unable to guess the VisDA-2017 dataset structure for semantic segmentation task

[vamsikrishna7909]

Hi there,

I am trying to run this project in my machine to experiment something. I setup the project and I wanted to train it with VisDA-2017 segmentation datasets like GTA5 and Cityscapes. But the script is expecting something and I'm not sure.

Here is the issue:

`Loading trainer: DAPL Loading dataset: VisDA17 Traceback (most recent call last): File "train.py", line 208, in main(args) File "train.py", line 127, in main trainer = build_trainer(cfg) File "/home/[email protected]/ENTER/envs/dassl/lib/python3.8/site-packages/dassl/engine/build.py", line 11, in build_trainer return TRAINER_REGISTRY.get(cfg.TRAINER.NAME)(cfg) File "/home/[email protected]/ENTER/envs/dassl/lib/python3.8/site-packages/dassl/engine/trainer.py", line 324, in init self.build_data_loader() File "/home/[email protected]/ENTER/envs/dassl/lib/python3.8/site-packages/dassl/engine/trainer.py", line 347, in build_data_loader dm = DataManager(self.cfg) File "/home/[email protected]/ENTER/envs/dassl/lib/python3.8/site-packages/dassl/data/data_manager.py", line 61, in init dataset = build_dataset(cfg) File "/home/[email protected]/ENTER/envs/dassl/lib/python3.8/site-packages/dassl/data/datasets/build.py", line 11, in build_dataset return DATASET_REGISTRY.get(cfg.DATASET.NAME)(cfg) File "/home/[email protected]/ENTER/envs/dassl/lib/python3.8/site-packages/dassl/data/datasets/da/visda17.py", line 31, in init train_x = self._read_data("synthetic") File "/home/[email protected]/ENTER/envs/dassl/lib/python3.8/site-packages/dassl/data/datasets/da/visda17.py", line 44, in _read_data with open(image_list, "r") as f: FileNotFoundError: [Errno 2] No such file or directory: '/vamsi_project/da_prompt_repo/DAPrompt/data/visda17/train/image_list.txt' ^C

(dassl) :/vamsi_project/da_prompt_repo/DAPrompt/scripts$ cd .. (dass)/vamsi_project/da_prompt_repo/DAPrompt$ cd data (dassl) :/vamsi_project/da_prompt_repo/DAPrompt/data$ ls get_gta.sh real synthetic test_data test.zip (dassl) /vamsi_project/da_prompt_repo/DAPrompt/data$ cd real (dassl):/vamsi_project/da_prompt_repo/DAPrompt/data/real$ ls cookie.txt gtFine leftImg8bit (dassl) /vamsi_project/da_prompt_repo/DAPrompt/data/real$ cd .. (dassl) :/vamsi_project/da_prompt_repo/DAPrompt/data$ cd synthetic/ (dassl) :/vamsi_project/da_prompt_repo/DAPrompt/data/synthetic$ ls extract_gta5_from_scratch.sh extraction_log.txt images labels mapping.mat readImage.m split.mat`

Please let me know

[John-Ge]

Our project only supports classification now.

[vamsikrishna7909]

Oh! yeah. Thanks for clarification. I thought we can use this experiment for segmentation track as well.

Thanks again for your prompt response. I really appreciate it.

[John-Ge]

Maybe some papers following our work develop method for segmentation. You could check it.

[vamsikrishna7909]

Okay. I'll start researching on it. Thanks!

By the way, I'm trying to run this project using VisDA-2017 classification dataset. I downloaded the datasets as suggested in readme.

when I try to run the train command, I'm getting the below response. `(dassl) vamsi_project/da_prompt_repo/DAPrompt/scripts$ bash main.sh visda17 ep25-32-csc 1.0 0.5 1.0 t0 Results are available in output/visda17/DAPL/ep25-32-csc/1.0_0.5_1.0_t0/seed_1. Skip this job Results are available in output/visda17/DAPL/ep25-32-csc/1.0_0.5_1.0_t0/seed_2. Skip this job Results are available in output/visda17/DAPL/ep25-32-csc/1.0_0.5_1.0_t0/seed_3. Skip this job Results are available in output/visda17/DAPL/ep25-32-csc/1.0_0.5_1.0_t0/seed_4. Skip this job Results are available in output/visda17/DAPL/ep25-32-csc/1.0_0.5_1.0_t0/seed_5. Skip this job

`

when I go and see log files. I don't know what's missing. Please let me know.

[vamsikrishna7909]

Never mind, I got to clear the previous log files.