AntoineH/smart_augmentation
1
0
Fork 0
mirror of https://github.com/AntoineHX/smart_augmentation.git synced 2025-05-04 04:00:46 +02:00
Code Issues Projects Releases Packages Wiki Activity

Sauvegarde both mesure memoire

Browse source
This commit is contained in:
Harle, Antoine (Contracteur) 2020-02-10 14:50:45 -05:00
parent 7d5aa7c6fb
commit 65e67addf6
2 changed files with 14 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

12
higher/smart_aug/benchmark.py
View file

@ -97,6 +97,7 @@ if __name__ == "__main__":
for m_setup in mag_setup: for m_setup in mag_setup:
torch.cuda.reset_max_memory_allocated() #reset_peak_stats torch.cuda.reset_max_memory_allocated() #reset_peak_stats
torch.cuda.reset_max_memory_cached() #reset_peak_stats
t0 = time.perf_counter() t0 = time.perf_counter()
model = getattr(model_type, model_name)(pretrained=False) model = getattr(model_type, model_name)(pretrained=False)
@ -126,7 +127,8 @@ if __name__ == "__main__":
save_sample_freq=None) save_sample_freq=None)
exec_time=time.perf_counter() - t0 exec_time=time.perf_counter() - t0
max_cached = torch.cuda.max_memory_allocated()/(1024.0 * 1024.0) #torch.cuda.max_memory_reserved() #MB max_allocated = torch.cuda.max_memory_allocated()/(1024.0 * 1024.0)
max_cached = torch.cuda.max_memory_cached()/(1024.0 * 1024.0) #torch.cuda.max_memory_reserved() #MB
#### ####
print('-'*9) print('-'*9)
times = [x["time"] for x in log] times = [x["time"] for x in log]
@ -134,7 +136,7 @@ if __name__ == "__main__":
"Time": (np.mean(times),np.std(times), exec_time), "Time": (np.mean(times),np.std(times), exec_time),
'Optimizer': optim_param, 'Optimizer': optim_param,
"Device": device_name, "Device": device_name,
"Memory": max_cached, "Memory": [max_allocated, max_cached],
"Param_names": aug_model.TF_names(), "Param_names": aug_model.TF_names(),
"Log": log} "Log": log}
print(str(aug_model),": acc", out["Accuracy"], "in:", out["Time"][0], "+/-", out["Time"][1]) print(str(aug_model),": acc", out["Accuracy"], "in:", out["Time"][0], "+/-", out["Time"][1])
@ -155,6 +157,7 @@ if __name__ == "__main__":
for model_name in model_list[model_type]: for model_name in model_list[model_type]:
for run in range(nb_run): for run in range(nb_run):
torch.cuda.reset_max_memory_allocated() #reset_peak_stats torch.cuda.reset_max_memory_allocated() #reset_peak_stats
torch.cuda.reset_max_memory_cached() #reset_peak_stats
t0 = time.perf_counter() t0 = time.perf_counter()
model = getattr(model_type, model_name)(pretrained=False).to(device) model = getattr(model_type, model_name)(pretrained=False).to(device)
@ -164,7 +167,8 @@ if __name__ == "__main__":
log= train_classic(model=model, opt_param=optim_param, epochs=epochs, print_freq=epochs/4) log= train_classic(model=model, opt_param=optim_param, epochs=epochs, print_freq=epochs/4)
exec_time=time.perf_counter() - t0 exec_time=time.perf_counter() - t0
max_cached = torch.cuda.max_memory_allocated()/(1024.0 * 1024.0) #torch.cuda.max_memory_reserved() #MB max_allocated = torch.cuda.max_memory_allocated()/(1024.0 * 1024.0)
max_cached = torch.cuda.max_memory_cached()/(1024.0 * 1024.0) #torch.cuda.max_memory_reserved() #MB
#### ####
print('-'*9) print('-'*9)
times = [x["time"] for x in log] times = [x["time"] for x in log]
@ -172,7 +176,7 @@ if __name__ == "__main__":
"Time": (np.mean(times),np.std(times), exec_time), "Time": (np.mean(times),np.std(times), exec_time),
'Optimizer': optim_param, 'Optimizer': optim_param,
"Device": device_name, "Device": device_name,
"Memory": max_cached, "Memory": [max_allocated, max_cached],
#"Rand_Aug": rand_aug, #"Rand_Aug": rand_aug,
"Log": log} "Log": log}
print(model_name,": acc", out["Accuracy"], "in:", out["Time"][0], "+/-", out["Time"][1]) print(model_name,": acc", out["Accuracy"], "in:", out["Time"][0], "+/-", out["Time"][1])

10
higher/smart_aug/test_dataug.py
View file

@ -82,7 +82,7 @@ if __name__ == "__main__":
} }
#Parameters #Parameters
n_inner_iter = 1 n_inner_iter = 1
epochs = 150 epochs = 2
dataug_epoch_start=0 dataug_epoch_start=0
optim_param={ optim_param={
'Meta':{ 'Meta':{
@ -147,6 +147,7 @@ if __name__ == "__main__":
#### Augmented Model #### #### Augmented Model ####
if 'aug_model' in tasks: if 'aug_model' in tasks:
torch.cuda.reset_max_memory_allocated() #reset_peak_stats torch.cuda.reset_max_memory_allocated() #reset_peak_stats
torch.cuda.reset_max_memory_cached() #reset_peak_stats
t0 = time.perf_counter() t0 = time.perf_counter()
tf_dict = {k: TF.TF_dict[k] for k in tf_names} tf_dict = {k: TF.TF_dict[k] for k in tf_names}
@ -163,10 +164,11 @@ if __name__ == "__main__":
print_freq=1, print_freq=1,
unsup_loss=1, unsup_loss=1,
hp_opt=False, hp_opt=False,
save_sample_freq=1) save_sample_freq=None)
exec_time=time.perf_counter() - t0 exec_time=time.perf_counter() - t0
max_cached = torch.cuda.max_memory_allocated()/(1024.0 * 1024.0) #torch.cuda.max_memory_reserved() #MB max_allocated = torch.cuda.max_memory_allocated()/(1024.0 * 1024.0)
max_cached = torch.cuda.max_memory_cached()/(1024.0 * 1024.0) #torch.cuda.max_memory_reserved() #MB
#### ####
print('-'*9) print('-'*9)
times = [x["time"] for x in log] times = [x["time"] for x in log]
@ -174,7 +176,7 @@ if __name__ == "__main__":
"Time": (np.mean(times),np.std(times), exec_time), "Time": (np.mean(times),np.std(times), exec_time),
'Optimizer': optim_param, 'Optimizer': optim_param,
"Device": device_name, "Device": device_name,
"Memory": max_cached, "Memory": [max_allocated, max_cached],
"Param_names": aug_model.TF_names(), "Param_names": aug_model.TF_names(),
"Log": log} "Log": log}
print(str(aug_model),": acc", out["Accuracy"], "in:", out["Time"][0], "+/-", out["Time"][1]) print(str(aug_model),": acc", out["Accuracy"], "in:", out["Time"][0], "+/-", out["Time"][1])