about running predict.py about lama HOT 22 CLOSED

Hi! I'm not sure that I understand your problem and question. Could you please post the command you run and the stack trace?

from lama.

@windj007 I'm having a similar issue while I run the model on the local machine. Let me know what I should do about it.

** Initial Lines from shell
Detectron v2 is not installed
[2021-12-13 01:26:17,758][saicinpainting.utils][WARNING] - Setting signal 10 handler <function print_traceback_handler at 0x7fee50be7040>
[2021-12-13 01:26:17,780][root][INFO] - Make training model default
[2021-12-13 01:26:17,780][saicinpainting.training.trainers.base][INFO] - BaseInpaintingTrainingModule init called
[2021-12-13 01:26:17,781][root][INFO] - Make generator ffc_resnet
[2021-12-13 01:26:18,175][saicinpainting.training.trainers.base][INFO] - Generator
FFCResNetGenerator(
  (model): Sequential(
    (0): ReflectionPad2d((3, 3, 3, 3))
    (1): FFC_BN_ACT(
      (ffc): FFC(
        (convl2l): Conv2d(4, 64, kernel_size=(7, 7), stride=(1, 1), bias=False, padding_mode=reflect)
        (convl2g): Identity()
        (convg2l): Identity()
        (convg2g): Identity()
        (gate): Identity()
      )
*****
Lines of model architecture
*****   
** last few lines
    (33): ReflectionPad2d((3, 3, 3, 3))
    (34): Conv2d(64, 3, kernel_size=(7, 7), stride=(1, 1))
    (35): Sigmoid()
  )
)
[2021-12-13 01:26:18,175][saicinpainting.training.trainers.base][INFO] - BaseInpaintingTrainingModule init done
[2021-12-13 01:26:18,830][saicinpainting.training.data.datasets][INFO] - Make val dataloader default from /path/to/lama_model/input/

from lama.

@yyyyojiro I have installed Detectron 2 as it was asking. Still, isn't working.

python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'

from lama.

Hi! I'm not sure that I understand your problem and question. Could you please post the command you run and the stack trace?

thanks for your brilliant work and reply. I find that when I don't use gpu for predicting, it works. @SaiNikhileshReddy

from lama.

Sorry for the late reply again))

Detectron v2 is not installed

This is just a warning - it should not break training/prediction

mismatched input '(' expecting

This means that there is an error in any of the YAML files. Did you change any of the configs?

from lama.

Thanks for your brilliant work, and when I run predict.py, the errors occured.

Detectron v2 is not installed mismatched input '(' expecting See https://hydra.cc/docs/next/advanced/override_grammar/basic for details

Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace. how could I solve it, thanks!

I'm facing the similar error:
`Detectron v2 is not installed

Bad key "text.kerning_factor" on line 4 in
/home/dmsheng/anaconda3/envs/lama/lib/python3.6/site-packages/matplotlib/mpl-data/stylelib/_classic_test_patch.mplstyle.
You probably need to get an updated matplotlibrc file from
https://github.com/matplotlib/matplotlib/blob/v3.1.3/matplotlibrc.template
or from the matplotlib source distribution
Traceback (most recent call last):
File "/home/dmsheng/anaconda3/envs/lama/lib/python3.6/site-packages/hydra/core/utils.py", line 77, in _save_config
file.write(OmegaConf.to_yaml(cfg))
TypeError: dump_all() got an unexpected keyword argument 'sort_keys'

Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.`

@windj007 any idea to solve the problem?

from lama.

Hi!

That's strange... The lines about detectron and matplotlib are just warnings, they should not break anything.

Did you check the versions of the libraries installed and that they do not conflict with those set in requirements.txt?
It seems that you have a different version of Hydra and/or OmegaConf, which we did not test our pipeline with.

Could you please run it as HYDRA_FULL_ERROR=1 ./bin/predict.py .... ?

from lama.

I had the same error as the OP on Windows, fixed it by enter the arguments into configs/prediction/default.yaml and commenting out line 41 of predict.py

from lama.

Hi!

That's strange... The lines about detectron and matplotlib are just warnings, they should not break anything.

Did you check the versions of the libraries installed and that they do not conflict with those set in requirements.txt? It seems that you have a different version of Hydra and/or OmegaConf, which we did not test our pipeline with.

Could you please run it as HYDRA_FULL_ERROR=1 ./bin/predict.py .... ?

I simply solve the problem by downgrading pyyaml to version 5.4. Thanks for your reply.

from lama.

Hi! iIs the problem resolved currently? If yes, please let me know so I can close the issue

from lama.

Hi, I got some problems. @windj007

Detectron v2 is not installed
[2022-07-24 12:33:59,096][saicinpainting.utils][WARNING] - Setting signal 1 handler <function print_traceback_handler at 0x000002D268183DC0>
[2022-07-24 12:33:59,105][main][CRITICAL] - Prediction failed due to invalid signal value:
Traceback (most recent call last):
File "bin/predict.py", line 41, in main
register_debug_signal_handlers() # kill -10 will result in traceback dumped into log
File "D:\D\github\lama\saicinpainting\utils.py", line 109, in register_debug_signal_handlers
signal.signal(sig, handler)
File "D:\D\Anaconda\envs\lama\lib\signal.py", line 47, in signal
handler = _signal.signal(_enum_to_int(signalnum), _enum_to_int(handler))
ValueError: invalid signal value

Can you help me fixing this problem, please?

from lama.

