在训练自己的数据集时，类别数报错。

[qjh666888]

Traceback (most recent call last):
File "main.py", line 371, in
main(args)
File "main.py", line 254, in main
missing_keys, unexpected_keys = model_without_ddp.load_state_dict(checkpoint['model'], strict=False)
File "/data/qjh/envs/meta-detr/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1052, in load_state_dict
self.class.name, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for MetaDETR:
size mismatch for category_codes_cls.L.weight_g: copying a param with shape torch.Size([91, 1]) from checkpoint, the shape in current model is torch.Size([10, 1]).
size mismatch for category_codes_cls.L.weight_v: copying a param with shape torch.Size([91, 256]) from checkpoint, the shape in current model is torch.Size([10, 256]).

有人遇到过吗，应该怎么修改

[NHeLv1]

您好，想和您讨论一下训练自己数据集的问题，只需要按照coco格式存放就可以了吗

[qjh666888]

我没有使用这个代码训练自己的数据集

…

------------------ 原始邮件 ------------------ 发件人: "Shen ***@***.***>; 发送时间: 2024年2月28日(星期三) 晚上8:55 收件人: ***@***.***>; 抄送: ***@***.***>; "State ***@***.***>; 主题: Re: [ZhangGongjie/Meta-DETR] 在训练自己的数据集时，类别数报错。 (Issue #78) 您好，想和您讨论一下训练自己数据集的问题，只需要按照coco格式存放就可以了吗 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: ***@***.***>