Could you please provide the training code?

[xfcxing]

Hi, thank you for sharing this great project! I’ve gone through the repository and found the inference code very helpful. However, I noticed that the training-related code (e.g., model training scripts, data preprocessing for training, loss functions, etc.) is not included in the repo.

Would it be possible for you to share the training code or give some guidance on how to train the model from scratch?

Any help would be greatly appreciated. Thanks again for your excellent work!

[littlespray]

Hey there,

Thank you for your interest in our work. I am sorry that the original training code has become quite messy after many iterations. I will work on cleaning it up as soon as possible, but it might take some time to make it presentable.

In the meantime, I can send you the raw code first. I hope it can offer a quick glimpse into our key modules.

Besides, I highly recommend refering to DOVER's training code. Our work builds directly on their codebase, and we didn’t modify the core training pipeline—so it might be clearer for your development.

[xfcxing]

Thank you so much for your prompt reply and kind offer to share the raw code! I truly appreciate your willingness to let me take a look at it. I'd be more than happy to receive the raw code you mentioned. Even in its current state, I'm confident it will provide valuable insights into the key modules of your work.   I'm also very familiar with Dr. Wu Haoning's DOVER training code. As you pointed out, since your work is built directly on their codebase without significant modifications to the core training pipeline, I've already studied DOVER's code in depth. This prior knowledge will surely help me better understand your work and make good use of the code you share.   No rush on cleaning up the code. I understand how time - consuming and meticulous that process can be. Whenever it's convenient for you, I'm looking forward to seeing the polished version as well.   Thank you again for your help and support. I'm excited to explore your code and learn from it.

…

---Original--- From: ***@***.***> Date: Wed, Jun 11, 2025 00:00 AM To: ***@***.***>; Cc: ***@***.******@***.***>; Subject: Re: [littlespray/VE-Bench] Could you please provide the trainingcode? (Issue #7) littlespray left a comment (littlespray/VE-Bench#7) Hey there, Thank you for your interest in our work. I am sorry that the original training code has become quite messy after many iterations. I will work on cleaning it up as soon as possible, but it might take some time to make it presentable. In the meantime, I can send you the raw code first. I hope it can offer a quick glimpse into our key modules. Besides, I highly recommend refering to DOVER's training code. Our work builds directly on their codebase, and we didn’t modify the core training pipeline—so it might be clearer for your development. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>