 Jeff Bolz
|
9bf20d8ac3
vulkan: Use BK=32 for coopmat2 mul_mat_id (#18332)
|
1 месяц назад |
|
Jeff Bolz
|
b96b82fc85
vulkan: Support UPSCALE w/antialias (#18327)
|
1 месяц назад |
|
Jeff Bolz
|
10dc500bdb
vulkan: handle rope with large number of rows (#18306)
|
1 месяц назад |
|
Jeff Bolz
|
2a9ea2020c
vulkan: fix command buffer corruption in ggml_backend_vk_event_wait (#18302)
|
1 месяц назад |
 Ruben Ortlam
|
7f459c98e7
vulkan: use fewer FA rows for small cache runs (#18280)
|
1 месяц назад |
|
Jeff Bolz
|
e3b35ddf1c
vulkan: Extend rope fusions to allow mrope (#18264)
|
1 месяц назад |
|
Jeff Bolz
|
e1f15b454f
vulkan: Implement set_tensor_async and the event interfaces (#18047)
|
1 месяц назад |
|
Jeff Bolz
|
fd05c51cec
vulkan: fix im2col overflowing maxworkgroupcount (#18180)
|
1 месяц назад |
|
Jeff Bolz
|
b365c3ff01
vulkan/cuda: fix topk_moe with exp_probs_b (#18071)
|
1 месяц назад |
|
Jeff Bolz
|
cb64222b0c
vulkan: support GGML_UNARY_OP_XIELU (#18062)
|
1 месяц назад |
|
Jeff Bolz
|
6eb7081860
vulkan: in graph_optimize, try to group ADD operations (#18060)
|
1 месяц назад |
|
Jeff Bolz
|
cdbada8d10
vulkan: Add perf logger mode with concurrency (#17944)
|
1 месяц назад |
|
Jeff Bolz
|
36255a2268
vulkan: support get_rows for i32 (#17941)
|
1 месяц назад |
|
Jeff Bolz
|
3229a23fa6
vulkan: support GGML_OP_DIAG (#17893)
|
1 месяц назад |
|
Jeff Bolz
|
303f8615e9
vulkan: Multi-pass softmax for large number of cols (#17892)
|
1 месяц назад |
|
Jeff Bolz
|
07a10c1090
vulkan: Allow non-pow2 n_experts in topk_moe (#17872)
|
1 месяц назад |
|
Jeff Bolz
|
db97837385
vulkan: perf_logger improvements (#17672)
|
1 месяц назад |
 Phylliida Dev
|
09c7c50e64
ggml : add circular tiling support to pad, for Vulkan, CUDA, and CPU (used for making seamless textures) (#16985)
|
1 месяц назад |
|
Jeff Bolz
|
2960eb2975
vulkan: Use one row per workgroup for f32 mmv (#17711)
|
1 месяц назад |
|
Jeff Bolz
|
c6c5e85979
vulkan: support solve_tri with larger N/K values (#17781)
|
1 месяц назад |
 Masato Nakasaka
|
67788f6846
vulkan: Replace deprecated VK_EXT_validation_features (#17637)
|
1 месяц назад |
|
Masato Nakasaka
|
d8c0a7b085
vulkan: Fix mismatch in TOPK_MOE unit test (#17541)
|
1 месяц назад |
|
Jeff Bolz
|
a0f3897d53
vulkan: fix top_k bug when there are ties in the input (#17659)
|
1 месяц назад |
 Acly
|
e15cd06a94
vulkan : support conv-2d with large output size (#17685)
|
1 месяц назад |
|
Jeff Bolz
|
6ab0d64960
vulkan: enable mmvq for q2_k on NVIDIA (#17675)
|
1 месяц назад |
|
Jeff Bolz
|
93bb92664e
vulkan: set all memory allocations to high priority (#17624)
|
1 месяц назад |
|
Jeff Bolz
|
61bde8e21f
vulkan: Reduce temporary memory usage for TOP_K (#17623)
|
2 месяцев назад |
 Tarek Dakhran
|
2ba719519d
model: LFM2-VL fixes (#17577)
|
2 месяцев назад |
|
Ruben Ortlam
|
47a268ea50
Vulkan: MMVQ Integer Dot K-Quant and MUL_MAT_ID support (#16900)
|
2 месяцев назад |
|
Jeff Bolz
|
59d8d4e963
vulkan: improve topk perf for large k, fix overflow in unit tests (#17582)
|
2 месяцев назад |
