FOTS.PyTorch icon indicating copy to clipboard operation
FOTS.PyTorch copied to clipboard

RuntimeError:set_storage is not allowed on Tensor created from .data or .detach()

Open isyanan1024 opened this issue 6 years ago • 1 comments

This is my config.json { "name": "FOTS", "cuda": true, "gpus": [0,1,2,3], "data_loader": { "dataset":"icdar2015", "data_dir": "./text_data/2015/training", "batch_size": 24, "shuffle": true, "workers": 4 }, "validation": { "validation_split": 0.2, "shuffle": true }, "lr_scheduler_type": "", "lr_scheduler_freq": 10000, "lr_scheduler": { "gamma": 0.94 },

"optimizer_type": "Adam",
"optimizer": {
    "lr": 0.001
},
"loss": "FOTSLoss",
"metrics": ["fots_metric"],
"trainer": {
    "epochs": 1000,
    "save_dir": "saved/",
    "save_freq": 1,
    "verbosity": 2,
    "monitor": "hmean",
    "monitor_mode": "max"
},
"arch": "FOTSModel",
"model": {
    "mode": "united"
}

} who can tell me?

isyanan1024 avatar Aug 27 '19 05:08 isyanan1024

when i use gpu[0,1,2,3],it's wrong. use gpu[2,3],it's ok! why?

isyanan1024 avatar Aug 27 '19 06:08 isyanan1024