Skip to content
Snippets Groups Projects
Select Git revision
  • summer default
1 result
You can move around the graph by using the arrow keys.
Created with Raphaël 2.2.013Jun12111098765432131May30292827262524232221201918171615141312111098765130Apr2928272625242322212019171615141312119876532131Mar302928262524232221201918171615141312111098765432128Feb27InstanceNorm module (#5131)Skip building cares executables (#5189)Add CMake argument PIP_INDEX_MIRROR (#5188)Always build test exe so CI can check it (#5187)Yejiaojiao/dev acos (#5031)Fix which nightly release read the doc uses (#5186)Dev sum/mean/min/max modules (#5022)add backward op for flatten and softmax; fix bug for softmax in dim 1 case (#5158)Delete squeeze and unsqueeze backward useless code (#5183)Fix target build dependency. (#5185)Refine imports in package mirror (#5184)forbig in source build (#5182)L1_Loss (#5134)Update generated functional files only if different. (#5177)Update version info 0.5.0 (#5161)Dev topk eager (#5142)Fix unpack_call. (#5180)Fix eq module bug (#5166)Use different tag for ci build and ci release (#5170)Use composed attrs when creating kernel state. (#5171)Dev functional interface. (#4932)Align dropout module with pytorch (#5154)Export Optimizer and _LRScheduler (#5156)Fix argwhere module bug (#5149)Dynamic link cudnn (#5152)refine and add test case (#5157)Add RecordingReplay debug tool (#5121)use cpu in read the doc to prevent fail (#5150)Fix slice bug (#5117)add eager experimental stack (#5133)fix bert backward bug (#5138)Add int8 support for expand and repeat (#5139)style(Optimzier): refine optimizer interface (#5145)Compile C++, CUDA targets in parallel (#5147)Add ceil & expm1 module (#4953)Fix doctest for modules (#5130)add pixel_shuffle module (#5135)Add convtranspose module (#5101)implement tensor.register_hook (#5128)add mseloss module (#5116)