Daqiu Shi
Daqiu Shi
如题 最小复现代码: ``` >>> oneflow.__version__ '0.8.0.dev20220705+cu112' ``` ``` import torch import oneflow as flow x_torch = torch.randn(5,5) x_flow = flow.tensor(x_torch.numpy()) # BUG: inplace x_torch[:,:2] += x_torch[:,4:] x_flow[:,:2] += x_flow[:,4:] #...
TODO LIST: - [x] coco_dataset预处理 - [x] modeling - [x] trainer - [x] torch权重加载测试(已对齐) - [x] eager global tensor parallel [evaluation结果对齐](https://github.com/Oneflow-Inc/libai/issues/288) - [x] 更libai的transformer实现,目前版本参考很多torch.nn.MultiHeadAttention - [ ] 推进训练 oneflow bug和不支持算子记录...
DETR结果对齐实验记录
Eager global 模型并行 参数对齐:https://github.com/facebookresearch/detr 问题排查TODO LIST: - [x] 继承libai attention实现的MultiHeadAttention是否对齐torch.nn.MultiHeadAttention (已对齐) - [x] PyTorch权重加载正确性 - [x] bakebone权重加载 - [x] libai-like transformer完善 - [x] [对某些input shape导致loss.backward报错"F20220602 14:17:25.050042 15603 shape.cpp:187] Check failed:...
该pr完善了`libai/optim/build.py/get_default_optimizer_params`中自定义不同层lr不方便的问题。
**该issue记录ResNeSt速度问题跟踪进展。** 单卡 OneFlow resnest50 time: 118.3ms (= 11832.9ms / 100, input_shape=[16, 3, 224, 224]) PyTorch resnest50 time: 79.6ms (= 7961.8ms / 100, input_shape=[16, 3, 224, 224]) ❌ Relative speed: 0.67...