invoke-ai
[bug]: Can't change model
Is there an existing issue for this?
- [X] I have searched the existing issues
OS
Windows
GPU
cuda
VRAM
24
What version did you experience this issue on?
2.3.5 post 2
What happened?
Got error when try to replace model
Model change requested: lofi_V21Inpaint Current VRAM usage: 2.17G Offloading lofi_V21 to CPU Loading lofi_V21Inpaint from C:/vlad/automatic/models/Stable-diffusion/lofi_V21Inpaint.safetensors Converting legacy checkpoint lofi_V21Inpaint into a diffusers model... | global_step key not found in model ** model lofi_V21Inpaint could not be loaded: Error(s) in loading state_dict for UNet2DConditionModel: size mismatch for conv_in.weight: copying a param with shape torch.Size([320, 9, 3, 3]) from checkpoint, the shape in current model is torch.Size([320, 4, 3, 3]). Traceback (most recent call last): File "C:\Users\derec\invokeai.venv\lib\site-packages\ldm\generate.py", line 964, in set_model model_data = cache.get_model(model_name) File "C:\Users\derec\invokeai.venv\lib\site-packages\ldm\invoke\model_manager.py", line 105, in get_model requested_model, width, height, hash = self._load_model(model_name) File "C:\Users\derec\invokeai.venv\lib\site-packages\ldm\invoke\model_manager.py", line 317, in _load_model model, width, height, model_hash = self._load_ckpt_model( File "C:\Users\derec\invokeai.venv\lib\site-packages\ldm\invoke\model_manager.py", line 376, in _load_ckpt_model pipeline = load_pipeline_from_original_stable_diffusion_ckpt( File "C:\Users\derec\invokeai.venv\lib\site-packages\ldm\invoke\ckpt_to_diffuser.py", line 1230, in load_pipeline_from_original_stable_diffusion_ckpt unet.load_state_dict(converted_unet_checkpoint) File "C:\Users\derec\invokeai.venv\lib\site-packages\torch\nn\modules\module.py", line 2041, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format( RuntimeError: Error(s) in loading state_dict for UNet2DConditionModel: size mismatch for conv_in.weight: copying a param with shape torch.Size([320, 9, 3, 3]) from checkpoint, the shape in current model is torch.Size([320, 4, 3, 3]).
** trying to reload previous model
Retrieving model lofi_V21 from system RAM cache Textual inversion triggers: Setting Sampler to k_lms (LMSDiscreteScheduler)
Screenshots
No response
Additional context
No response
Contact Details
No response
This is an inpainting model - when you add it via the model manager, make sure you select the 'inpainting' type.
