How to load a pretrained model with a different output dimension

[priyamw]

i want to train a model for custom 22 key points out of which 17 are similar to coco can i use coco weights to my training as my model have 22 keypoints it says
The model and loaded state dict do not match exactly

size mismatch for head.final_layer.weight: copying a param with shape torch.Size([17, 768, 7, 7]) from checkpoint, the shape in current model is torch.Size([22, 768, 7, 7]).
size mismatch for head.final_layer.bias: copying a param with shape torch.Size([17]) from checkpoint, the shape in current model is torch.Size([22]).
The model and loaded state dict do not match exactly

size mismatch for head.final_layer.weight: copying a param with shape torch.Size([17, 768, 7, 7]) from checkpoint, the shape in current model is torch.Size([22, 768, 7, 7]).
size mismatch for head.final_layer.bias: copying a param with shape torch.Size([17]) from checkpoint, the shape in current model is torch.Size([22]).