-
### Paper
Implement [MnasNet](https://arxiv.org/abs/1807.11626)
### TODOs
- [ ] develop the model in a scalable way (using `*Block` syntax and `nn` modules)
- [ ] draw the model on figma
- [ ] wr…
-
Thanks for you work! Have you reimplement mnasnet-ssd network?
-
What is the training schedule for MNASNet 1.0 that gives 73.5 top-1? I tried the mobilenet schedule but it not giving the same accuracy as mentioned in torchvision page.
On a side note, mnasnet pa…
-
在cifar10数据上做了个小测验， base_lr=0.1训练60000轮后，测试精度只有58.74%， 日志显示train loss过大，应该上训练未充分， base_lr=0.0001训练60000轮后，只有28.17%的精度。
`
I1013 11:38:29.757405 3189 caffe.cpp:330] Softmax1 = 1.43176 (* 1 = 1.43176 l…
lqian updated
6 years ago
-
I applied OpenCL Tuner to the acl_cl backend.
OpenCL Tuner finds the optimal local work size for GPU acceleration tuning parameters.
https://developer.arm.com/solutions/machine-learning-on-arm/devel…
intom updated
3 years ago
-
How to get pretrain model weight of Mnasnet-small, d1, d1-320?
-
I experiment your framework with different backbone. Surprisingly, the performance of resnet34 is quite close to resnet50. Renest34 have mAP of about 0.705 (and 0.893 AP@0.5) on 2017 coco keypoints fo…
-
I downloaded the `saved_model` format mnasnet model from https://storage.googleapis.com/mnasnet/checkpoints/mnasnet-a1.tar.gz. A text file in the folder claims the given `tflite` and `saved_model` mod…
-
Using channels first data format causes the script to fail because of a number of reasons
1. The pre-processing steps assumes the data will always be channels last (even after the data format is se…
-
this is my parameters
Namespace(batch_size=32, database_dir='/ssd_data/yanghang/palm_vein/POLYU/NIR_polyu_crop/', embedding_size=512, evaluate=False, learning_rate=0.01, logdir='./data_polyup/CSVFil…