Someone else can setting up blip and tagger missing linaqruf/lora trainer 1.5version method?

[gcoox]

any hero can settle this?

[upright2003]

Give it a try https://colab.research.google.com/drive/1SOoYachQhOKgbq3hSmZYu4z36FQdmktO?usp=sharing

[gcoox]

Give it a try https://colab.research.google.com/drive/1SOoYachQhOKgbq3hSmZYu4z36FQdmktO?usp=sharing

ok i have a try，but when i starting blip caption steps the execute cell come out this errors Traceback (most recent call last): File "/content/kohya-trainer/finetune/make_captions.py", line 20, in from blip.blip import blip_decoder, is_url File "/content/kohya-trainer/finetune/blip/blip.py", line 13, in from blip.vit import VisionTransformer, interpolate_pos_embed File "/content/kohya-trainer/finetune/blip/vit.py", line 16, in from timm.models.vision_transformer import _cfg, PatchEmbed ModuleNotFoundError: No module named 'timm'

[upright2003]

I did not use the 4.2.1 BLIP Captioning step, How do you use the training steps? My steps are 1.1> 2.2>2.3>3.1> 4.3>4.4> 5.1>5.2>5.3>5.4>5.5 What object are you training on? Custom Model?

[gcoox]

I did not use the 4.2.1 BLIP Captioning step, How do you use the training steps? My steps are 1.1> 2.2>2.3>3.1> 4.3>4.4> 5.1>5.2>5.3>5.4>5.5 What object are you training on? Custom Model?

yes,i have using custom model to change the model type,is that the reason of error?

[upright2003]

Not sure about your situation, or can you post a notebook link? Training data? custom model link, and steps, or operation video?

[hollowstrawberry]

The taggers should be fixed now. Continue discussion in #216

[gcoox]

Not sure about your situation, or can you post a notebook link? Training data? custom model link, and steps, or operation video?

nvm,host has been fix it,but anyway tq for you help with it ,very very much

[gcoox]

The taggers should be fixed now. Continue discussion in #216

Okay tq >.<