FJDEV

Results 10 comments of FJDEV

ValueError: operands could not be broadcast together with shapes (3,3) (0,0)

Hi, just check other issues reported last year and found it is same issue, but I still can;t figure it out, here is some output: print(delD_norm.min(),delD_norm.max()) 0.0 380.9160012391183 with np.errstate(divide='ignore',invalid='ignore'):...

Fine tune BLIP Image Captioning to custom dataset

> Hi @labenz, thanks for your question. > > 1. Yes it is possible to use variable prompt. > 2. The maximum number of tokens that BLIP accepts is the...

RuntimeError: expected scalar type Half but found Float

change to "fp16=False" in trainer.py, should work. training work, but inference break. May need help on this.

ValueError(f"Could not parse LLM output: `{llm_output}`")

Is there any solution to this? Thanks.

How to modify the training and sampling process to generate enhanced image, not mask?

Thank you so much, will try.

Remove Something From The Photo, 'NoneType' object has no attribute 'resize'

> Did you get my error? AttributeError: 'NoneType' object has no attribute 'resize' > > are you using other models than ImageCaptioning, ImageEditing and Text2Image? I have the exactly same...

Exception: TypeError('Operator() takes no arguments')

Hi, the system memory is good enough. I tried to define this first: `op_imaging = ImagingOperator(geometry, image, space_order, save=False)` then import those: ``` from devito import DevitoCheckpoint, CheckpointOperator from devito...

How to use the code for Inference?

+1, I was trying to use val.py, but no luck. May need author's help.

Read a large segy file

Thanks for your reply. You are right, I missed the eval_range part.

When refinement, does it finetune the model?

Ok I see. That's a really good paper too. Do you know if there is any open-source repository for that work? I couldn't see. Thanks.