Command: python bin/predict.py model.path=$(pwd)/big-lama indir=$(pwd)/LaMa_test_images outdir=$(pwd)/output
ERROR
Detectron v2 is not installed
mismatched input '(' expecting
See https://hydra.cc/docs/next/advanced/override_grammar/basic for details

Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.

Still getting this error.
@windj007

from lama.

I had the same error as the OP on Windows, fixed it by enter the arguments into configs/prediction/default.yaml and commenting out line 41 of predict.py

Can you please tell us the changes you make in configs/prediction/default.yaml in detail? @callanrocks

from lama.

I have the same error,did someone solved it?

from lama.

@abdullah-rashid62 I just changed the indir/outdir/path lines from no to the paths I wanted to use instead of inputting them through CLI.

from lama.

@windj007 I'm having a similar issue while I run the model on the local machine. Let me know what I should do about it.

** Initial Lines from shell
Detectron v2 is not installed
[2021-12-13 01:26:17,758][saicinpainting.utils][WARNING] - Setting signal 10 handler <function print_traceback_handler at 0x7fee50be7040>
[2021-12-13 01:26:17,780][root][INFO] - Make training model default
[2021-12-13 01:26:17,780][saicinpainting.training.trainers.base][INFO] - BaseInpaintingTrainingModule init called
[2021-12-13 01:26:17,781][root][INFO] - Make generator ffc_resnet
[2021-12-13 01:26:18,175][saicinpainting.training.trainers.base][INFO] - Generator
FFCResNetGenerator(
  (model): Sequential(
    (0): ReflectionPad2d((3, 3, 3, 3))
    (1): FFC_BN_ACT(
      (ffc): FFC(
        (convl2l): Conv2d(4, 64, kernel_size=(7, 7), stride=(1, 1), bias=False, padding_mode=reflect)
        (convl2g): Identity()
        (convg2l): Identity()
        (convg2g): Identity()
        (gate): Identity()
      )
*****
Lines of model architecture
*****   
** last few lines
    (33): ReflectionPad2d((3, 3, 3, 3))
    (34): Conv2d(64, 3, kernel_size=(7, 7), stride=(1, 1))
    (35): Sigmoid()
  )
)
[2021-12-13 01:26:18,175][saicinpainting.training.trainers.base][INFO] - BaseInpaintingTrainingModule init done
[2021-12-13 01:26:18,830][saicinpainting.training.data.datasets][INFO] - Make val dataloader default from /path/to/lama_model/input/

It's completely not the similar issue, man

from lama.

@abdullah-rashid62 I just changed the indir/outdir/path lines from no to the paths I wanted to use instead of inputting them through CLI.

Put all configs into configs/prediction/defaults.yaml instead of CLI could solve the issue

from lama.

Anyone found solution for windows?

from lama.

@abdullah-rashid62 I just changed the indir/outdir/path lines from no to the paths I wanted to use instead of inputting them through CLI.

Put all configs into configs/prediction/defaults.yaml instead of CLI could solve the issue

Not working :(

from lama.

To solve this issue you need to use this two steps,

Step-1: As @callanrocks mentioned, comment out line 41 of predict.py.

Step-2: Issue on windows OS is, it requires full path instead of relative path. So in predict.py, go to line 45 where train_config_path is being set and add a line below it as train_config_path = os.path.join('<your_full_path_to_lama_base_directory>', train_config_path) and then go to line 54 where checkpoint_path is being set and add a line below it as, checkpoint_path = os.path.join('<your_full_path_to_lama_base_directory>', checkpoint_path). This will allow the program to use full path instead of relative path.

Doing so shall solve the issue. @windj007 @arnavmehta7 @abdullah-rashid62 @yumengWang112

from lama.

在本地计算机上运行模型时，我遇到了类似的问题。让我知道我应该怎么做。

** Initial Lines from shell
Detectron v2 is not installed
[2021-12-13 01:26:17,758][saicinpainting.utils][WARNING] - Setting signal 10 handler <function print_traceback_handler at 0x7fee50be7040>
[2021-12-13 01:26:17,780][root][INFO] - Make training model default
[2021-12-13 01:26:17,780][saicinpainting.training.trainers.base][INFO] - BaseInpaintingTrainingModule init called
[2021-12-13 01:26:17,781][root][INFO] - Make generator ffc_resnet
[2021-12-13 01:26:18,175][saicinpainting.training.trainers.base][INFO] - Generator
FFCResNetGenerator(
  (model): Sequential(
    (0): ReflectionPad2d((3, 3, 3, 3))
    (1): FFC_BN_ACT(
      (ffc): FFC(
        (convl2l): Conv2d(4, 64, kernel_size=(7, 7), stride=(1, 1), bias=False, padding_mode=reflect)
        (convl2g): Identity()
        (convg2l): Identity()
        (convg2g): Identity()
        (gate): Identity()
      )
*****
Lines of model architecture
*****   
** last few lines
    (33): ReflectionPad2d((3, 3, 3, 3))
    (34): Conv2d(64, 3, kernel_size=(7, 7), stride=(1, 1))
    (35): Sigmoid()
  )
)
[2021-12-13 01:26:18,175][saicinpainting.training.trainers.base][INFO] - BaseInpaintingTrainingModule init done
[2021-12-13 01:26:18,830][saicinpainting.training.data.datasets][INFO] - Make val dataloader default from /path/to/lama_model/input/

这完全不是类似的问题，伙计

detectron2
I am also experiencing such a problem

from lama.

TY @moniltusharshah !

from lama.