sheldon04 / cabm-pytorch

CABM: Content-Aware Bit Mapping for Single Image Super-Resolution Network with Large Input

🙋 Hello

🤺 About Me   Hi! I'm Sheldon，a programmer who enjoys poetry.   喜欢运动，喜欢诗词~

🏢 Education Experience Beijing Jiaotong University   📌 2019-09-01 —— Now B.E. & M.S

🤾‍♂️ Recent Interests 周琦加盟广东！ 🔥🔥🔥🔥🔥- 2023-08-19 八角笼中 🌟🌟🌟🌟🌟- 2023-07-09 肖申克的救赎 🌟🌟🌟🌟🌟 - 2023-02-07

📊 More TODO 💪 Using 🧠 Plan 🧰 Tools

cabm-pytorch's Issues

Questions about BitOPs

I am highly interested in innovative ideas and performance as well.
However, I have a few questions regarding BitOPs. It seems that the formula for calculating BitOPs in Table 2 of the CABM paper is
EDSR-CABM_x4 BitOPs (3.75G) = 114.23G x (5.80 / 32) x (5.80 / 32).
114.23G for 32bit model
5.80 / 32 for activation
5.80 / 32 for weight
However, the paper and code mention the use of fixed-precision for weights.
Could you please explain how the BitOPs are calculated?

How to run inference on single images?

student_weights

When I tried to use test**.bash, but found that this file (student_weights) is missing, may I ask where to get this file ?
Thanks.

where can I find EDSR-[4]-['div2k_valid'].json?

When I run the code, I encounter this below error:
with open(f"{args.model}-[4]-['div2k_valid'].json", "r") as f:
FileNotFoundError: [Errno 2] No such file or directory: "EDSR-[4]-['div2k_valid'].json"
I'm looking forward for your reply.

question for student network

It's a nice work! Can I just use the student and teacher networks edsr_×4_pams_w8a8.pt which provided in CADyQ from Google Cloud to train the edsrbaseline_cabm_simple?

Can it be applied to other SR networks?

Can it be applied to other networks besides the network provided in the paper? What part of the code should I change?

I wonder if this will also apply to transformer models such as SwinIR.