AttributeError: module diffusers has no attribute ORTStableDiffusionXLPipeline. Did you mean: 'StableDiffusionXLPipeline about diffusers

[fabiorigano]

I see, you should maybe ask which opset (or ONNX, ONNXRuntime and pytorch versions) has been used to export the model. Also try a different execution provider (CPU, CUDA)

[fabiorigano]

hi @neonarc4
The "How to use" documentation is definitely wrong
ORTStableDiffusionXLPipeline is not a diffusers pipeline, but a special pipeline for models in the ONNX format and belongs to the optimum library (an extension of transformers and diffusers).

Install optimum and load the pipeline as follows:

python -m pip install optimum

from optimum.onnxruntime import ORTStableDiffusionXLPipeline
pipeline = ORTStableDiffusionXLPipeline.from_pretrained("greentree/SDXL-olive-optimized")

[neonarc4]

@fabiorigano
tried earlier just getting similar issue encounter

2024-05-12 22:48:45.277336: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2024-05-12 22:48:46.160987: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
WARNING:tensorflow:From Z:\software\python11\Lib\site-packages\tf_keras\src\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.

Traceback (most recent call last):
  File "X:\sad\practice\test\aii\neo.py", line 259, in <module>
    pipeline = ORTStableDiffusionXLPipeline.from_pretrained("greentree/SDXL-olive-optimized")
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "Z:\software\python11\Lib\site-packages\optimum\onnxruntime\modeling_ort.py", line 669, in from_pretrained
    return super().from_pretrained(
           ^^^^^^^^^^^^^^^^^^^^^^^^
  File "Z:\software\python11\Lib\site-packages\optimum\modeling_base.py", line 402, in from_pretrained
    return from_pretrained_method(
           ^^^^^^^^^^^^^^^^^^^^^^^
  File "Z:\software\python11\Lib\site-packages\optimum\onnxruntime\modeling_diffusion.py", line 337, in _from_pretrained    vae_decoder, text_encoder, unet, vae_encoder, text_encoder_2 = cls.load_model(
                                                                   ^^^^^^^^^^^^^^^
  File "Z:\software\python11\Lib\site-packages\optimum\onnxruntime\modeling_diffusion.py", line 214, in load_model
    vae_decoder = ORTModel.load_model(vae_decoder_path, provider, session_options, provider_options)
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "Z:\software\python11\Lib\site-packages\optimum\onnxruntime\modeling_ort.py", line 375, in load_model
    return ort.InferenceSession(
           ^^^^^^^^^^^^^^^^^^^^^
  File "Z:\software\python11\Lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py", line 419, in __init__
    self._create_inference_session(providers, provider_options, disabled_optimizers)
  File "Z:\software\python11\Lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py", line 483, in _create_inference_session
    sess.initialize_session(providers, provider_options, disabled_optimizers)
onnxruntime.capi.onnxruntime_pybind11_state.NotImplemented: [ONNXRuntimeError] : 9 : NOT_IMPLEMENTED : Could not find an implementation for GroupNorm(1) node with name 'GroupNorm_0'