Skip to content

dlc3.0 cannot find 'pose_estimation_pytorch/config/dlcrnet/dlcrnet_ms5.yaml' when creating training data #2655

@chz31

Description

@chz31

Is there an existing issue for this?

  • I have searched the existing issues

Bug description

I installed deeplabcut 3.0 with pytorch in both Windows 11 and Ubuntu 22.04 in WSL2 in Windows 11 following the instruction here: #2644 (comment). I created a project and labeled video. However, when I tried to create a training dataset by simply running deeplabcut.create_multianimaltraining_dataset(config_path)

It showed below error.
No such file or directory: '/home/chz31/anaconda3/envs/deeplabcut3/lib/python3.11/site-packages/deeplabcut/pose_estimation_pytorch/config/dlcrnet/dlcrnet_ms5.yaml'

Just by looking at the pytorch-dlc branch, the location only has dlcrnet_stride16_ms5.yaml and dlcrnet_stride17_ms5.yaml

Using conda env create -f DEEPLABCUT.yaml produced the same error.

Operating System

Windows 11 and Ubuntu 22.04 WSL2 in Windows 11

DeepLabCut version

deeplabcut 3.0 with pytorch

DeepLabCut mode

multi animal

Device type

GPU should be able to be accessed. python -c "import torch;print(torch.cuda.is_available()) returns True.

Steps To Reproduce

No response

Relevant log output

FileNotFoundError                         Traceback (most recent call last)
Cell In[8], line 1
----> 1 deeplabcut.create_multianimaltraining_dataset(config_path)

File ~/anaconda3/envs/deeplabcut3/lib/python3.11/site-packages/deeplabcut/generate_training_dataset/multiple_individuals_trainingsetmanipulation.py:581, in create_multianimaltraining_dataset(config, num_shuffles, Shuffles, windows2linux, net_type, numdigits, crop_size, crop_sampling, paf_graph, trainIndices, testIndices, n_edges_threshold, paf_graph_degree, userfeedback, weight_init, engine)
    574         pytorch_cfg = make_super_animal_finetune_config(
    575             project_config=cfg,
    576             pose_config_path=path_train_config,
    577             net_type=net_type,
    578             weight_init=weight_init,
    579         )
    580     else:
--> 581         pytorch_cfg = make_pytorch_pose_config(
    582             project_config=cfg,
    583             pose_config_path=path_train_config,
    584             net_type=net_type,
    585             top_down=top_down,
    586             weight_init=weight_init,
    587         )
    589     auxiliaryfunctions.write_plainconfig(pose_cfg_path, pytorch_cfg)
    591 print(
    592     "The training dataset is successfully created. Use the function 'train_network' to start training. Happy training!"
    593 )

File ~/anaconda3/envs/deeplabcut3/lib/python3.11/site-packages/deeplabcut/pose_estimation_pytorch/config/make_pose_config.py:107, in make_pytorch_pose_config(project_config, pose_config_path, net_type, top_down, weight_init)
    104         default_value_kwargs.update(_get_paf_parameters(project_config, bodyparts))
    106     cfg_path = configs_dir / architecture / f"{net_type}.yaml"
--> 107     model_cfg = read_config_as_dict(cfg_path)
    108     model_cfg = replace_default_values(
    109         model_cfg,
    110         num_bodyparts=len(bodyparts),
    111         num_individuals=len(individuals),
    112         **default_value_kwargs,
    113     )
    115 is_top_down = model_cfg.get("method", "BU").upper() == "TD"

File ~/anaconda3/envs/deeplabcut3/lib/python3.11/site-packages/deeplabcut/pose_estimation_pytorch/config/utils.py:186, in read_config_as_dict(config_path)
    178 def read_config_as_dict(config_path: str | Path) -> dict:
    179     """
    180     Args:
    181         config_path: the path to the configuration file to load
   (...)
    184         The configuration file with pure Python classes
    185     """
--> 186     with open(config_path, "r") as f:
    187         cfg = YAML(typ='safe', pure=True).load(f)
    189     return cfg

FileNotFoundError: [Errno 2] No such file or directory: '/home/chz31/anaconda3/envs/deeplabcut3/lib/python3.11/site-packages/deeplabcut/pose_estimation_pytorch/config/dlcrnet/dlcrnet_ms5.yaml'

Anything else?

No response

Code of Conduct

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions