How to freeze a model?

I use

        for p in model.parameters():
            p.requires_grad = False

to freeze a T5 model, but when I print parameters that require grad, there is still one parameter with the size 32000x512. What is this? Is it embeddings matrix? Should I freeze it too? It seems backward gradients affect this one remaining parameter

It seems I have called model.resize_token_embeddings(len(tokenizer)) after freezing parameters, and it can affect the embeddings require_grad