glenn-jocher
commited on
Commit
•
6e46617
1
Parent(s):
cf298fb
AutoBatch checks against failed solutions (#8159)
Browse files* AutoBatch checks against failed solutions
@kalenmike this is a simple improvement to AutoBatch to verify that returned solutions have not already failed, i.e. return batch-size 8 when 8 already produced CUDA out of memory.
This is a halfway fix until I can implement a 'final solution' that will actively verify the solved-for batch size rather than passively assume it works.
* Update autobatch.py
* Update autobatch.py
- utils/autobatch.py +19 -10
utils/autobatch.py
CHANGED
@@ -8,7 +8,7 @@ from copy import deepcopy
|
|
8 |
import numpy as np
|
9 |
import torch
|
10 |
|
11 |
-
from utils.general import LOGGER, colorstr
|
12 |
from utils.torch_utils import profile
|
13 |
|
14 |
|
@@ -26,6 +26,7 @@ def autobatch(model, imgsz=640, fraction=0.9, batch_size=16):
|
|
26 |
# model = torch.hub.load('ultralytics/yolov5', 'yolov5s', autoshape=False)
|
27 |
# print(autobatch(model))
|
28 |
|
|
|
29 |
prefix = colorstr('AutoBatch: ')
|
30 |
LOGGER.info(f'{prefix}Computing optimal batch size for --imgsz {imgsz}')
|
31 |
device = next(model.parameters()).device # get model device
|
@@ -33,25 +34,33 @@ def autobatch(model, imgsz=640, fraction=0.9, batch_size=16):
|
|
33 |
LOGGER.info(f'{prefix}CUDA not detected, using default CPU batch-size {batch_size}')
|
34 |
return batch_size
|
35 |
|
|
|
36 |
gb = 1 << 30 # bytes to GiB (1024 ** 3)
|
37 |
d = str(device).upper() # 'CUDA:0'
|
38 |
properties = torch.cuda.get_device_properties(device) # device properties
|
39 |
-
t = properties.total_memory / gb #
|
40 |
-
r = torch.cuda.memory_reserved(device) / gb #
|
41 |
-
a = torch.cuda.memory_allocated(device) / gb #
|
42 |
-
f = t - (r + a) # free
|
43 |
LOGGER.info(f'{prefix}{d} ({properties.name}) {t:.2f}G total, {r:.2f}G reserved, {a:.2f}G allocated, {f:.2f}G free')
|
44 |
|
|
|
45 |
batch_sizes = [1, 2, 4, 8, 16]
|
46 |
try:
|
47 |
img = [torch.zeros(b, 3, imgsz, imgsz) for b in batch_sizes]
|
48 |
-
|
49 |
except Exception as e:
|
50 |
LOGGER.warning(f'{prefix}{e}')
|
51 |
|
52 |
-
|
53 |
-
|
54 |
-
p = np.polyfit(batch_sizes, y, deg=1) # first degree polynomial fit
|
55 |
b = int((f * fraction - p[1]) / p[0]) # y intercept (optimal batch size)
|
56 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
57 |
return b
|
|
|
8 |
import numpy as np
|
9 |
import torch
|
10 |
|
11 |
+
from utils.general import LOGGER, colorstr, emojis
|
12 |
from utils.torch_utils import profile
|
13 |
|
14 |
|
|
|
26 |
# model = torch.hub.load('ultralytics/yolov5', 'yolov5s', autoshape=False)
|
27 |
# print(autobatch(model))
|
28 |
|
29 |
+
# Check device
|
30 |
prefix = colorstr('AutoBatch: ')
|
31 |
LOGGER.info(f'{prefix}Computing optimal batch size for --imgsz {imgsz}')
|
32 |
device = next(model.parameters()).device # get model device
|
|
|
34 |
LOGGER.info(f'{prefix}CUDA not detected, using default CPU batch-size {batch_size}')
|
35 |
return batch_size
|
36 |
|
37 |
+
# Inspect CUDA memory
|
38 |
gb = 1 << 30 # bytes to GiB (1024 ** 3)
|
39 |
d = str(device).upper() # 'CUDA:0'
|
40 |
properties = torch.cuda.get_device_properties(device) # device properties
|
41 |
+
t = properties.total_memory / gb # GiB total
|
42 |
+
r = torch.cuda.memory_reserved(device) / gb # GiB reserved
|
43 |
+
a = torch.cuda.memory_allocated(device) / gb # GiB allocated
|
44 |
+
f = t - (r + a) # GiB free
|
45 |
LOGGER.info(f'{prefix}{d} ({properties.name}) {t:.2f}G total, {r:.2f}G reserved, {a:.2f}G allocated, {f:.2f}G free')
|
46 |
|
47 |
+
# Profile batch sizes
|
48 |
batch_sizes = [1, 2, 4, 8, 16]
|
49 |
try:
|
50 |
img = [torch.zeros(b, 3, imgsz, imgsz) for b in batch_sizes]
|
51 |
+
results = profile(img, model, n=3, device=device)
|
52 |
except Exception as e:
|
53 |
LOGGER.warning(f'{prefix}{e}')
|
54 |
|
55 |
+
# Fit a solution
|
56 |
+
y = [x[2] for x in results if x] # memory [2]
|
57 |
+
p = np.polyfit(batch_sizes[:len(y)], y, deg=1) # first degree polynomial fit
|
58 |
b = int((f * fraction - p[1]) / p[0]) # y intercept (optimal batch size)
|
59 |
+
if None in results: # some sizes failed
|
60 |
+
i = results.index(None) # first fail index
|
61 |
+
if b >= batch_sizes[i]: # y intercept above failure point
|
62 |
+
b = batch_sizes[max(i - 1, 0)] # select prior safe point
|
63 |
+
|
64 |
+
fraction = np.polyval(p, b) / t # actual fraction predicted
|
65 |
+
LOGGER.info(emojis(f'{prefix}Using batch-size {b} for {d} {t * fraction:.2f}G/{t:.2f}G ({fraction * 100:.0f}%) ✅'))
|
66 |
return b
|