chokevin8

Question on how to train a stable diffusion image variation model

11

Hello, I'm trying to come up with an image variation model, where the input would be an unstained pathology image and the model would output a stained version of the...

I was wondering if ControlNet could be used to virtually stain an unstained image for my research purposes. For those not aware of how unstained and stained H&E images look...

Pix2Pix Input Inference Image Looks Different From the Original Inference Image

2

When running inference on Pix2Pix, in the results folder, you get real_A, real_B, and fake_B. The problem is that all of these images look different compared to the images that...

Hi, I am aware that the authors utilized random cropping, random horizontal flipping, label-smoothing regularization, mixup, and random erasing as data augmentations. However, there hasn't been an ablation study on...

Using all three color spaces

4

So I calculated my dataset statistics by using dataset_statistics.py for all three color spaces (RGB, HSV, HED), and then created a yaml file that looked exactly like [this](https://github.com/yiqings/RandStainNA/blob/master/segmentation/norm_jitter/MoNuSeg_LAB%2BHED%2BHSV_randomTrue_n0.yaml). When trying...

Like the title suggests, can we use simba and ignore the labels (just pass all image label as zero) and then train via self-distillation (SSL)? Would you expect the results...

If anyone has done bf16 training on Vim, can you let me know your environment for CUDA version, pytorch, causal_conv1d, and mamba_ssm? I cannot seem to get the environment right...

I wonder if anyone has tried an implementation of FSDP, it would help train larger Vim models for larger datasets since FSDP will shard the models and its parameters across...