可以使用方向键移动图形。
Created with Raphaël 2.2.011Jan518Dec11427Nov23212013530Oct262220141317Sep161529Aug26252412115423Jul212014fix(mge): fix cpp trace function releaserefactor(mge): add specialized grad rule supporttest(mge/autodiff): add test for weird errorrefactor(imperative): remove multidispatch, raw_tensor, registerfeat(mge): restore Functionfeat(mge): add python custom oprefactor(mge): refactor to prepare for custom grad rulestest(mge): fix megbrain_graph/cgtools testfix(mge): fix none return value for attrs, add test_correctnessrefactor(mge/imperative): fix compilation for Python 3.5perf(mge/imperative): move convert_inputs from python to C++feat(mge/imperative): implement trace and dump under new core implementationperf(mge/imperative): add mini graph to partially replace proxy graphrefactor(mge/tensor): remove old implementationrefactor(mge/imperative): migrate to new core implementationrefactor(mge/imperative): move detach into C++refactor(mge/imperative): implement new tensor systemchore(third_party): add range-v3 as third_party dependencyfix(mgb/opr): move NVOF opr's shape inference to execute partfeat(opr): add param(axis) for GetVarShapefix(mge/imperative): fix op specializations and tuple hashtest(mge/autodiff): validate grad is set to None when no dependencyfeat(mge/imperative): basic impl of python c extension for opdefrefactor(mge/distributed): sync interpreter for distribtued launcherfeat(mgb/gopt): add opt pass for fusing convolution and reformatfeat(dnn/cuda): add cutlass impls for fused convolution reformat operationfeat(dnn/cuda): add a cutlass impl for fusing convolution and dimshufflefeat(dnn/cuda): add impl for fusing warp perspective and dimshufflefix(mgb/core): fix memory management release cambricon var issuefix(mgb/jit): lower FuncOp to GPUFuncOp directlybuild(imperative): clean develop before build imperative and mgbrefactor(mge/virtualdep): remove virtualdep opdeffeat(mge/trace): tracing use id to set priorityfix(mgb/jit): add bind_shape feature to MLIRCompilerperf(imperative): add apply_on_physical_tensor for Elemwisefix(asan/thread): fix asan issuefix(dnn): forbid pooling window size smaller than paddingfix(imperative): add param(axis) for GetVarShapefix(mge/quantization): fix `qat.conv_bn` parameter issuetest(mge/quantization): qat module, quantized op and module
反馈
建议
客服 返回
顶部