Open3DA/LL3DA

Training time

kuaileqipaoshui opened this issue · 5 comments

kuaileqipaoshui commented

How long it takes you to train?

ch3cook-fdu commented

The training time takes about a day or two depending on your devices. The evaluation time takes much longer. You can track the training process using the logged output.

kuaileqipaoshui commented

The training time takes about a day or two depending on your devices. The evaluation time takes much longer. You can track the training process using the logged output.

Thanks! Does a day or two days include the tune time?

ch3cook-fdu commented

Not included. The tuning time usually takes less than a day.

kuaileqipaoshui commented

Not included. The tuning time usually takes less than a day.

Thanks very much! When I try to train, I meet a problem, that is

Original Traceback (most recent call last):
File "/opt/conda/envs/ll3da/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 302, in _worker_loop
data = fetcher.fetch(index)
File "/opt/conda/envs/ll3da/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 58, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/opt/conda/envs/ll3da/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 58, in
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/data/private/LL3DA/datasets/unified_3dllm_scene_description.py", line 135, in getitem
ret_dict['box_query'] = box_query.astype(np.float32)
TypeError: 'NoneType' object does not support item assignment

I don't know how to solve it. How could I fix it?

ch3cook-fdu commented

You should check whether you have processed the scene data right and loaded the scene data properly in ret_dict.