Multiple GPU's issue.

modeling_chatglm.py

@@ -858,7 +858,7 @@ class ChatGLMModel(ChatGLMPreTrainedModel):
                         self.config.eoi_token_id)
                     assert eoi_token_pos - boi_token_pos == 2
                     new_input_embeds.append(torch.cat(
-                        (inputs_embeds[i, :boi_token_pos], images_features[i], inputs_embeds[i, eoi_token_pos + 1:])))
+                        (inputs_embeds[i, :boi_token_pos], images_features[i].to(inputs_embeds.device), inputs_embeds[i, eoi_token_pos + 1:])))
                     new_position_ids.append(torch.cat(
                         (position_ids[i, :boi_token_pos + 1], position_ids[i, boi_token_pos + 1].repeat(num_patches),
                          position_ids[i, eoi_token_pos:])