Thanks a great job. I want to run bash urbangpt_eval.sh, so I changed the setting from urbangpt_eval.sh. For example, output_model=./vicuna-7b-v1.5, which is a checkpoint download from huggingface, and I change the datapath and so on. By the way, I don't have ./checkpoints/tw2t_multi_reg-cla-gird. And then, I ran and got an error:
File "/nfs/code/LLMs/UrbanGPT/urbangpt/model/STLlama.py", line 123, in set_st_tower st_tower = st_tower.to(dtype=torch.float32) ^^^^^^^^^^^ AttributeError: 'NoneType' object has no attribute 'to'
I think the problem is output_model. Can anyone provide?
Thanks for your attention. Please use the checkpoint of UrbanGPT for evaluation. If you want to train from scratch, you can use vicuna-7b-v1.5 as the base model.
Thanks a great job. I want to run
bash urbangpt_eval.sh
, so I changed the setting from urbangpt_eval.sh. For example,output_model=./vicuna-7b-v1.5
, which is a checkpoint download from huggingface, and I change the datapath and so on. By the way, I don't have./checkpoints/tw2t_multi_reg-cla-gird
. And then, I ran and got an error:File "/nfs/code/LLMs/UrbanGPT/urbangpt/model/STLlama.py", line 123, in set_st_tower st_tower = st_tower.to(dtype=torch.float32) ^^^^^^^^^^^ AttributeError: 'NoneType' object has no attribute 'to'
I think the problem is output_model. Can anyone provide?