |
commit | commitdiff | tree |
| 2024-04-07 |
Georgi Gerganov | sync : whisper.cpp (grammar-parser, skip) |
commit | commitdiff | tree |
| 2024-04-07 |
Georgi Gerganov | sync : whisper.cpp |
commit | commitdiff | tree |
| 2024-04-07 |
Slava Primenko | ggml: bypass code incompatible with CUDA < 11.1 (whispe... |
commit | commitdiff | tree |
| 2024-04-07 |
ulatekh | main : add command-style grammar (whisper/1998) |
commit | commitdiff | tree |
| 2024-04-07 |
Georgi Gerganov | scripts : sync ggml-cuda folder |
commit | commitdiff | tree |
| 2024-04-06 |
Georgi Gerganov | ci : temporary build embeded metal library |
commit | commitdiff | tree |
| 2024-04-06 |
Georgi Gerganov | sync : llama.cpp (skip) |
commit | commitdiff | tree |
| 2024-04-06 |
Ouadie EL FAROUKI | Fixed minor bug when enabling FP16 for non intel target... |
commit | commitdiff | tree |
| 2024-04-06 |
slaren | ggml : mul_mat_id use the same tensor for all the exper... |
commit | commitdiff | tree |
| 2024-04-06 |
Meng, Hengyu | Disable iqx on windows as WA (llama/6435) |
commit | commitdiff | tree |
| 2024-04-06 |
0cc4m | Vulkan k-quant mmq and ggml-backend offload functionali... |
commit | commitdiff | tree |
| 2024-04-06 |
Neo Zhang Jianyu | fix set main gpu crash (llama/6339) |
commit | commitdiff | tree |
| 2024-04-06 |
slaren | ggml : fix bounds checking of zero size views (llama... |
commit | commitdiff | tree |
| 2024-04-03 |
Daniel Bevenius | backend : fix typo in scheduler documentation (#781) |
commit | commitdiff | tree |
| 2024-04-03 |
Andrei | cmake : add `GGML_METAL_EMBED_LIBRARY` (#780) |
commit | commitdiff | tree |
| 2024-04-03 |
Georgi Gerganov | logo : add files (#782) |
commit | commitdiff | tree |
| 2024-03-27 |
JacobLinCool | gguf : update type enum (#775) |
commit | commitdiff | tree |
| 2024-03-27 |
Georgi Gerganov | sync : whisper.cpp |
commit | commitdiff | tree |
| 2024-03-27 |
Georgi Gerganov | sync : whisper.cpp |
commit | commitdiff | tree |
| 2024-03-27 |
Georgi Gerganov | whisper : improve handling of prompts (whisper/1981) |
commit | commitdiff | tree |
| 2024-03-27 |
Sanchit Gandhi | whisper : improve support for distil-large-v3 (whisper... |
commit | commitdiff | tree |
| 2024-03-27 |
denersc | whisper : token-level timestamps with DTW (whisper... |
commit | commitdiff | tree |
| 2024-03-27 |
Jo Liss | examples : rename --audio-context to --audio-ctx per... |
commit | commitdiff | tree |
| 2024-03-27 |
Georgi Gerganov | examples : more CUDA leftovers (#0) |
commit | commitdiff | tree |
| 2024-03-27 |
Georgi Gerganov | examples : fix CUBLAS leftovers (#0) |
commit | commitdiff | tree |
| 2024-03-27 |
Georgi Gerganov | sync : adapt to CUDA changes (#0) |
commit | commitdiff | tree |
| 2024-03-27 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| 2024-03-27 |
AidanBeltonS | Fix batched impl for NVidia GPU (llama/6164) |
commit | commitdiff | tree |
| 2024-03-27 |
Kawrakow | Make IQ1_M work for QK_K = 64 (llama/6327) |
commit | commitdiff | tree |
| 2024-03-27 |
compilade | llama : greatly reduce output buffer memory usage ... |
commit | commitdiff | tree |
| 2024-03-27 |
Kawrakow | IQ1_M: 1.75 bpw quantization (llama/6302) |
commit | commitdiff | tree |
| 2024-03-27 |
slaren | cuda : rename build flag to LLAMA_CUDA (llama/6299) |
commit | commitdiff | tree |
| 2024-03-27 |
Kawrakow | tests : include IQ2_XXS and IQ2_XS in test-quantize... |
commit | commitdiff | tree |
| 2024-03-27 |
slaren | cuda : refactor into multiple files (llama/6269) |
commit | commitdiff | tree |
| 2024-03-27 |
Justine Tunney | ggml : support AVX512VNNI (llama/6280) |
commit | commitdiff | tree |
| 2024-03-27 |
Rick G | Fix heap corruption from wmode out-of-bound writes... |
commit | commitdiff | tree |
| 2024-03-27 |
Meng, Hengyu | offload op (llama/6217) |
commit | commitdiff | tree |
| 2024-03-27 |
Jared Van Bortel | use _wfopen instead of fopen on Windows (llama/6248) |
commit | commitdiff | tree |
| 2024-03-27 |
slaren | cuda : add LLAMA_CUDA_NO_PEER_COPY to workaround broken... |
commit | commitdiff | tree |
| 2024-03-27 |
Georgi Gerganov | metal : proper assert for mat-mat memory alignment... |
commit | commitdiff | tree |
| 2024-03-27 |
Georgi Gerganov | metal : pad n_ctx by 32 (llama/6177) |
commit | commitdiff | tree |
| 2024-03-27 |
slaren | cuda : disable host register by default (llama/6206) |
commit | commitdiff | tree |
| 2024-03-27 |
slaren | cuda : fix LLAMA_CUDA_F16 build (llama/6197) |
commit | commitdiff | tree |
| 2024-03-27 |
Kawrakow | ggml : same IQ4_NL quantization for CPU/CUDA/Metal... |
commit | commitdiff | tree |
| 2024-03-27 |
Kawrakow | Add ability to use Q5_0, Q5_1, and IQ4_NL for quantized... |
commit | commitdiff | tree |
| 2024-03-27 |
AidanBeltonS | Add nvidia and amd backends (llama/6157) |
commit | commitdiff | tree |
| 2024-03-27 |
slaren | cuda : fix conflict with std::swap (llama/6186) |
commit | commitdiff | tree |
| next |
overview / pkg / ggml / sources / ggml / shortlog