艾梦

error of sparse_categorical_crossentropy when using theano backend

5

It's totally no problem when using tensorflow backend. Now I test the theano. When running train_model of tutorial.ipynb,we get 1d~2d tensor but not Tensortype(float32,3D) error from T.nnet.softmax() of K.sparse_categorical_crossentropy ```...

Without box.numpy(), nms raise `TypeError: operand 'Tensor' does not support ufuncs (__array_ufunc__=None)`.

Fix HQQ model param device transfer issue

1

# What does this PR do? Fixes #36254 And this PR makes it possible to infer bnb-4bit Flux with hqq-4bit quantized T5 model in the diffusers pipeline. I uploaded models...