memleak.py

gistfile1.txt

import torch
import gc
def get_memory_mb():
    import resource
    return resource.getrusage(resource.RUSAGE_SELF).ru_maxrss / 1024
initial_mem = get_memory_mb()
print(f"Initial memory: {initial_mem:.2f} MB")
# Repeated operations cause memory leak
for i in range(10000):
    t = torch.randn(2, 3, 4, 5, 6)  # 5 dims
    t = t.view(1, 2, 3, 4, 5, 6)    # 6 dims - triggers leak
gc.collect()  # This is OUTSIDE the loop
final_mem = get_memory_mb()
print(f"Final memory: {final_mem:.2f} MB")
print(f"Memory leaked: {final_mem - initial_mem:.2f} MB")