The ego4d_nlq_2022_1st_place_solution from nnnnai

ego4d_nlq_2022_1st_place_solution's People

Contributors

Stargazers

Watchers

ego4d_nlq_2022_1st_place_solution's Issues

Code for CLIP video+text feature extraction

Hi,

Thanks for sharing this repo to reproduce the NLQ results. I'm interested in extracting CLIP features from raw videos + text similar to your pipeline. I was hoping you could share the code to do this.

Evaluation: FileNotFoundError: [Errno 2] No such file or directory: './results/video_tef-vlen600_slowfast/opt.json'

Hi,

could you provide the opt.json file that comes along with the already provided model weights ./results/video_tef-vlen600_slowfast/model_0146.ckpt?

$ python -m ms_cm.inference_ego4d_slowfast --resume ./results/video_tef-vlen600_slowfast/model_0146.ckpt --split val
Better speed can be achieved with apex installed from https://www.github.com/nvidia/apex .
2023-11-16 08:52:23.084:INFO:__main__ - Setup config, data and model...
Traceback (most recent call last):
  File "<frozen runpy>", line 198, in _run_module_as_main
  File "<frozen runpy>", line 88, in _run_code
  File "/reler/ms_cm/inference_ego4d_slowfast.py", line 291, in <module>
    start_inference()
  File "/reler/ms_cm/inference_ego4d_slowfast.py", line 210, in start_inference
    opt = TestOptions().parse()
          ^^^^^^^^^^^^^^^^^^^^^
  File "/reler/ms_cm/configs.py", line 250, in parse
    saved_options = load_json(os.path.join(opt.model_dir, self.saved_option_filename))
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/utils/basic_utils.py", line 21, in load_json
    with open(filename, "r") as f:
         ^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: './results/video_tef-vlen600_slowfast/opt.json'

Thanks a lot! 😄

Best,
Maximotus

Recommend Projects