|
|
@ -8,8 +8,7 @@ from lightning_sdk import Machine, Studio # noqa: F401 |
|
|
|
# consistency of randomly sampled experiments. |
|
|
|
seed(19920921) |
|
|
|
|
|
|
|
NUM_JOBS = 1 |
|
|
|
|
|
|
|
NUM_JOBS = 10 |
|
|
|
|
|
|
|
# Define the ranges or sets of values for each hyperparameter |
|
|
|
# alpha_values = list(np.round(np.linspace(2, 4, 21), 4)) |
|
|
@ -21,11 +20,12 @@ learning_rate_values = [1e-3] |
|
|
|
alpha_values = [0, 0.1] |
|
|
|
widths = [2**k for k in range(4, 13)] |
|
|
|
depths = [1, 2, 4, 8, 16] |
|
|
|
dropouts = [0, 0.25, 0.5] |
|
|
|
# widths, depths = [512], [4] |
|
|
|
|
|
|
|
batch_size_values = [256] |
|
|
|
max_epochs_values = [420] # at 12 fps, around 35s |
|
|
|
seeds = list(range(20, 1992)) |
|
|
|
seeds = list(range(21, 1992)) |
|
|
|
optimizers = [ |
|
|
|
# "Adagrad", |
|
|
|
"Adam", |
|
|
@ -39,7 +39,7 @@ optimizers = [ |
|
|
|
|
|
|
|
# Generate all possible combinations of hyperparameters |
|
|
|
all_params = [ |
|
|
|
(alpha, lr, bs, me, s, w, d, opt) |
|
|
|
(alpha, lr, bs, me, s, w, d, opt, dr) |
|
|
|
for alpha in alpha_values |
|
|
|
for lr in learning_rate_values |
|
|
|
for bs in batch_size_values |
|
|
@ -48,6 +48,7 @@ all_params = [ |
|
|
|
for w in widths |
|
|
|
for d in depths |
|
|
|
for opt in optimizers |
|
|
|
for dr in dropouts |
|
|
|
] |
|
|
|
|
|
|
|
|
|
|
@ -58,7 +59,7 @@ search_params = sample(all_params, min(NUM_JOBS, len(all_params))) |
|
|
|
# --trainer.callbacks.init_args.monitor hp_metric \ |
|
|
|
|
|
|
|
for idx, params in enumerate(search_params): |
|
|
|
a, lr, bs, me, s, w, d, opt = params |
|
|
|
a, lr, bs, me, s, w, d, opt, dr = params |
|
|
|
# cmd = f"cd ~/colors && python main.py --alpha {a} --lr {lr} --bs {bs} --max_epochs {me} --seed {s} --width {w}" |
|
|
|
cmd = f""" |
|
|
|
cd ~/colors && python newmain.py fit \ |
|
|
@ -72,7 +73,7 @@ cd ~/colors && python newmain.py fit \ |
|
|
|
--model.bias true \ |
|
|
|
--model.loop true \ |
|
|
|
--model.transform tanh \ |
|
|
|
--model.dropout 0.5 \ |
|
|
|
--model.dropout {dr} \ |
|
|
|
--trainer.min_epochs 10 \ |
|
|
|
--trainer.max_epochs {me} \ |
|
|
|
--trainer.log_every_n_steps 3 \ |
|
|
@ -84,19 +85,19 @@ cd ~/colors && python newmain.py fit \ |
|
|
|
--optimizer torch.optim.{opt} \ |
|
|
|
--optimizer.init_args.lr {lr} \ |
|
|
|
--trainer.callbacks+ lightning.pytorch.callbacks.LearningRateFinder |
|
|
|
# --lr_scheduler lightning.pytorch.cli.ReduceLROnPlateau \ |
|
|
|
# --lr_scheduler.init_args.monitor hp_metric \ |
|
|
|
# --lr_scheduler.init_args.factor 0.05 \ |
|
|
|
# --lr_scheduler.init_args.patience 5 \ |
|
|
|
# --lr_scheduler.init_args.cooldown 10 \ |
|
|
|
# --lr_scheduler.init_args.verbose true |
|
|
|
""" |
|
|
|
# --lr_scheduler lightning.pytorch.cli.ReduceLROnPlateau \ |
|
|
|
# --lr_scheduler.init_args.monitor hp_metric \ |
|
|
|
# --lr_scheduler.init_args.factor 0.05 \ |
|
|
|
# --lr_scheduler.init_args.patience 5 \ |
|
|
|
# --lr_scheduler.init_args.cooldown 10 \ |
|
|
|
# --lr_scheduler.init_args.verbose true |
|
|
|
print(f"Running {params}: {cmd}") |
|
|
|
try: |
|
|
|
studio = Studio("colors-animate-jobs") |
|
|
|
studio.install_plugin("jobs") |
|
|
|
job_plugin = studio.installed_plugins["jobs"] |
|
|
|
job_name = f"color-animate-{idx}-{s}" |
|
|
|
job_name = f"colors-animate-20240303-{idx}" |
|
|
|
job_plugin.run(cmd, machine=Machine.T4, name=job_name) |
|
|
|
|
|
|
|
# Run the command and wait for it to complete |
|
|
|