I have some remapped issues

[GrainSack]

Below script is my terminal error for running

python precompute_noises_and_conditionings.py \ --config ./config/parameter_estimation.yaml \ --inversion_subfolder noise \ --token_subfolder tokens \ --triplet_file triplets.csv \ --data_path ./dataset/data/

Model loaded
/hdd1/kss/home/miniconda3/envs/dia_env/lib/python3.8/site-packages/torchvision/transforms/functional_pil.py:42: DeprecationWarning: FLIP_LEFT_RIGHT is deprecated and will be removed in Pillow 10 (2023-07-01). Use Transpose.FLIP_LEFT_RIGHT instead.
  return img.transpose(Image.FLIP_LEFT_RIGHT)
ftfy or spacy is not installed using BERT BasicTokenizer instead of ftfy.
Selected timesteps: tensor([4, 0, 5, 2, 3, 6, 7, 1])
  0%|                                                                                                                                                                                        | 0/10 [00:00<?, ?it/s]
Traceback (most recent call last):
  File "estimate_CLIP_features.py", line 65, in <module>
    output = invertor.perform_cond_inversion_individual_timesteps(file_path, None, optimize_tokens=True)
  File "/hdd1/kss/home/DIA/ddim_invertor.py", line 275, in perform_cond_inversion_individual_timesteps
    noise_prediction = self.ddim_sampler.model.apply_model(noisy_samples, steps_in, cond_init.expand(self.config.conditioning_optimization.batch_size, -1 , -1))
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/models/diffusion/ddpm.py", line 987, in apply_model
    x_recon = self.model(x_noisy, t, **cond)
  File "/hdd1/kss/home/miniconda3/envs/dia_env/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/models/diffusion/ddpm.py", line 1410, in forward
    out = self.diffusion_model(x, t, context=cc)
  File "/hdd1/kss/home/miniconda3/envs/dia_env/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/modules/diffusionmodules/openaimodel.py", line 732, in forward
    h = module(h, emb, context)
  File "/hdd1/kss/home/miniconda3/envs/dia_env/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input,  #**kwargs)
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/modules/diffusionmodules/openaimodel.py", line 85, in forward
    x = layer(x, context)
  File "/hdd1/kss/home/miniconda3/envs/dia_env/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/modules/attention.py", line 258, in forward
    x = block(x, context=context)
  File "/hdd1/kss/home/miniconda3/envs/dia_env/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/modules/attention.py", line 209, in forward
    return checkpoint(self._forward, (x, context), self.parameters(), self.checkpoint)
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/modules/diffusionmodules/util.py", line 114, in checkpoint
    return CheckpointFunction.apply(func, len(inputs), *args)
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/modules/diffusionmodules/util.py", line 127, in forward
    output_tensors = ctx.run_function(*ctx.input_tensors)
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/modules/attention.py", line 213, in _forward
    x = self.attn2(self.norm2(x), context=context) + x
  File "/hdd1/kss/home/miniconda3/envs/dia_env/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/hdd1/kss/home/DIA/stable-diffusion/ldm/modules/attention.py", line 180, in forward
    sim = einsum('b i d, b j d -> b i j', q, k) * self.scale
  File "/hdd1/kss/home/miniconda3/envs/dia_env/lib/python3.8/site-packages/torch/functional.py", line 330, in einsum
    return _VF.einsum(equation, operands)  # type: ignore[attr-defined]
RuntimeError: einsum(): operands do not broadcast with remapped shapes [original->remapped]: [64, 4096, 40]->[64, 4096, 1, 40] [8, 77, 40]->[8, 1, 77, 40]

---

And did't found Token inversion at last

Traceback (most recent call last):
  File "estimate_input_noise.py", line 70, in <module>
    outputs = invertor.perform_inversion(file_name, cond = None, init_noise_init = None, loss_weights= {'latents': 1. , 'pixels':1.} )
  File "/hdd1/kss/home/DIA/ddim_invertor.py", line 93, in perform_inversion
    assert cond_out is not None, 'Token inversion was not found...'
AssertionError: Token inversion was not found...

---

My torch version is same as repo(1.11.0) And nvidia-smi cuda version in 12.0

[subrtadel]

Did you change the batch size? There was a bug, I believe it is fixed now.