slaren
|
0d56246f4b
ggml : group all experts in a single ggml_mul_mat_id (#6505)
|
1 rok temu |
Shijie
|
f4dea7da18
llama : add qwen2moe (#6074)
|
1 rok temu |
Dave
|
422c2aff1c
Added support for GGML_OP_CLAMP in Metal (#6662)
|
1 rok temu |
slaren
|
fbbc030ba9
metal : unify mul_mv_id kernels (#6556)
|
1 rok temu |
slaren
|
08a0c02060
ggml : mul_mat_id use the same tensor for all the experts (#6387)
|
1 rok temu |
compilade
|
557410b8f0
llama : greatly reduce output buffer memory usage (#6122)
|
1 rok temu |
Kawrakow
|
55c1b2a3bb
IQ1_M: 1.75 bpw quantization (#6302)
|
1 rok temu |
Georgi Gerganov
|
b3e94f26ba
metal : proper assert for mat-mat memory alignment (#6225)
|
1 rok temu |
Kawrakow
|
76aa30a263
Add ability to use Q5_0, Q5_1, and IQ4_NL for quantized K cache (#6183)
|
1 rok temu |
slaren
|
2bf8d0f7c4
backend : offload large batches to GPU (#6083)
|
1 rok temu |
Georgi Gerganov
|
381da2d9f0
metal : build metallib + fix embed path (#6015)
|
1 rok temu |
slaren
|
f30ea47a87
llama : add pipeline parallelism support (#6017)
|
1 rok temu |
Georgi Gerganov
|
8030da7afe
ggml : reuse quantum structs across backends (#5943)
|
1 rok temu |
Georgi Gerganov
|
bb6d00bbf9
metal : move mm_id indices to shared mem (#5982)
|
1 rok temu |
Michael Podvitskiy
|
9fa2627347
ggml : introduce ggml_status (ggml/750)
|
1 rok temu |
leejet
|
7d43c585dc
add some new ops, fix some operators and add batch operations to certain operators. (ggml/747)
|
1 rok temu |
UEXTM.com
|
5f70671856
Introduce backend GUIDs (ggml/743)
|
1 rok temu |
Kawrakow
|
0becb22ac0
IQ4_XS: a 4.25 bpw quantization (#5747)
|
1 rok temu |
Kawrakow
|
a33e6a0d2a
Adding IQ2_S and IQ2_M to complete coverage of the 2-3 bit quantization range (#5721)
|
1 rok temu |
Georgi Gerganov
|
ab336a9d5e
code : normalize enum names (#5697)
|
1 rok temu |
Kawrakow
|
4c4cb30736
IQ3_S: a much better alternative to Q3_K (#5676)
|
1 rok temu |
Kawrakow
|
a14679cc30
IQ4_NL: 4-bit non-linear quants with blocks of 32 (#5590)
|
1 rok temu |
Didzis Gosko
|
890559ab28
metal : option to embed MSL source into compiled binary (whisper/1842)
|
1 rok temu |
Kawrakow
|
bd2d4e393b
1.5 bit quantization (#5453)
|
1 rok temu |
Georgi Gerganov
|
8f1be0d42f
ggml : add ALiBi support for ggml_soft_max_ext (#5488)
|
1 rok temu |
Ananta Bastola
|
6e4e973b26
ci : add an option to fail on compile warning (#3952)
|
1 rok temu |
Ian Bull
|
f026f8120f
metal : use autoreleasepool to avoid memory leaks (#5437)
|
1 rok temu |
Georgi Gerganov
|
efb7bdbbd0
metal : add im2col F32 dst support (#5132)
|
1 rok temu |
Georgi Gerganov
|
549a1e6cd5
ci : fix yolo URLs + fix metal capture (ggml/712)
|
2 lat temu |
Jack Mousseau
|
5f14ee0b0c
metal : add debug capture backend function (ggml/694)
|
2 lat temu |