cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
sitting
Voyager
Voyager
639 Views
Registered: ‎05-04-2014

Resnet 50 with custom datasets

Jump to solution

Hi,

We have custom datasets and want to use xilinx resnet50 directly, is it possible?

If not, do we need to train new resnet50 model with our custom datasets? Is there any tutorial?

 

 

Thanks

Sitting

 

 

0 Kudos
1 Solution

Accepted Solutions
zhijiexu
Moderator
Moderator
616 Views
Registered: ‎09-29-2020

Hi, @sitting 

You also can use other framework like pytorch and tensorflow, but different framework have different API it supported.

And you can train it in your own environment ,and quantified, compile in the docker environment 

Best regard,

Zhijie

View solution in original post

3 Replies
zhijiexu
Moderator
Moderator
626 Views
Registered: ‎09-29-2020

hi, @sitting 

Yes, you need to train new resnet50 model with your custom Datasets.

The related Tutorial you can refer to was showed below:

https://github.com/Xilinx/Vitis-Tutorials/blob/master/Machine_Learning/Design_Tutorials/04-Keras_GoogleNet_ResNet/README.md

Also, you can read ug1414 and pg338,

Best regard,

Zhijie 

-----------------------------------------------------------------------------

Don't forget to reply, give kudos and accept as a solution!

----------------------------------------------------------------------------

sitting
Voyager
Voyager
621 Views
Registered: ‎05-04-2014

Hi @zhijiexu ,

I need to prepare the origin resent50 model and use caffe-xilinx to train it?

 

Thanks

Sitting

 

Thanks

Sitting

0 Kudos
zhijiexu
Moderator
Moderator
617 Views
Registered: ‎09-29-2020

Hi, @sitting 

You also can use other framework like pytorch and tensorflow, but different framework have different API it supported.

And you can train it in your own environment ,and quantified, compile in the docker environment 

Best regard,

Zhijie

View solution in original post