ValueError: Cannot assign to variable block_1_1_ds_conv1/filters:0 due to variable shape (1, 1, 64, 256) and value shape (64,) are incompatible

[AndreyStille]

Please help.

Cannot load weights using infer,py. Weights are taken from google drive
(lightweight_openpose_resnet50.npz)

import cv2
import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
from hyperpose import Config,Model,Dataset
from hyperpose.Dataset import imread_rgb_float,imwrite_rgb_float
Config.set_model_name("LightweightOpenpose")
Config.set_model_backbone(Config.BACKBONE.Resnet50)
Config.set_model_type(Config.MODEL.LightweightOpenpose)
config=Config.get_config()

#get and load model
model=Model.get_model(config)
weight_path=f"{config.model.model_dir}/newest_model.npz"
model.load_weights('/content/hyperpose/save_dir/openpose/model_dir/newest_model.npz')

[orestis-z]

Tried setting Config.set_model_backbone(Config.BACKBONE.Vggtiny) but the same error occurs

[Gyx-One]

Hello! @AndreyStille
Sorry to response so late.
The infer.py is reported to have several issues and thus deprecated now, please use python_demo.py (which is tested to make sure it can work with the existing model weights).
As for the shape mismatch issue, this has been fixed in the newest commit, please download the newest model weights here and they should work well with the python_demo.py