transformers

fa6d12f7 - Allow to train dinov2 with different dtypes like bf16 (#28504)

Allow to train dinov2 with different dtypes like bf16 (#28504) I want to train dinov2 with bf16 but I get the following error in https://github.com/huggingface/transformers/blob/bc72b4e2cdcbc80d5f56731f35dbc9c18b4c8de6/src/transformers/models/dinov2/modeling_dinov2.py#L635: ``` RuntimeError: Input type (float) and bias type (c10::BFloat16) should be the same ``` Since the input dtype is torch.float32, the parameter dtype has to be torch.float32... @LZHgrla and I checked the code of clip vision encoder and found there is an automatic dtype transformation (https://github.com/huggingface/transformers/blob/bc72b4e2cdcbc80d5f56731f35dbc9c18b4c8de6/src/transformers/models/clip/modeling_clip.py#L181-L182). So I add similar automatic dtype transformation to modeling_dinov2.py.