 Molly Sophia
|
ee7136c6d1
llama: add support for QRWKV6 model architecture (#11001)
|
1 year ago |
 Jeff Bolz
|
716bd6dec3
vulkan: optimize mul_mat for small values of N (#10991)
|
1 year ago |
|
Jeff Bolz
|
a813badbbd
vulkan: im2col and matmul optimizations for stable diffusion (#10942)
|
1 year ago |
 Georgi Gerganov
|
0006f5a74a
ggml : update ggml_backend_cpu_device_supports_op (#10867)
|
1 year ago |
 HimariO
|
ba1cb19cdd
llama : add Qwen2VL support + multimodal RoPE (#10361)
|
1 year ago |
 PAB
|
a8cbab201d
ggml: add `GGML_SET` Metal kernel + i32 CPU kernel (ggml/1037)
|
1 year ago |
|
PAB
|
c2082d93a8
ggml : add `GGML_PAD_REFLECT_1D` operation (ggml/1034)
|
1 year ago |
|
Jeff Bolz
|
2759916d86
vulkan: Implement "fast divide" (mul+shift) for unary ops like copy (#10642)
|
1 year ago |
|
PAB
|
efb6ae9630
feat: add `GGML_UNARY_OP_ARGMAX` Metal kernel (ggml/1019)
|
1 year ago |
|
Georgi Gerganov
|
0115df2f65
metal : small-batch mat-mul kernels (#10581)
|
1 year ago |
|
Georgi Gerganov
|
f0678c5ff4
ggml : fix I8MM Q4_1 scaling factor conversion (#10562)
|
1 year ago |
|
Jeff Bolz
|
904109ed0d
vulkan: fix group_norm (#10496)
|
1 year ago |
 Diego Devesa
|
5931c1f233
ggml : add support for dynamic loading of backends (#10469)
|
1 year ago |
|
Diego Devesa
|
a5e47592b6
cuda : optimize argmax (#10441)
|
1 year ago |
 Johannes Gäßler
|
02e4eaf22f
ggml-opt: fix data corruption (ggml/1022)
|
1 year ago |
|
Jeff Bolz
|
b3e585988f
vulkan: Optimize soft_max (#10301)
|
1 year ago |
|
Johannes Gäßler
|
8a43e940ab
ggml: new optimization interface (ggml/988)
|
1 year ago |
|
Jeff Bolz
|
80dd7ff22f
vulkan: Optimize contiguous copies (#10254)
|
1 year ago |
|
Georgi Gerganov
|
841f27abdb
metal : optimize FA kernels (#10171)
|
1 year ago |
 Zhiyuan Li
|
3bcd40b3c5
Optimize RWKV6 Operator Naming and Implement Multi-core CPU/ SYCL Acceleration (#10133)
|
1 year ago |
|
Georgi Gerganov
|
5c333e0140
metal : add BF16 support (#8439)
|
1 year ago |
|
Diego Devesa
|
9f40989351
ggml : move CPU backend to a separate file (#10144)
|
1 year ago |
|
Johannes Gäßler
|
c39665f589
CUDA: fix MMQ for non-contiguous src0, add tests (#10021)
|
1 year ago |
|
Johannes Gäßler
|
80273a306d
CUDA: fix 1D im2col, add tests (ggml/993)
|
1 year ago |
 Jun Hee Yoo
|
4c9388fb96
metal : add POOL2D and fix IM2COL (#9943)
|
1 year ago |
|
Diego Devesa
|
dca1d4b58a
ggml : fix BLAS with unsupported types (#9775)
|
1 year ago |
|
Diego Devesa
|
6374743747
ggml : add backend registry / device interfaces to BLAS backend (#9752)
|
1 year ago |
|
Johannes Gäßler
|
fabdc3bda3
ggml/ex: calculate accuracy in graph, adapt MNIST (ggml/980)
|
1 year ago |
|
Diego Devesa
|
c83ad6d01e
ggml-backend : add device and backend reg interfaces (#9707)
|
1 year ago |
|
Johannes Gäßler
|
e98c1c188e
test: fix OPT_STEP_ADAMW for test-backend-ops (ggml/974)
|
1 year ago |
