Inferrence errors, help!

[RYHSmmc]

Traceback (most recent call last):
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/demo/demo_lazy.py", line 135, in
demo = VisualizationDemo(cfg, args=args)
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/demo/predictor_lazy.py", line 177, in init
self.predictor = DefaultPredictor(cfg)
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/ape/engine/defaults.py", line 56, in init
self.model = instantiate(cfg.model)
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/gitFile/detectron2-main/detectron2/config/instantiate.py", line 67, in instantiate
cfg = {k: instantiate(v) for k, v in cfg.items()}
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/gitFile/detectron2-main/detectron2/config/instantiate.py", line 67, in
cfg = {k: instantiate(v) for k, v in cfg.items()}
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/gitFile/detectron2-main/detectron2/config/instantiate.py", line 83, in instantiate
return cls(**cfg)
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/ape/modeling/text/clip_wrapper_eva01.py", line 19, in init
self.net, _ = build_eva_model_and_transforms(clip_model, pretrained=cache_dir)
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/ape/modeling/text/eva01_clip/eva_clip.py", line 165, in build_eva_model_and_transforms
model = create_model(
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/ape/modeling/text/eva01_clip/eva_clip.py", line 110, in create_model
load_checkpoint(model, pretrained)
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/ape/modeling/text/eva01_clip/eva_clip.py", line 82, in load_checkpoint
state_dict = load_state_dict(checkpoint_path, model_key=model_key)
File "/cpfs/user/mingcanma/workspace/code/openseg/APE-main/ape/modeling/text/eva01_clip/eva_clip.py", line 69, in load_state_dict
checkpoint = torch.load(checkpoint_path, map_location=map_location)
File "/opt/conda/lib/python3.10/site-packages/torch/serialization.py", line 771, in load
with _open_file_like(f, 'rb') as opened_file:
File "/opt/conda/lib/python3.10/site-packages/torch/serialization.py", line 270, in _open_file_like
return _open_file(name_or_buffer, mode)
File "/opt/conda/lib/python3.10/site-packages/torch/serialization.py", line 251, in init
super(_open_file, self).init(open(name, mode))
FileNotFoundError: [Errno 2] No such file or directory: 'models/BAAI/EVA/eva_clip_psz14.pt'
munmap_chunk(): invalid pointer
Aborted (core dumped)

[shenyunhang]

It seems that the train.init_checkpoint is not set in the configure, and the model uses the default path models/BAAI/EVA/eva_clip_psz14.pt, which is not found.

Please download the pre-trained model and add the path to train.init_checkpoint .

[RYHSmmc]

python demo/demo_lazy.py
--config-file configs/LVISCOCOCOCOSTUFF_O365_OID_VGR_SA1B_REFCOCO_GQA_PhraseCut_Flickr30k/ape_deta/ape_deta_vitl_eva02_clip_vlf_lsj1024_cp_16x4_1080k.py
--input image1.jpg image2.jpg image3.jpg
--output ./out
--confidence-threshold 0.1
--text-prompt 'person,car,chess piece of horse head'
--with-box
--with-mask
--with-sseg
--opts
train.init_checkpoint=models/model_final.pth
model.model_vision.select_box_nums_for_evaluation=500
model.model_vision.text_feature_bank_reset=True
..................
I have set train.init_checkpoint to the path of APE-D. But encounter the same error: RuntimeError: Pretrained weights (models/QuanSun/EVA-CLIP/EVA02_CLIP_E_psz14_plus_s9B.pt) not found for model EVA02-CLIP-bigE-14-plus.Available pretrained tags (['eva', 'eva02', 'eva_clip', 'eva02_clip'].

[shenyunhang]

Thank you for the feedback.
It tries to load the text model from EVA02_CLIP_E_psz14_plus_s9B.pt.
You can set model.model_language.cache_dir="" to disable it, and the text model will be initialized from APE-D.
We will update the README.