JiahuiYu / slimmable_networks

Slimmable Networks, AutoSlim, and Beyond, ICLR 2019, and ICCV 2019

JiahuiYu/slimmable_networks Issues

how to run this code with cpu
Updated 2 years ago
model profiling
Closed 4 years ago3
How can I evaluate standalone smaller width model?
Closed 2 years ago1
the issues of fuction SlimableConv2D
Updated 2 years ago1
Some questions with the code.
Closed 3 years ago
No module named 'mmdet.ops.nms.gpu_nms'
Updated 3 years ago
Mobile Implementation
Updated 3 years ago
Slimmable Network using Mixed Precision
Updated 3 years ago
Slimming to lower width results in poor accuracy on ARM architectures
Updated 3 years ago
BN initialization for tracking performance during training
Closed 3 years ago
reproducing CIFAR10 results for AutoSlim
Updated 3 years ago8
AutoSlim search
Updated 3 years ago1
US-WDSR
Updated 3 years ago1
Size mismatch when loading us_mobilenet_v1_calibrated.pt
Closed 5 years ago2
top1 error = 0.99 when testing
Closed 4 years ago1
Size mismatch for bn running_mean and running_var, us_mobilenet_v1
Updated 4 years ago
why use slimmable_training instead of universally_slimmable_training: True in Autoslim
Closed 4 years ago7
Would you share Image Super Resolution example model for Universally Slimmable Network training?
Closed 4 years ago1
A Question
Closed 4 years ago1
memory increasing during testing
Closed 4 years ago1
Question about slimming
Closed 4 years ago1
Run the experiment with distributed mode
Closed 4 years ago1
width_mult = None when training us-Net
Closed 4 years ago3
Using Cifar-100 to train slimmable_networks
Closed 4 years ago1
Question about arbitrary width in a universally slimmable network.
Closed 4 years ago8
When will you release Greedy Slimming?
Closed 4 years ago1
channel configuration of auto-slim mobilenet-v2
Closed 4 years ago1
The accuracy decrease sharply after calibration, even on the given Model zoo
Closed 4 years ago1
Size mismatch when loading us_mobilenet_v2_calibrated.pt
Closed 4 years ago3
us.mobilenet_v2 cifar
Closed 4 years ago5
how can i load us_model at a certain ratio
Closed 4 years ago5
Calibration with non-uniform widths at test time
Closed 5 years ago6
Why bn layers do not have weight_decay?
Closed 4 years ago1
Is num_sample_training=2 enough for training ?
Closed 4 years ago3
How to retrain slim-MobilenetV2
Closed 5 years ago5
Good job!
Closed 5 years ago10
US-Net BatchNorm
Closed 5 years ago6
How to collect post statistics of BN
Closed 5 years ago8
USNet for Obeject Detection
Closed 5 years ago1
Memory Leak when Training US-Net
Closed 5 years ago1
Sliming Object detection models
Closed 5 years ago1
Training Speed, S-Net vs. US-Net
Closed 5 years ago1
Training Speed
Closed 5 years ago4
some questions about pertained model
Closed 5 years ago3
Exciting work! Some questions about universally slimmable networks
Closed 5 years ago3
How did you combine the hard label loss and the Distillation loss
Closed 5 years ago7
Reproduced Result
Closed 5 years ago1
weight decay = 0 for BN parameters
Closed 5 years ago5
is the shufflenet implemented v1 or v2? Thanks.
Closed 5 years ago1
Add code
Closed 5 years ago1