I am following a hugging face guide for fine tuning whisper but I run into error when training

I am following this blog by hugging face: https://huggingface.co/blog/fine-tune-whisper

I am using my dataset and followed all the steps until the last step where I got an error

Error: RuntimeError: The size of tensor a (802) must match the size of tensor b (448) at non-singleton dimension 1.

Here is the colab notebook: Google Colab

With the help of colab ai it suggests that the error is originating in the compute_metrics function but I am using the same compute_metrics function as the one in the hugging face blog.

How can I fix this issue?