menghao666 / hdr Goto Github PK

Dear author,
I have encountered an error when I try to call the segmentation network. I sincerely need your help, thank you!

import sys
sys.path.insert(0,'/mnt/workspace/project/2022AW/09/HDR/')
import numpy as np
import cv2
from utils.preprocessing import process_bbox, generate_patch_image
from PIL import Image
from mmseg.apis import inference_segmentor, init_segmentor
import matplotlib.pyplot as plt
%matplotlib inline


# segmentor
seg_cfg_file = "./configs/segformer/segformer_mit-b5_256x256_interhand_1101.py"
seg_model = init_segmentor(
    config=seg_cfg_file,
    checkpoint='./demo_work_dirs/Interhand_seg/iter_237500.pth',
)

Error

/mnt/workspace/project/2022AW/09/HDR/mmseg/models/backbones/mit.py:311: UserWarning: DeprecationWarning: pretrained is a deprecated, please use "init_cfg" instead
  warnings.warn('DeprecationWarning: pretrained is a deprecated, '
Use load_from_local loader
---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-6-db1a521dfe39> in <module>
      3 seg_model = init_segmentor(
      4     config=seg_cfg_file,
----> 5     checkpoint='./demo_work_dirs/Interhand_seg/iter_237500.pth',
      6 )

/mnt/workspace/project/2022AW/09/HDR/mmseg/apis/inference.py in init_segmentor(config, checkpoint, device)
     33     model = build_segmentor(config.model, test_cfg=config.get('test_cfg'))
     34     if checkpoint is not None:
---> 35         checkpoint = load_checkpoint(model, checkpoint, map_location='cpu')
     36         model.CLASSES = checkpoint['meta']['CLASSES']
     37         model.PALETTE = checkpoint['meta']['PALETTE']

/home/pai/lib/python3.6/site-packages/mmcv/runner/checkpoint.py in load_checkpoint(model, filename, map_location, strict, logger, revise_keys)
    525         dict or OrderedDict: The loaded checkpoint.
    526     """
--> 527     checkpoint = _load_checkpoint(filename, map_location, logger)
    528     # OrderedDict is a subclass of dict
    529     if not isinstance(checkpoint, dict):

/home/pai/lib/python3.6/site-packages/mmcv/runner/checkpoint.py in _load_checkpoint(filename, map_location, logger)
    464            information, which depends on the checkpoint.
    465     """
--> 466     return CheckpointLoader.load_checkpoint(filename, map_location, logger)
    467 
    468 

/home/pai/lib/python3.6/site-packages/mmcv/runner/checkpoint.py in load_checkpoint(cls, filename, map_location, logger)
    242         class_name = checkpoint_loader.__name__
    243         mmcv.print_log(f'Use {class_name} loader', logger)
--> 244         return checkpoint_loader(filename, map_location)
    245 
    246 

/home/pai/lib/python3.6/site-packages/mmcv/runner/checkpoint.py in load_from_local(filename, map_location)
    259     if not osp.isfile(filename):
    260         raise IOError(f'{filename} is not a checkpoint file')
--> 261     checkpoint = torch.load(filename, map_location=map_location)
    262     return checkpoint
    263 

/home/pai/lib/python3.6/site-packages/torch/serialization.py in load(f, map_location, pickle_module, **pickle_load_args)
    598             # reset back to the original position.
    599             orig_position = opened_file.tell()
--> 600             with _open_zipfile_reader(opened_file) as opened_zipfile:
    601                 if _is_torchscript_zip(opened_zipfile):
    602                     warnings.warn("'torch.load' received a zip file that looks like a TorchScript archive"

/home/pai/lib/python3.6/site-packages/torch/serialization.py in __init__(self, name_or_buffer)
    240 class _open_zipfile_reader(_opener):
    241     def __init__(self, name_or_buffer) -> None:
--> 242         super(_open_zipfile_reader, self).__init__(torch._C.PyTorchFileReader(name_or_buffer))
    243 
    244 

RuntimeError: PytorchStreamReader failed reading zip archive: invalid header or archive is corrupted

I am replacing Tzionas_ Dataset-02-1-rgb-100.png took a picture of my own crossed hands, but the effect was not good after running. May I ask if it is related to Tzionas_ Dataset-02-1-joints_ 2D_ Is GT-100.txt related to this file? Please provide guidance.

Hi,

Thank you for producing such amazing work for the community. Can we get the training and testing scripts soon?

Thank you!

I wonder whether segmentation module was trained on the whole interhand2.6m dataset including both one hand and two hands. I run this demo using data of interhand2.6m and get some results below. The segmentation seems nice while the prediction should be None about right hand.

Thank you for your great work and look forward to the release of your training code!

Whether the pre-trained model is the model of the results in the paper

Thank you for open sourcing the code, it's a really interesting job.

Thanks for open sourcing Google Drive pretrained models.

Does this pre-training model correspond to the experimental results in the paper? Or just a demo model, not the SOTA model in the paper?

Thanks!

Hello, how do I perform a real-time test?

OSError： ./dome_work_dirs/Interhand_seg/iter_237500.pth is not a checkpoint file

Very interesting job. Can you open source the code for generating amodal InterHand (AIH) Dataset? I will be very grateful!

Thank you for your interesting research.

I implemented the demo file you provided about the interhand 2.6m dataset, and the result was not good as below.

I wonder if you can tell me the path of the image to see the result well when I put in the image.

Hi,I'm a student and very interested in your "3D Interacting Hand Pose Estimation by Hand De-occlusion and Removal" paper.

i'm getting error on running demo.py

(hdr_hope) C:\HDR>python demo/demo.py
load checkpoint from local path: ./demo_work_dirs/Interhand_seg/iter_237500.pth => loading checkpoint './demo_work_dirs/TDR_fintune/checkpoints\ckpt_iter_138000.pth.tar' => loading checkpoint './demo_work_dirs/All_train_SingleRightHand\checkpoints\ckpt_iter_261000.pth.tar' img.shape= torch.Size([1, 3, 256, 256]) Traceback (most recent call last): File "demo/demo.py", line 163, in <module> seg_result = inference_segmentor(seg_model, rgb.cuda()) File "C:\HDR\mmseg\apis\inference.py", line 168, in inference_segmentor seg_logits = model.encode_decode(img, None) File "C:\HDR\mmseg\models\segmentors\encoder_decoder.py", line 77, in encode_decode x = self.extract_feat(img) File "C:\HDR\mmseg\models\segmentors\encoder_decoder.py", line 73, in extract_feat return x UnboundLocalError: local variable 'x' referenced before assignment

i think maybe the problem is in "extract_feat" so i also post it on here

def extract_feat(self, img): """Extract features from images.""" # print(self.backbone(img)) # print(img) try: x = self.backbone(img) except : print("img.shape=", img.shape) if self.with_neck: x = self.neck(self.backbone(img))

Thank you for the good research.

According to the paper, you said you pre-train SegFormer, but did you use only two hand images except one hand image on the Interhand 2.6M dataset?

Thank you for your perfect work. I would like to test on my own image, so how could I get the 2D GT (.txt) for my own image?

Thx~