search

dengyecode / T-former_image_inpainting

MIT License
23 stars 6 forks source link

trian problem! #2

Open 1225743219 opened 1 year ago

1225743219 commented 1 year ago

hello! Can you tell me how to train this model? I meet a issues

train.py [-h] [--name NAME] [--model MODEL] [--mask_type MASK_TYPE] [--checkpoints_dir CHECKPOINTS_DIR] [--which_iter WHICH_ITER] [--gpu_ids GPU_IDS] [--img_file IMG_FILE] [--mask_file MASK_FILE] [--loadSize LOADSIZE] [--fineSize FINESIZE] [--resize_or_crop RESIZE_OR_CROP] [--no_flip] [--no_rotation] [--no_augment] [--batchSize BATCHSIZE] [--nThreads NTHREADS] [--no_shuffle] [--display_winsize DISPLAY_WINSIZE] [--display_id DISPLAY_ID] [--display_port DISPLAY_PORT] [--display_single_pane_ncols DISPLAY_SINGLE_PANE_NCOLS] [--iter_count ITER_COUNT] [--niter NITER] [--niter_decay NITER_DECAY] [--continue_train] [--lr_policy LR_POLICY] [--lr LR] [--beta1 BETA1] [--beta2 BETA2] [--weight_decay WEIGHT_DECAY] [--gan_mode {wgan-gp,hinge,lsgan}] [--display_freq DISPLAY_FREQ] [--print_freq PRINT_FREQ] [--save_latest_freq SAVE_LATEST_FREQ] [--save_iters_freq SAVE_ITERS_FREQ] [--no_html] [--output_scale OUTPUT_SCALE] [--train_paths TRAIN_PATHS] [--lambda_per LAMBDA_PER] [--lambda_l1 LAMBDA_L1] [--lambda_g LAMBDA_G] [--lambda_sty LAMBDA_STY] train.py: error: unrecognized arguments: --image_file

dengyecode commented 1 year ago

Sorry, i have made mistakes. you may use ""--img_file"" instead of ""--image_file"".

1225743219 commented 1 year ago

Thank you very much for your reply. But after solving this problem, I encountered a new problem:

Requests. exceptions. ConnectionError: HTTPConnectionPool (host='localhost ', port=8097): Max retries exceeded with URL:/env/main (Caused by NewConnectionError ('<urllib3. connection. HTTPConnection object at 0x7efdda195590>: Failed to establish a new connection: [Errno 111] Connection reused ')

Why do I need to connect to the internet here

dengyecode commented 1 year ago

This is due to not running the visualization module, so please open an additional terminal and run "python - m visdom.server"

xukaili123 commented 1 year ago

I would like to ask why epoch starts from 1 when using the "fine tuning" command, instead of starting from epoch after training. Does this result in restarting training?

dengyecode commented 1 year ago

I would like to ask why epoch starts from 1 when using the "fine tuning" command, instead of starting from epoch after training. Does this result in restarting training?

In fact, due to a flaw in our code, it just doesn't inherit the "number" but actually inherits the model parameters after "xxxxxxx" iterations. To avoid confusion, we recommend you to create a new folder and copy the model files ("xx_net_D.pth" and "xx_net_G.pth") that you want to fine_tune into the new folder. Then continue training in the new folder