Jiahao Zhao

MODEL.PRETRAINED_DETECTOR_CKPT set to /upload_causal_motif_sgdet and modify `last_checkpoint` to your path. https://github.com/KaihuaTang/Scene-Graph-Benchmark.pytorch/issues/7#issue-578052449

Yes, it's custom image. I think`visual_info.json` is for mapping bbox index in `custom_prediction.pytorch` with image file name. So I just changed this `__getitem__` function to not return index but image...

@nyj-ocean Have you solved the problem? I also only get `custom_prediction.pytorch` using `SGDet on Custom Images`. And another question is `torch.load('custom_prediction.pytorch')` return a list of `BoxList`, how to know which...

Without changing the source peft code, in my case, you can just remove the following line in finetuning.py to solve save `adapter_model.bin` twice and only get the last empty checkpoint...

> @clxyder, with the latest main branch, you can simply do `model = model.merge_and_unload()` to get the base model with lora weights merged into it. @pacman100 @younesbelkada Hi, thanks for...

@younesbelkada Cool! A big thanks to you! This explanation solved my confusion in a very simple and concrete way. Thanks for this beautiful diagram and math explanation! 🌹🌹🌹

Hi, @lxe. I'm new to peft. The model.merge_and_unload() method looks like magic to me. Could you give a quick introduction about model.merge_and_unload()？ In my view, LoRA adds new trainable parameters/layers...

@activezhao A hotfix would be modify the `is_linear` function to skip 'Rotary' layer. ``` def is_linear(module: nn.Module) -> bool: """Returns whether the module is a linear layer.""" return any([k in...

run `pip install scikit-fuzzy`

Build Precondition: 1. Install SWIG 2. Using python=2.7 3. Change setup.py `os.environ["CC"] = "g++-5 -std=c++11` to your version 4. Change setup.py `build_path` to your path And here you go: run...