Commits · dkebudi/lora-ease-dk

upping the limit to 1000

24cb3e6
verified

dkebudi commited on Aug 6

increased max images to 650

d49f159
verified

dkebudi commited on Aug 6

revert to original

8a0d5c0
verified

dkebudi commited on Aug 5

syntax error

bd3d196
verified

dkebudi commited on Aug 5

comment out error catcher

ccca96c
verified

dkebudi commited on Aug 5

generate a folder on the fly for prior preservation

cfc0981
verified

dkebudi commited on Aug 5

remove the Raise check for class_image_directory being empty

f4612c7
verified

dkebudi commited on Aug 5

revert changes for class_images

646dca6
verified

dkebudi commited on Aug 5

taking out -- if class_images: --

8ec4f7e
verified

dkebudi commited on Aug 5

class_image if statement moved to not block the creation of a new data path

7a4b48d
verified

dkebudi commited on Aug 5

Update app.py

f4f3601
verified

dkebudi commited on Aug 5

printing training data's pixel values with """print(f"{batch['pixel_values'].shape=}")"""

3c525b4
verified

dkebudi commited on Aug 1

save images to tensorboard by default

effc90e
verified

dkebudi commited on Jul 31

removed print statements

18e9ffa
verified

dkebudi commited on Jul 31

pipeline args all updated to have hard coded dimensions in validation and prediction

e8383f5
verified

dkebudi commited on Jul 31

passing in dimensions here as well: pipeline_args

439be08
verified

dkebudi commited on Jul 31

don't delete vae from memory

0ef0958
verified

dkebudi commited on Jul 31

more print statements

547f26b
verified

dkebudi commited on Jul 31

print for testing

5e3e4ba
verified

dkebudi commited on Jul 31

crop based on hard coded resolution: 1080 (h) x 1920 (w) effectively resulting in no crop

c954ee0
verified

dkebudi commited on Jul 31

comment out y1, x1 lines

7d5b9b0
verified

dkebudi commited on Jul 31

added huggingface-hub to packages for GPU container

b564f87
verified

dkebudi commited on Jul 31

removed version from packages

aa5a1dc
verified

dkebudi commited on Jul 31

remove version from git diffusers import for additional GPU container

d5ed76a
verified

dkebudi commited on Jul 31

Update requirements.txt

cde849d
verified

dkebudi commited on Jul 31

changing the diffusers clone

ead37ab
verified

dkebudi commited on Jul 31

adding git to diffusers

1b431a3
verified

dkebudi commited on Jul 31

removed peft version issue due to incompatibility with autotrain

2691c08
verified

dkebudi commited on Jul 31

new requirements for the training script

ba49b38
verified

dkebudi commited on Jul 31

adding pass statements

9ac1b66
verified

dkebudi commited on Jul 31

hard code 1080x1920

c30061b
verified

dkebudi commited on Jul 30

switch out the training script to custom training

e055b3b
verified

dkebudi commited on Jul 30

remove resizing and cropping (#6)

e374933
verified

dkebudi commited on Jul 30

Create train_dreambooth_lora_sdxl_advanced.py

8e61a1b
verified

dkebudi commited on Jul 30

center-crop clean up

7e63d03
verified

dkebudi commited on Jul 30

comment out center_crop as an input fro the UI (#5)

ced3f4a
verified

dkebudi commited on Jul 30

fix for crop_center variable (#4)

07a1fda
verified

dkebudi commited on Jul 30

center_crop variable add to everywhere (#3)

8352565
verified

dkebudi commited on Jul 30

increase max sample size, center_crop = True (#2)

4acbb3a
verified

dkebudi commited on Jul 30

A10G --> A100 (#1)

56b9662
verified

dkebudi commited on Jul 30

Update README.md

60c7a9e
verified

multimodalart HF staff commited on Jul 12

Update requirements.txt

c203ef4
verified

multimodalart HF staff commited on Jul 2

Update requirements_local.txt (#20)

8a225b2
verified

multimodalart HF staff