For generating crops having fixed aspect ratio, you need to change the generate_crop function to generate a set of candidate crops having fixed aspect ratio. You can refer to the implementations in testGAIC_qualitative_customer.m.

[andriyrizhiy]

For generating crops having fixed aspect ratio, you need to change the generate_crop function to generate a set of candidate crops having fixed aspect ratio. You can refer to the implementations in testGAIC_qualitative_customer.m.

Originally posted by @HuiZeng in #10 (comment)

[andriyrizhiy]

How generate small number crops with specific aspect ration? Because in your implementations in aspect ration 1_1 i get 137602 crops and it is too much for 125 fps(

[HuiZeng]

You may not use it in the right way. Please read the pre-processing and post-processing of test_GAIC_qualitative_customer.m rather than simply call the generate_crop function. -------- 原始邮件 --------主题：Re: [HuiZeng/Grid-Anchor-based-Image-Cropping] For generating crops having fixed aspect ratio, you need to change the generate_crop function to generate a set of candidate crops having fixed aspect ratio. You can refer to the implementations in testGAIC_qualitative_customer.m. (#15)发件人：andriyrizhiy 收件人：HuiZeng/Grid-Anchor-based-Image-Cropping 抄送：Hui Zeng ,Mention how generate small number crops with specific aspect ration. Because in your implementations in aspect ration 1_1 i get 137602 crops and it is too much for 125 fps( —You are receiving this because you were mentioned.Reply to this email directly, view it on GitHub, or mute the thread.

[Quershi]

Hi @HuiZeng Can we smoothly run Python version code in Google Colab ?... Because we have not many tools as you mention in paper. I asking about GPU and RAM etc , ..
respect of your response..

[HuiZeng]

Hi @HuiZeng Can we smoothly run Python version code in Google Colab ?... Because we have not many tools as you mention in paper. I asking about GPU and RAM etc , ..
respect of your response..

Hi，we have not tested on Google Colab. But it should be ok. Our model actually does not consume much computational resource. You can try the latest pytorch code based on shufflenet or mobilenet. 4G memory is enough to run the code. Or you can test it on CPU.