the bug in convert_model_to_long.ipynb

There are two bugs in create_long_model() functions.
The first is adding one line
model.roberta.embeddings.position_ids.data = torch.tensor([i for i in range(max_pos)]).reshape(1, max_pos)
after
model.roberta.embeddings.position_embeddings.weight.data = new_pos_embed

The second is needing copy.deepcopy

longformer_self_attn = LongformerSelfAttention(config, layer_id=i)
longformer_self_attn.query = layer.attention.self.query
longformer_self_attn.key = layer.attention.self.key
longformer_self_attn.value = layer.attention.self.value

longformer_self_attn.query_global = layer.attention.self.query
longformer_self_attn.key_global = layer.attention.self.key
longformer_self_attn.value_global = layer.attention.self.value

--->

longformer_self_attn = LongformerSelfAttention(config, layer_id=i)
longformer_self_attn.query = copy.deepcopy(layer.attention.self.query)
longformer_self_attn.key = copy.deepcopy(layer.attention.self.key)
longformer_self_attn.value = copy.deepcopy(layer.attention.self.value)

longformer_self_attn.query_global = copy.deepcopy(layer.attention.self.query)
longformer_self_attn.key_global = copy.deepcopy(layer.attention.self.key)
longformer_self_attn.value_global = copy.deepcopy(layer.attention.self.value)