我使用的是Closed-set finetuning微调方法，在test中能正常输出
Traceback (most recent call last):
File "/home/gxdong/yolo/WeDetect-main/infer_wedetect.py", line 213, in
model.reparameterize(texts)
File "/home/gxdong/anaconda3/envs/openmmlab_env/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1729, in getattr
AttributeError: 'SimpleYOLOWorldDetector' object has no attribute 'reparameterize'. Did you mean: 'reparameterized'?
我是用的是tiny
python3 infer_wedetect.py
--config /home/gxdong/yolo/WeDetect-main/work_dirs/wedetect_tiny_all/20260407_065304/vis_data/config.py
--checkpoint /home/gxdong/yolo/WeDetect-main/work_dirs/wedetect_tiny_all/best_coco_bbox_mAP_epoch_20.pth
--image images/aaa.jpg
--text '大车,小车,非机动车,人'
--threshold 0.1
更换成命令 不报错 但是加载
python3 infer_wedetect.py
--config /home/gxdong/yolo/WeDetect-main/config/wedetect_tiny.py
--checkpoint /home/gxdong/yolo/WeDetect-main/work_dirs/wedetect_tiny_all/best_coco_bbox_mAP_epoch_20.pth
--image images/aaa.jpg
--text '大车,小车,非机动车,人'
--threshold 0.1
The model and loaded state dict do not match exactly
missing keys in source state_dict: backbone.text_model.model.embeddings.word_embeddings.weight, backbone.text_model.model.embeddings.position_embeddings.weight, backbone.text_model.model.embeddings.token_type_embeddings.weight,****
我看了下都是 backbone.text_model. 开头的 然后输出的图片 有检测框 但是很多很乱 并且类别不对
我使用的是Closed-set finetuning微调方法，在test中能正常输出
Traceback (most recent call last):
File "/home/gxdong/yolo/WeDetect-main/infer_wedetect.py", line 213, in
model.reparameterize(texts)
File "/home/gxdong/anaconda3/envs/openmmlab_env/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1729, in getattr
AttributeError: 'SimpleYOLOWorldDetector' object has no attribute 'reparameterize'. Did you mean: 'reparameterized'?
我是用的是tiny
python3 infer_wedetect.py
--config /home/gxdong/yolo/WeDetect-main/work_dirs/wedetect_tiny_all/20260407_065304/vis_data/config.py
--checkpoint /home/gxdong/yolo/WeDetect-main/work_dirs/wedetect_tiny_all/best_coco_bbox_mAP_epoch_20.pth
--image images/aaa.jpg
--text '大车,小车,非机动车,人'
--threshold 0.1
更换成命令 不报错 但是加载
python3 infer_wedetect.py
--config /home/gxdong/yolo/WeDetect-main/config/wedetect_tiny.py
--checkpoint /home/gxdong/yolo/WeDetect-main/work_dirs/wedetect_tiny_all/best_coco_bbox_mAP_epoch_20.pth
--image images/aaa.jpg
--text '大车,小车,非机动车,人'
--threshold 0.1
The model and loaded state dict do not match exactly
missing keys in source state_dict: backbone.text_model.model.embeddings.word_embeddings.weight, backbone.text_model.model.embeddings.position_embeddings.weight, backbone.text_model.model.embeddings.token_type_embeddings.weight,****
我看了下都是 backbone.text_model. 开头的 然后输出的图片 有检测框 但是很多很乱 并且类别不对