How to apply gradient clip in fused optimizer?

I noticed that my embedding bag parameters exploded. Is there a way I could apply gradient clip. 
I'm using `EmbOptimType.EXACT_ROWWISE_ADAGRAD`

Here is the code


```
    sharder_with_optim_params = EmbeddingBagCollectionSharder(
        fused_params={
            'optimizer': EmbOptimType.EXACT_ROWWISE_ADAGRAD,
            'learning_rate': 0.01,
            'eps': 1e-8,
        },
    )
```