19 Commits (c0a94cd9cabacb686b2423702a42e4a4765b746f)

Author SHA1 Message Date
  nihui c0a94cd9ca
fix armv7 without neon (#3514) 4 years ago
  nihui 6941ec8fc9
arm neon optimization for general packed convolution (#3426) 4 years ago
  nihui cdf45a6512
cmake option NCNN_BF16 (#3068) 5 years ago
  Leighton Choi 44518f457a
Support negative axis in concat, slice and softmax (#2365) 5 years ago
  nihui 46565025b6 x86 slice pack4 5 years ago
  Leo 5afd318b86
Support remove libstdc++ denpendency (#2030) 5 years ago
  nihui aa68246dc7 more test coverage 6 years ago
  nihui e7abc5fbd7 concat slice arm fp16sa pack8 6 years ago
  nihuini f23122bb3f since fp16 storage option is on by default, upper-level function may pass fp32 storage with default option, guard with element bits checking 6 years ago
  nihuini 47ae0c151a some shared arm bf16s fp16s implementation 6 years ago
  nihui b5e288b521 layer creator function is not necessary for built-in layers 6 years ago
  nihui 01b8b79ed2 packing layout option respect support_packing property 6 years ago
  nihui 3ef995ed1e
format code style and setup restyled.io (#1840) 6 years ago
  nihui 7a89ce6223 slice bf16s 6 years ago
  nihuini 1ad1949fe7 slice pack4 6 years ago
  nihuini aee6552360 memcpy is fast :) 8 years ago
  dong 6ea09ebf2c Use aarch64 assembly to replace arm intrinsics 8 years ago
  nihuini a84ba8fc0f element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
  nihuini b7db8be4f6 add ncnn source qwq 9 years ago