mirror of
https://github.com/hanzo-ml/kserve.git
synced 2026-10-30 20:53:49 +00:00
SivananthamandGitHub
a869657f25
* Fix python linter Signed-off-by: Sivanantham Chinnaiyan <[email protected]> * Format e2e tests Signed-off-by: Sivanantham Chinnaiyan <[email protected]> * Format hack, docs Signed-off-by: Sivanantham Chinnaiyan <[email protected]> * Add py-fmt target Signed-off-by: Sivanantham Chinnaiyan <[email protected]> * Update python lint workflow Signed-off-by: Sivanantham Chinnaiyan <[email protected]> * Rebase master Signed-off-by: Sivanantham Chinnaiyan <[email protected]> --------- Signed-off-by: Sivanantham Chinnaiyan <[email protected]>
70 lines
2.0 KiB
Python
70 lines
2.0 KiB
Python
import requests
|
|
import json
|
|
from matplotlib import pyplot as plt
|
|
import numpy as np
|
|
from aix360.datasets import MNISTDataset
|
|
import time
|
|
import sys
|
|
|
|
if len(sys.argv) < 3:
|
|
raise Exception("No endpoint specified. ")
|
|
endpoint = sys.argv[1]
|
|
headers = {"Host": sys.argv[2]}
|
|
|
|
data = MNISTDataset()
|
|
test_num = 349
|
|
is_file = False
|
|
if len(sys.argv) > 3:
|
|
try:
|
|
test_num = int(sys.argv[3])
|
|
except Exception: # pylint: disable=broad-except
|
|
is_file = True
|
|
|
|
if is_file:
|
|
inputs = open(sys.argv[3])
|
|
payload = json.load(inputs)
|
|
input_image = payload["instances"][0]
|
|
plt_image = np.array(input_image[0])
|
|
label = payload["instances"][1][0]
|
|
else:
|
|
plt_image = data.test_data[test_num]
|
|
input_image = np.array([data.test_data[test_num]])
|
|
input_label = data.test_labels[test_num]
|
|
input_image = input_image.tolist()
|
|
label = input_label.argmax()
|
|
payload = {"instances": [input_image, [input_label.argmax().item()]]}
|
|
|
|
x = time.time()
|
|
|
|
res = requests.post(endpoint, json=payload, headers=headers)
|
|
print("TIME TAKEN: ", time.time() - x)
|
|
|
|
print(res)
|
|
res_json = res.json()
|
|
|
|
adv_im = np.asarray(res_json["explanations"]["adversarial_example"])
|
|
adv_class = res_json["explanations"]["adversarial_prediction"]
|
|
image_class = res_json["explanations"]["prediction"]
|
|
|
|
is_diff = False
|
|
for x in range(0, len(input_image[0])):
|
|
for y in range(0, len(input_image[0][0])):
|
|
for z in range(0, len(input_image[0][0][0])):
|
|
if input_image[0][x][y][z] != adv_im[0][x][y][z]:
|
|
is_diff = True
|
|
|
|
if is_diff:
|
|
print("Found an adversarial example.")
|
|
fig0 = (plt_image[:, :, 0] + 0.5) * 255
|
|
fig_adv = (adv_im[0, :, :, 0] + 0.5) * 255
|
|
|
|
f, axarr = plt.subplots(1, 2, figsize=(10, 10))
|
|
axarr[0].set_title("Original (" + str(label) + ")")
|
|
axarr[1].set_title("Adversarial (" + str(adv_class) + ")")
|
|
|
|
axarr[0].imshow(fig0, cmap="gray")
|
|
axarr[1].imshow(fig_adv, cmap="gray")
|
|
plt.show()
|
|
else:
|
|
print("Unable to find an adversarial example.")
|