Spaces:
Running
Running
Commit History
llama/ggml: add LLM training support (llama/10544)
8d3b3c1
Add `--no-op-offload` to improve `-ot` pp perf in MoE models like llama4 400B (llama/13386)
418769d
David Huang
commited on
CUDA: fix bad asserts for partial offload (llama/13337)
23e676b
CUDA: fix logic for clearing padding with -ngl 0 (llama/13320)
c3e51a2
CUDA: fix q_nope_absorbed prec for DS 2 Lite f16 (llama/13137)
e9c9d4b
ggml: move fp16/bf16 conversion optimizations to CPU backend + export conversion APIs (llama/13107)
c47823e
rpc : do not wait for response when sending RPC_CMD_SET_TENSOR (llama/12943)
691c071
ggml : fix ggml_gallocr_ptr type (ggml/1205)
cf46d5c
Diego Devesa
commited on
rpc : add RPC_CMD_HELLO (llama/12955)
ff22836
ggml : Depthwise 2D convolution (ggml/1152)
0c950d5
ggml : add bilinear upscale support (ggml/1185)
4c5e449
Diego Devesa
commited on
ggml : add more generic custom op, remove deprecated custom ops (ggml/1183)
ba7a5f8
Diego Devesa
commited on
metal : improve FA + improve MoE (llama/12612)
04a3389
rpc : send hash when tensor data is above some fixed threshold (llama/12496)
c39f9c4
llama: Add support for RWKV v7 architecture (llama/12412)
727de7e
ggml-cpu: Faster IQ1 mul_mat_vec on AVX2 using BMI2 instructions (llama/12154)
05466a9
Rémy O
commited on
ggml : portability fixes for VS 2017 (llama/12150)
49e3343
mgroeber9110
Marcus Groeber
commited on
ggml : upgrade init_tensor API to return a ggml_status (llama/11854)
d6b6852
William Tambellini
slaren
commited on
ggml-cpu: Support s390x SIMD Instruction Set (llama/12019)
4aa54ec
Aaron Teo
Jinyang He
junchao-zhao
commited on
ggml-cpu: Add CPU backend support for KleidiAI library (llama/11390)
9de6d81
Charles Xu
commited on
repo : update links to new url (llama/11886)
9705bb5
cleanup: fix compile warnings associated with gnu_printf (llama/11811)
ef6a968
bandoti
commited on
vulkan: Make Vulkan optional at runtime (ggml/11493). (llama/11494)
762f497
CUDA: use mma PTX instructions for FlashAttention (llama/11583)
f328957
rpc : early register backend devices (llama/11262)
4134077
CUDA: backwards pass for misc. ops, add tests (llama/11257)
2fbcec1
RoPE: fix back, CUDA support for back + noncont. (llama/11240)
131a21e
GGUF: C++ refactor, backend support, misc fixes (skip) (llama/11030)
92311a3
GGUF: C++ refactor, backend support, misc fixes (llama/11030)
21c5b64
tts : add OuteTTS support (llama/10784)
8d0f0ac
Introducing experimental OpenCL backend with support for Qualcomm Adreno GPUs (llama/10693)
83a0899
lhez
Skyler Szot
Shangqing Gu
Alexander Angus
Hongqiang Wang
Max Krasnyansky
commited on
ggml: load all backends from a user-provided search path (llama/10699)
c6de218
Gilad S
Diego Devesa
commited on
ggml : refactor online repacking (llama/10446)
163128e
ggml : remove old files (skip) (#0)
6284570
unverified
ggml : add `GGML_PAD_REFLECT_1D` operation (ggml/1034)
154bbc0
ggml-cpu: support IQ4_NL_4_4 by runtime repack (llama/10541)
bf73242
ggml : add support for dynamic loading of backends (llama/10469)
b73266f
ggml: new optimization interface (ggml/988)
dd33ace
backend cpu: add online flow for aarch64 Q4_0 GEMV/GEMM kernels (llama/9921)
3541ee8
Charles Xu
Diego Devesa
commited on
ggml : build backends as libraries (llama/10256)
3dc93f3
metal : optimize FA kernels (llama/10171)
44ff932
ggml : move CPU backend to a separate file (llama/10144)
0f447f2
Diego Devesa
commited on
llama : add simple-chat example (llama/10124)
41ff26f
Diego Devesa
Xuan Son Nguyen
commited on
llama : use smart pointers for ggml resources (llama/10117)
6b82135
Diego Devesa
commited on
