Exemplos de slurm_distributed_context em Python

Exemplo n.º 1

Arquivo: distributed_train.py Projeto: sawravchy/ParlAI

 def run(self):
     with distributed_utils.slurm_distributed_context(self.opt) as opt:
         self.train_loop = single_train.TrainLoop(opt)
         self.parser = self.parser
         self.parser.opt = self.train_loop.agent.opt
         self.parser.print_args()
         return self.train_loop.train()

Exemplo n.º 2

Arquivo: distributed_eval.py Projeto: sawravchy/ParlAI

def main():
    parser = eval_model.setup_args()
    parser.add_distributed_training_args()
    parser.add_argument('--port', type=int, default=61337, help='TCP port number')
    opt = parser.parse_args(print_args=(os.environ['SLURM_PROCID'] == '0'))

    with distributed_utils.slurm_distributed_context(opt) as opt:
        return eval_model.eval_model(opt)

Exemplo n.º 3

Arquivo: distributed_tod_world_script.py Projeto: sagar-spkt/ParlAI

 def run(self):
     with distributed_utils.slurm_distributed_context(self.opt) as opt:
         return TodWorldScript(opt).run()

Exemplo n.º 4

 def run(self):
     with distributed_utils.slurm_distributed_context(self.opt) as opt:
         self.train_loop = single_train.TrainLoop(opt)
         return self.train_loop.train()

Exemplo n.º 5

Arquivo: distributed_eval.py Projeto: advi1012/ParlAITest

 def run(self):
     with distributed_utils.slurm_distributed_context(self.opt) as opt:
         return eval_model.eval_model(opt)