neve33210

very slow

2

I'm on RTX4090, to drag a face generated by default config. One step will take about 1s, is this normal or it's just what this model is now?

unselected sdxl styles panel by default

2

### Is there an existing issue for this? - [X] I have searched the existing issues and checked the recent builds/commits ### What would your feature do ? As the...

SwinV2 for downstream

1

### Branch main branch (mmpretrain version) ### Describe the bug I tried to adapt swinv2 in mmpretrain to dino in mmdet. The config is swinv2 tiny and dino and I...

### What's the feature? I try to train biggan in custom datasets. It turns out to be buggy. Besides some problems about modeling which I fixed, the main problem is...

I make a distilled faster rcnn r50 from r101 with PKD. Now the result weight is all arch about pkd. How can I extract the weight so that the weight...

yolo format unsupported

1

import numpy as np import albumentations as A album=A.Compose([ A.Resize(height=64, width=64) , A.RandomScale(p=1.0, scale_limit=(-0.5, 0.5)) ],bbox_params=A.BboxParams(format="yolo", label_fields=["bboxes"])) img=np.random.randint(112, 128, [1280, 1280, 3], dtype=np.uint8) boxes = [[0.1, 0.1, 0.01, 0.01],[0.1, 0.1,...

[Bug]: DAT upscaler in extra tab is not working

3

### Checklist - [ ] The issue exists after disabling all extensions - [ ] The issue exists on a clean installation of webui - [ ] The issue is...

I'm doing segmentation on very large images, which produce big flop even with your b0 scale. I decide to scale backbone to smaller one. IMO segformer can go smaller for...

where is lora_pti???

2

As readme says, export MODEL_NAME="runwayml/stable-diffusion-v1-5" export INSTANCE_DIR="./data/data_disney" export OUTPUT_DIR="./exps/output_dsn" lora_pti \ --pretrained_model_name_or_path=$MODEL_NAME \ --instance_data_dir=$INSTANCE_DIR \ --output_dir=$OUTPUT_DIR \ --train_text_encoder \ --resolution=512 \ --train_batch_size=1 \ --gradient_accumulation_steps=4 \ --scale_lr \ --learning_rate_unet=1e-4 \ --learning_rate_text=1e-5...

mm_grounding_dino num_class=256?

9

configs\mm_grounding_dino\grounding_dino_swin-t_pretrain_obj365.py, in this cfg, the num_classes in dino head is 256, and then this is inherited by all cfgs. why is 256 for 365 classes in object365? How does it...