关于 训练 fnlp/cpt-large，出现 维度不一致问题

[km1994]

我用以下命令 训练 fnlp/cpt-large
python train.py --train_file /mnt/kaimo/data/summary/bidding/train.json --dev_file /mnt/kaimo/data/summary/bidding/dev.json --batch_size 2 --max_epochs 10 --max_source_length 1024 --max_target_length 1024 --model_path fnlp/cpt-large --gpus 1 --lr 5e-5 --model_type cpt

但是报错了，报错内容为“

RuntimeError: Error(s) in loading state_dict for CPTForConditionalGeneration: size mismatch for model.encoder.embeddings.position_ids: copying a param with shape torch.Size([1, 1024]) from checkpoint, the shape in current model is torch.Size([1, 512]). size mismatch for model.encoder.embeddings.position_embeddings.weight: copying a param with shape torch.Size([1024, 1024]) from checkpoint, the shape in current model is torch.Size([512, 1024]). You may consider adding ignore_mismatched_sizes=Truein the modelfrom_pretrained method.

请问是哪个地方出问题？

[renmada]

可能我下载的权重比较早，cpt后来模型有更新。代码已更新，适配最新的fnlp/cpt-large