Misleading error when trying to load models with underscore

Thanks for this repo!

Quick remark: I tried passing `efficientnet_b0` but you get `KeyError: 'efficientnet_b0'` if you try to pass this.   

Yet, if you pass a wrong model name such as  `efficientnet_0` you get the following message.

```
ValueError: model_name should be one of: efficientnet_b0, efficientnet_b1, efficientnet_b2, efficientnet_b3, efficientnet_b4, efficientnet_b5, efficientnet_b6, efficientnet_b7  
```   

The error should tell you to provide a name with  a dash  instead.

```
ValueError: model_name should be one of: efficientnet-b0, efficientnet-b1, efficientnet-b2, efficientnet-b3, efficientnet-b4, efficientnet-b5, efficientnet-b6, efficientnet-b7   
```     

I think the proposed small change fixes the issue.

Author: MichelML

efficientnet_pytorch/model.py

@@ -207,6 +207,6 @@ def _check_model_name_is_valid(cls, model_name, also_need_pretrained_weights=Fal
         """ Validates model name. None that pretrained weights are only available for
         the first four models (efficientnet-b{i} for i in 0,1,2,3) at the moment. """
         num_models = 4 if also_need_pretrained_weights else 8
-        valid_models = ['efficientnet_b'+str(i) for i in range(num_models)]
-        if model_name.replace('-','_') not in valid_models:
+        valid_models = ['efficientnet-b'+str(i) for i in range(num_models)]
+        if model_name not in valid_models:
             raise ValueError('model_name should be one of: ' + ', '.join(valid_models))