oneflow icon indicating copy to clipboard operation
oneflow copied to clipboard

OneFlow is a deep learning framework designed to be user-friendly, scalable and efficient.

Results 591 oneflow issues
Sort by recently updated
recently updated
newest added

优化诸如形状 [a, 1, b] , perm [0, 2, 1] 的情形

enhancement
op

针对 UpsampleNearest2D 2X 情况进行优化 * 减少坐标换算的开销 * 优化访存指令的数量

enhancement
op
need-all-tests-even-fail

enhancement
op
need-all-tests-even-fail

打开 ONEFLOW_DEBUG_MODE=1 环境变量后, 每次后向计算都会生成 AutogradEngine 执行图到 log 目录下的 dot file。可以从图中看到后向执行的算子以及拓扑结构,方便算法人员和研发人员 debug 后向的问题。 ### autograd.grad ```python import oneflow as flow # autograd.grad a = flow.rand(2, 3).requires_grad_() b = flow.rand(2, 3).requires_grad_() c...

enhancement
feature
automerge
eager

pytorch version: 1.12.1+cu102 oneflow: ``` import torch a = torch.Tensor(133, 1, 15) b = torch.Tensor(133, 2, 1) idx = 0 pos = torch.tensor(0) a[:, idx, pos] = b[:, 1, idx]...

bug
community

- 重构 eye 相关的 initializer,使用 oneflow 自己的 kernel,加速初始化 - init.py涉及的算子已经全部使用oneflow 自己的 kernel,因此废弃之前的方案,删除 initializer_conf.proto 文件

按 https://github.com/Oneflow-Inc/OneTeam/issues/1737 文档中实现随机初始化算子的重构,保证数据的正确性和一致性。 涉及rand,randint,randn算子

some erros: 1. mlir-pdll -h error 2. cmake circular dependency. 3. error: ‘parseSourceString’ is not a member of ‘mlir’.