neurips21-cifar-centralized-reps.py

#!/usr/bin/env python3

import os
import json
import subprocess

from shared import remote_exec, register_job, mongo, upload_code_package

code_package = upload_code_package()

gpus_per_node = 4
num_workers = 16

experiment = os.path.splitext(os.path.basename(__file__))[0]

description = "Replicating with different seeds"

base_config = {
    "seed": 1,
    "task": "Cifar",
    "model_name": "VGG-11",
    "topology": "fully-connected",
    "algorithm": "all-reduce",
    "overlap_communication": False,
    "base_optimizer": "SGD",
    "num_epochs": 200,
    "num_lr_warmup_epochs": 5,
    "lr_schedule_milestones": [(150, 0.1), (180, 0.1)],
    "batch_size": 32,
    "momentum": 0.9,
    "weight_decay": 1e-4,
    "data_split_method": "random",
    "non_iid_alpha": None,
    "distributed_world_size": num_workers, 
    "gpus_per_node": gpus_per_node,
}

for seed in [2, 3]:
    for lr in [0.1]:
        config = {**base_config, "learning_rate": lr, "seed": seed}
        job_name = "{algorithm}-mom{momentum}-lr{learning_rate}-sd{seed}".format(**config)
        if mongo.job.count_documents({"job": job_name, "experiment": experiment, **{f"config.{key}": value for key, value in config.items()}}) > 0:
            # We have this one already
            continue
        job_id = register_job(
            user="anonymized",
            project="average-routing",
            experiment=experiment,
            job=job_name,
            priority=10,
            n_workers=num_workers,
            config_overrides=config,
            runtime_environment={"clone": {"code_package": code_package}, "script": "train.py"},
            annotations={"description": description},
        )
        remote_exec(f'sbatch --nodes 1 --ntasks {num_workers} --gres gpu:{gpus_per_node} --cpus-per-task 2 --job-name="{job_name}" --wrap="srun jobrun {job_id} --mpi"')

for seed in [1, 2, 3]:
    for lr in [0.1]:
        config = {**base_config, "learning_rate": lr, "momentum": 0.0, "seed": seed}
        job_name = "{algorithm}-mom{momentum}-lr{learning_rate}sd{seed}".format(**config)
        if mongo.job.count_documents({"job": job_name, "experiment": experiment, **{f"config.{key}": value for key, value in config.items()}}) > 0:
            # We have this one already
            continue
        job_id = register_job(
            user="anonymized",
            project="average-routing",
            experiment=experiment,
            job=job_name,
            priority=10,
            n_workers=num_workers,
            config_overrides=config,
            runtime_environment={"clone": {"code_package": code_package}, "script": "train.py"},
            annotations={"description": description},
        )
        remote_exec(f'sbatch --nodes 1 --ntasks {num_workers} --gres gpu:{gpus_per_node} --cpus-per-task 2 --job-name="{job_name}" --wrap="srun jobrun {job_id} --mpi"')