OpenI
/
MegEngine
Not watched
1
0
Fork 0
Code
Releases 31 Wiki Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain
2612 Commits
28 Branches
25 MB
3774 Download
Tree: f0088335bb
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'f0088335bb'
${ noResults }
Commit Graph

1149 Commits (f0088335bbe5e91d3463ca756e7012cfcc2d645a)

Author SHA1 Message Date
  Megvii Engine Team 657db8dce2 chore(tools): remove dump_with_testcase_mge, user should use jit.dump instead 3 years ago
  Megvii Engine Team d3247bee5a fix(dtr): always write shape when tensor produced 3 years ago
  Megvii Engine Team 1192a9a61b fix(imperative): fix adaptive pool2d 3 years ago
  Megvii Engine Team 7b4b94fd93 fix(imperative): fix the segmentfault when reduce backward 3 years ago
  Megvii Engine Team 24c5c19bf0 fix(imperative): make functional ops support negative axis 3 years ago
  Megvii Engine Team c76e80bc8e fix(mge): fix mem leak of numpy method of DeviceTensorND 3 years ago
  Megvii Engine Team f96429c031 feat(imperative): support empty tensor in roi_align 3 years ago
  Megvii Engine Team 2f829aaa7b test(imperative): speed up integration test 3 years ago
  Megvii Engine Team 07bdb3bf1e feat(imperative): add swapaxes 3 years ago
  Megvii Engine Team e694301721 fix(ops/jit): skip lookup include path when nvcc executable not found 3 years ago
  Megvii Engine Team 02c1a0c3de fix(imperative/format): warn once when parameter without format been attached 3 years ago
  Megvii Engine Team c2deef1a97 feat(mge): aad atlas710 support 3 years ago
  Megvii Engine Team df5ebd3da7 fix(imperative/ops): fix the vmemory problem in 1.9 3 years ago
  Megvii Engine Team b36b5bd8cb refactor(mgb): check input when profiling 3 years ago
  Megvii Engine Team 8563f51404 fix(imperative): fix buildin reduce keepdim 3 years ago
  Megvii Engine Team eef0308b29 feat(serialization): add no_change_graph and version param whem dump model 3 years ago
  Megvii Engine Team 4ab5f970e9 fix(build): fix ci error 3 years ago
  Megvii Engine Team b9a6932341 feat(imperative): channel default model format to fbs v2 3 years ago
  Megvii Engine Team da91e650a5 refactor(ops/layer_norm): speed up the host speed of layer_norm 3 years ago
  Megvii Engine Team 67cfce9f2f fix(imperative/amp): add is_scalar check in elemwise and concat 3 years ago
  Megvii Engine Team d313f92610 fix(imperative/amp): fix format transformation for symbol trans 3 years ago
  Megvii Engine Team 261a5bce23 feat(imperative/amp): add dimshuffle in set_format for nhwc 3 years ago
  Megvii Engine Team c9e56f4987 feat(imperative/amp): add dimshuffle before creating nhwc tensor 3 years ago
  Megvii Engine Team d57a071271 feat(imperative/amp): add fallback for op not supported for nhwc tensor 3 years ago
  Megvii Engine Team 38a9aa9faf feat(imperative/amp): add auto dimshuffle for elemwise and concat 3 years ago
  Megvii Engine Team cd26376549 style(imperative/amp): reformat code 3 years ago
  Megvii Engine Team 3892aa0b6e fix(imperative/amp): fix bn params for nhwc amp 3 years ago
  Megvii Engine Team 6f0b582064 chore(imperative/amp): adapt dev 3 years ago
  Megvii Engine Team ee984e8608 fix(imperative/amp): fix distributed backward callback for nhwc amp 3 years ago
  Megvii Engine Team 15c6da6218 feat(imperative/amp): add nhwc support for adaptive pooling 3 years ago
  Megvii Engine Team c28a875fac fix(imperative/amp): adapt new transformation 3 years ago
  Megvii Engine Team fd41302cc1 feat(imperative/amp): add set_format 3 years ago
  Megvii Engine Team fc633ce4ff fix(imperative/amp): fix custom grad in Subgraph 3 years ago
  Megvii Engine Team 673b295d75 feat(imperative/amp): remove conv_format and bn param_dim configs 3 years ago
  Megvii Engine Team 7e9aa742e6 feat(imperative/amp): enable auto_convert_format by default 3 years ago
  Megvii Engine Team fc0f454685 fix(dnn/check_non_finite): adjust some details of CheckNonFinite 3 years ago
  Megvii Engine Team e393d1cf65 feat(mge/amp): add convert_format module for NHWC training 3 years ago
  Megvii Engine Team 533fb5bf49 feat(imperative): support formatted tensor and add special op rules 3 years ago
  Megvii Engine Team 4aa79c453b perf(mge): override grad of matmul 3 years ago
  Megvii Engine Team 98b5ee78c1 feat(mge/dnn): add lamb optimizer 3 years ago
  Megvii Engine Team a926878c01 feat(imperative): remove symbolvar of imperative 3 years ago
  Megvii Engine Team c2e9860feb chore(license): remove all license in file header 3 years ago
  Megvii Engine Team 1783b8977a feat(profiler): integrate cupti backend 3 years ago
  Megvii Engine Team 58df717e37 fix(mge/autodiff): fix attaching tensor already in gradient path 3 years ago
  Megvii Engine Team 8182af6eb6 fix(mgb): fix strategy of grad_op and opr_attr 3 years ago
  Megvii Engine Team 597a1e791b refactor(imperative): add interface to clear algorithm cache 3 years ago
  Megvii Engine Team e2f5156b69 refactor(megbrain): save fastrun result to algorithm cache 3 years ago
  Megvii Engine Team dbce6526d6 fix(mge/functional): fix return dtype of comparison function 3 years ago
  Megvii Engine Team b92866d2c2 fix(build): fix build depends dirty file issue 3 years ago
  Megvii Engine Team 4b27e861f4 fix(ops): implement from_op_node for reshape 3 years ago