 Georgi Gerganov
|
6aeff24f8b
metal : fix indent (ggml/0)
|
1 year ago |
 Justina Cho
|
f5ef34e428
feat: implemented sigmoid function (ggml/806)
|
1 year ago |
|
Georgi Gerganov
|
9cb317f77e
ggml : full ALiBi support (#7192)
|
1 year ago |
|
Georgi Gerganov
|
18e437665c
metal : fix flash attention kernel requirements (#7169)
|
1 year ago |
 Gilad S
|
26458af1d6
metal : use `vm_allocate` instead of `posix_memalign` on macOS (#7078)
|
1 year ago |
 Justine Tunney
|
3855416027
ggml : introduce bfloat16 support (#6412)
|
1 year ago |
 Kevin Gibbons
|
f364eb6fb5
switch to using localizedDescription (#7010)
|
1 year ago |
|
Georgi Gerganov
|
77e15bec62
metal : remove deprecated error code (#7008)
|
1 year ago |
|
Kevin Gibbons
|
a68a1e7ed0
metal : log more info on error (#6987)
|
1 year ago |
|
Georgi Gerganov
|
9c67c2773d
ggml : add Flash Attention (#5021)
|
1 year ago |
 slaren
|
0d56246f4b
ggml : group all experts in a single ggml_mul_mat_id (#6505)
|
1 year ago |
 Shijie
|
f4dea7da18
llama : add qwen2moe (#6074)
|
1 year ago |
 Dave
|
422c2aff1c
Added support for GGML_OP_CLAMP in Metal (#6662)
|
1 year ago |
|
slaren
|
fbbc030ba9
metal : unify mul_mv_id kernels (#6556)
|
1 year ago |
|
slaren
|
08a0c02060
ggml : mul_mat_id use the same tensor for all the experts (#6387)
|
1 year ago |
 compilade
|
557410b8f0
llama : greatly reduce output buffer memory usage (#6122)
|
1 year ago |
 Kawrakow
|
55c1b2a3bb
IQ1_M: 1.75 bpw quantization (#6302)
|
1 year ago |
|
Georgi Gerganov
|
b3e94f26ba
metal : proper assert for mat-mat memory alignment (#6225)
|
1 year ago |
|
Kawrakow
|
76aa30a263
Add ability to use Q5_0, Q5_1, and IQ4_NL for quantized K cache (#6183)
|
1 year ago |
|
slaren
|
2bf8d0f7c4
backend : offload large batches to GPU (#6083)
|
1 year ago |
|
Georgi Gerganov
|
381da2d9f0
metal : build metallib + fix embed path (#6015)
|
1 year ago |
|
slaren
|
f30ea47a87
llama : add pipeline parallelism support (#6017)
|
1 year ago |
|
Georgi Gerganov
|
8030da7afe
ggml : reuse quantum structs across backends (#5943)
|
1 year ago |
|
Georgi Gerganov
|
bb6d00bbf9
metal : move mm_id indices to shared mem (#5982)
|
1 year ago |
 Michael Podvitskiy
|
9fa2627347
ggml : introduce ggml_status (ggml/750)
|
1 year ago |
 leejet
|
7d43c585dc
add some new ops, fix some operators and add batch operations to certain operators. (ggml/747)
|
1 year ago |
 UEXTM.com
|
5f70671856
Introduce backend GUIDs (ggml/743)
|
1 year ago |
|
Kawrakow
|
0becb22ac0
IQ4_XS: a 4.25 bpw quantization (#5747)
|
1 year ago |
|
Kawrakow
|
a33e6a0d2a
Adding IQ2_S and IQ2_M to complete coverage of the 2-3 bit quantization range (#5721)
|
1 year ago |
|
Georgi Gerganov
|
ab336a9d5e
code : normalize enum names (#5697)
|
1 year ago |
