No matter what I do, I can't get to use the SDXL ones that require the 1.5 clipvision

[PurpleBlueAloeVera]

Error occurred when executing IPAdapter:

Error(s) in loading state_dict for Resampler:
    size mismatch for latents: copying a param with shape torch.Size([1, 16, 1280]) from checkpoint, the shape in current model is torch.Size([1, 16, 2048]).
    size mismatch for proj_in.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([2048, 1280]).
    size mismatch for proj_in.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for proj_out.weight: copying a param with shape torch.Size([2048, 1280]) from checkpoint, the shape in current model is torch.Size([2048, 2048]).
    size mismatch for layers.0.0.norm1.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.0.0.norm1.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.0.0.norm2.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.0.0.norm2.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.0.0.to_q.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([768, 2048]).
    size mismatch for layers.0.0.to_kv.weight: copying a param with shape torch.Size([2560, 1280]) from checkpoint, the shape in current model is torch.Size([1536, 2048]).
    size mismatch for layers.0.0.to_out.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([2048, 768]).
    size mismatch for layers.0.1.0.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.0.1.0.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.0.1.1.weight: copying a param with shape torch.Size([5120, 1280]) from checkpoint, the shape in current model is torch.Size([8192, 2048]).
    size mismatch for layers.0.1.3.weight: copying a param with shape torch.Size([1280, 5120]) from checkpoint, the shape in current model is torch.Size([2048, 8192]).
    size mismatch for layers.1.0.norm1.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.1.0.norm1.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.1.0.norm2.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.1.0.norm2.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.1.0.to_q.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([768, 2048]).
    size mismatch for layers.1.0.to_kv.weight: copying a param with shape torch.Size([2560, 1280]) from checkpoint, the shape in current model is torch.Size([1536, 2048]).
    size mismatch for layers.1.0.to_out.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([2048, 768]).
    size mismatch for layers.1.1.0.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.1.1.0.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.1.1.1.weight: copying a param with shape torch.Size([5120, 1280]) from checkpoint, the shape in current model is torch.Size([8192, 2048]).
    size mismatch for layers.1.1.3.weight: copying a param with shape torch.Size([1280, 5120]) from checkpoint, the shape in current model is torch.Size([2048, 8192]).
    size mismatch for layers.2.0.norm1.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.2.0.norm1.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.2.0.norm2.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.2.0.norm2.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.2.0.to_q.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([768, 2048]).
    size mismatch for layers.2.0.to_kv.weight: copying a param with shape torch.Size([2560, 1280]) from checkpoint, the shape in current model is torch.Size([1536, 2048]).
    size mismatch for layers.2.0.to_out.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([2048, 768]).
    size mismatch for layers.2.1.0.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.2.1.0.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.2.1.1.weight: copying a param with shape torch.Size([5120, 1280]) from checkpoint, the shape in current model is torch.Size([8192, 2048]).
    size mismatch for layers.2.1.3.weight: copying a param with shape torch.Size([1280, 5120]) from checkpoint, the shape in current model is torch.Size([2048, 8192]).
    size mismatch for layers.3.0.norm1.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.3.0.norm1.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.3.0.norm2.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.3.0.norm2.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.3.0.to_q.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([768, 2048]).
    size mismatch for layers.3.0.to_kv.weight: copying a param with shape torch.Size([2560, 1280]) from checkpoint, the shape in current model is torch.Size([1536, 2048]).
    size mismatch for layers.3.0.to_out.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([2048, 768]).
    size mismatch for layers.3.1.0.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.3.1.0.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([2048]).
    size mismatch for layers.3.1.1.weight: copying a param with shape torch.Size([5120, 1280]) from checkpoint, the shape in current model is torch.Size([8192, 2048]).
    size mismatch for layers.3.1.3.weight: copying a param with shape torch.Size([1280, 5120]) from checkpoint, the shape in current model is torch.Size([2048, 8192]).

  File "F:\AI_repos\Stable_Comfy_UI\ComfyUI_windows_portable\ComfyUI\execution.py", line 152, in recursive_execute
    output_data, output_ui = get_output_data(obj, input_data_all)
  File "F:\AI_repos\Stable_Comfy_UI\ComfyUI_windows_portable\ComfyUI\execution.py", line 82, in get_output_data
    return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
  File "F:\AI_repos\Stable_Comfy_UI\ComfyUI_windows_portable\ComfyUI\execution.py", line 75, in map_node_over_list
    results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
  File "F:\AI_repos\Stable_Comfy_UI\ComfyUI_windows_portable\ComfyUI\custom_nodes\IPAdapter-ComfyUI\ip_adapter.py", line 163, in adapter
    self.ipadapter = IPAdapterModel(
  File "F:\AI_repos\Stable_Comfy_UI\ComfyUI_windows_portable\ComfyUI\custom_nodes\IPAdapter-ComfyUI\ip_adapter.py", line 103, in __init__
    self.image_proj_model.load_state_dict(state_dict["image_proj"])
  File "F:\AI_repos\Stable_Comfy_UI\ComfyUI_windows_portable\python_embeded\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(`

I always get this error, I even tried the workflow that you have posted for the SDXL model, but nothing makes this work it's driving me nuts. All the other ipadapters work fine. But the ones with vit-h that require the 1.5 image encoder, always print me this error.

Please help, I'd appreciate it a lot. Thanks in advance. Also, you're doing fantastic work with Ipadapter. Thank you for your contribution

[cubiq]

please be sure to have the latest version installed. then close comfy reopen and refresh your browser. if you had an old version of ipadapter installed you may need to delete the ipadapter node and recreate the workflow

Minimal working workflow:

[suede299]

I thought I'd add some detail to the question. When I use the latest model ip-adapter-plus_sdxl_vit-h.bin. I get the same error regardless of which clipvision model I pair it with. But with the two previous sdxl models, both work fine.

[cubiq]

that should happen only if you haven't updated the ipadapter extension to the latest version. if you use the manager remember that this is the comfyui_ipadapter_plus extension.

[GamingDaveUk]

I had the same issue, tried the work flow above and it wouldnt load into comfy, replicated it and found the issue. in load clip vision i used the sdxl model thats provided in the github instructions... looking at your image your using 1.5 one. tried it with a sdxl model and using the 1.5 clip vision model and it works.

On a ide note so many things coming out that are still using the out dated, low resolution 1.,5 models. I totally got that with 2.0 coming so crippled but sdxl has proven to be better in every way its insane so many are still focused on 1.5 lol

[cubiq]

@GamingDaveUk all *vit-h models (SDXL and otherwise) are trained with smaller images so they need the 1.5 clipvision. Those are images only used for the clip vision encoder. That doesn't impact the final image quality that is dictated by the main checkpoint.

[GamingDaveUk]

Hopefully soon these things will start being more sdxl centric. Very handy node though, probably wont replace LORA creation for me but certainly help for quick projects

[cubiq]

closing this for inactivity, feel free to reopen if the problem persists