divamgupta
Lora merge error: model.diffusion.model.input.blocks.1.0
Full error in 2.5.1. merging a 1.5 model and LORA:
Error : ('model.diffusion.model.input.blocks.1.0', 'lora_unet_down_blocks_O_resnets_O_convi', 'lora_unet_down_blocks_O_resnets_0_conv1.alpha', ['lora', 'unet', 'down' 'blocks", 'O', 'resnets'", 'O', 'conv1'])
Which models are tho trying to merge?
Does importing the LORA and using it directly without merging work for you?
On Sun, Feb 4, 2024 at 4:25 PM Dax Monson @.***> wrote:
Full error in 2.5.1. merging a 1.5 model and LORA:
Error : ('model.diffusion.model.input.blocks.1.0', 'lora_unet_down_blocks_O_resnets_O_convi', 'lora_unet_down_blocks_O_resnets_0_conv1.alpha', ['lora', 'unet', 'down' 'blocks", 'O', 'resnets'", 'O', 'conv1'])
— Reply to this email directly, view it on GitHub https://github.com/divamgupta/diffusionbee-stable-diffusion-ui/issues/505, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAONR5LBCG3HCQK4LDJC7RTYR74FRAVCNFSM6AAAAABCZE5XT6VHI2DSMVQWIX3LMV43ASLTON2WKOZSGEYTOMZRGU4DCNA . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Trying SDXLrender_v2.0.safetensors with any 1.5 model I get that error. When I try uploading that model directly to the main page the LoRa doesn't show up. Just trying to get more realistic portraits for a project so maybe it's time to move onto XL. Feel free to close this, but if you get a chance or see something incompatible I'd be curious to know. Thanks for your work.
It will show up when you enable advanced options
On Sun, Feb 4, 2024 at 7:17 PM Dax Monson @.***> wrote:
Trying SDXLrender_v2.0.safetensors with any 1.5 model I get that error. When I try uploading that model directly to the main page the LoRa doesn't show up. Just trying to get more realistic portraits for a project so maybe it's time to move onto XL. Feel free to close this, but if you get a chance or see something incompatible I'd be curious to know. Thanks for your work.
— Reply to this email directly, view it on GitHub https://github.com/divamgupta/diffusionbee-stable-diffusion-ui/issues/505#issuecomment-1926007283, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAONR5M7TMXGYX3NZYOVL43YSAQKLAVCNFSM6AAAAABCZE5XT6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMRWGAYDOMRYGM . You are receiving this because you commented.Message ID: @.*** com>
That is working with other LoRA, they're importing to the models page and in advanced options where it wasn't before. It must have been something to do with that particular LoRA. Chinese, so maybe something lost in translation. Thanks again
Not sure if this is related, but I just tried to add the "Juggernaut Cinematic XL LoRA from Civit.ai and got this error.
I never understood how the Lora thing works on Diffusionbee. On Civit.ai you use images and a txt file in a zip folder but with diffusion bee you are told to upload a safetensors file and there is nothing that tells you how to make your zip folder with images and txt into a safetensors file.
