Yubin Wang

same problem here。

I have install mxnet by pip, the problem also occurs. System: Windows 7 Python: 2.7.13 miniconda nightly build version

@yajiedesign here are more details: ```shell python -dv hello-mxnet.py (import mxnet as mx) ``` ```shell # cleanup[2] stat # cleanup[2] cv2.multicalib # cleanup[2] numpy.lib.format # cleanup[2] numpy.testing.nosetester # cleanup[2] cryptography.hazmat.primitives.asymmetric.rsa...

yes, i have found that, the cudnn depthwise's weights shape mismatch with other methods. when using checkpoint to initialize(such as resume train, or test accuracy), mismatch shape error would occur

@bruinxiong any news?

@bruinxiong thanks. I am also doing some experiments on resnet-18 with batch size 256. 256 is worse than 512. and there is no theoretical relationship between batchsize and dataset size,...

@bruinxiong, I have trained a renset-18 using this code with batch-size 256, after 4 days training, the result is a little better(30.4% error rate) than 512 batchsize. You may try...

@bruinxiong lr decreases at 30, 60, 90, and at 98 remove the imgressive data augmentation.

@LiJunY `get_model_tokenizer` 函数里指定你训练后的位置给`model_id_or_path` 参数