 Diego Devesa
|
dca1d4b58a
ggml : fix BLAS with unsupported types (#9775)
|
vor 1 Jahr |
|
Diego Devesa
|
6374743747
ggml : add backend registry / device interfaces to BLAS backend (#9752)
|
vor 1 Jahr |
 Johannes Gäßler
|
fabdc3bda3
ggml/ex: calculate accuracy in graph, adapt MNIST (ggml/980)
|
vor 1 Jahr |
|
Diego Devesa
|
c83ad6d01e
ggml-backend : add device and backend reg interfaces (#9707)
|
vor 1 Jahr |
|
Johannes Gäßler
|
e98c1c188e
test: fix OPT_STEP_ADAMW for test-backend-ops (ggml/974)
|
vor 1 Jahr |
|
Johannes Gäßler
|
7254cdf7e8
ggml: fix gradient allocation logic (ggml/966)
|
vor 1 Jahr |
|
slaren
|
1b2f992cd2
test-backend-ops : use flops for some performance tests (#9657)
|
vor 1 Jahr |
|
Johannes Gäßler
|
a5b57b08ce
CUDA: enable Gemma FA for HIP/Pascal (#9581)
|
vor 1 Jahr |
 Molly Sophia
|
2a63caaa69
RWKV v6: RWKV_WKV op CUDA implementation (#9454)
|
vor 1 Jahr |
|
Johannes Gäßler
|
424c5d00a9
ggml/examples: add backend support for numerical optimization (ggml/949)
|
vor 1 Jahr |
 Georgi Gerganov
|
d6a04f872d
ggml : hide ggml_object, ggml_cgraph, ggml_hash_set (#9408)
|
vor 1 Jahr |
|
Georgi Gerganov
|
a876861455
metal : update support condition for im2col + fix warning (#0)
|
vor 1 Jahr |
|
Johannes Gäßler
|
202084d31d
tests: add gradient tests for all backends (ggml/932)
|
vor 1 Jahr |
 Salvatore Mesoraca
|
efe6a83e30
ggml : fix cont with transposed tensors when one dimension is 1 (ggml/934)
|
vor 1 Jahr |
 compilade
|
9bc6db28d0
ggml-quants : ternary packing for TriLMs and BitNet b1.58 (#8151)
|
vor 1 Jahr |
|
Georgi Gerganov
|
231cff5f6f
sync : ggml
|
vor 1 Jahr |
|
Georgi Gerganov
|
fc18425b6a
ggml : add SSM Metal kernels (#8546)
|
vor 1 Jahr |
|
slaren
|
0c41e03ceb
metal : gemma2 flash attention support (#9159)
|
vor 1 Jahr |
|
Johannes Gäßler
|
e11bd856d5
CPU/CUDA: Gemma 2 FlashAttention support (#8542)
|
vor 1 Jahr |
 zhentaoyu
|
4f8d19ff17
[SYCL] Fix SYCL `im2col` and `convert` Overflow with Large Dims (#9052)
|
vor 1 Jahr |
|
Molly Sophia
|
2d5dd7bb3f
ggml : add epsilon as a parameter for group_norm (#8818)
|
vor 1 Jahr |
 0cc4m
|
064cdc265f
vulkan : fix Qantized Mat-Vec Mul on AMD GPUs for ncols < 64 (#8855)
|
vor 1 Jahr |
 Mengqing Cao
|
e09a800f9a
cann: Fix ggml_cann_im2col for 1D im2col (#8819)
|
vor 1 Jahr |
|
slaren
|
7a11eb3a26
cuda : fix dmmv cols requirement to 2*GGML_CUDA_DMMV_X (#8800)
|
vor 1 Jahr |
|
slaren
|
2b1f616b20
ggml : reduce hash table reset cost (#8698)
|
vor 1 Jahr |
|
slaren
|
87e397d00b
ggml : fix quant dot product with odd number of blocks (#8549)
|
vor 1 Jahr |
 hipudding
|
1bdd8ae19f
[CANN] Add Ascend NPU backend (#6035)
|
vor 1 Jahr |
|
Georgi Gerganov
|
6847d54c4f
tests : fix whitespace (#0)
|
vor 1 Jahr |
 John Balis
|
fde13b3bb9
feat: cuda implementation for `ggml_conv_transpose_1d` (ggml/854)
|
vor 1 Jahr |
|
slaren
|
0e0590adab
cuda : update supports_op for matrix multiplication (#8245)
|
vor 1 Jahr |
