File size: 3,370 Bytes
35199db e20f6b1 10021fb 35199db e20f6b1 35199db e20f6b1 35199db 10021fb 35199db 10021fb 35199db e20f6b1 35199db e20f6b1 35199db 35eb1b4 35199db 35eb1b4 35199db 35eb1b4 35199db e20f6b1 35eb1b4 35199db |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 |
import numpy as np
import time
import os, sys
from pathlib import Path
from concrete.ml.deployment import FHEModelClient
import requests
def to_json(python_object):
if isinstance(python_object, bytes):
return {"__class__": "bytes", "__value__": list(python_object)}
raise TypeError(repr(python_object) + " is not JSON serializable")
def from_json(python_object):
if "__class__" in python_object:
return bytes(python_object["__value__"])
# TODO: put the right link `API_URL` for your entry point
API_URL = "https://yw1dgyuig6ff5pft.us-east-1.aws.endpoints.huggingface.cloud"
headers = {
"Authorization": "Bearer " + os.environ.get("HF_TOKEN"),
"Content-Type": "application/octet-stream",
}
def query(payload):
response = requests.post(API_URL, headers=headers, data=payload)
return response.json()
path_to_model = Path("compiled_model")
# Decision-tree in FHE
from sklearn.datasets import fetch_openml
from sklearn.model_selection import train_test_split
import numpy
features, classes = fetch_openml(data_id=44, as_frame=False, cache=True, return_X_y=True)
classes = classes.astype(numpy.int64)
_, X_test, _, Y_test = train_test_split(
features,
classes,
test_size=0.15,
random_state=42,
)
NB_SAMPLES = 2
X_test = X_test[:NB_SAMPLES]
Y_test = Y_test[:NB_SAMPLES]
# Recover parameters for client side
fhemodel_client = FHEModelClient(path_to_model)
# Generate the keys
fhemodel_client.generate_private_and_evaluation_keys()
evaluation_keys = fhemodel_client.get_serialized_evaluation_keys()
# Test the handler
nb_good = 0
nb_samples = len(X_test)
verbose = False
time_start = time.time()
duration = 0
is_first = True
for i in range(nb_samples):
# Quantize the input and encrypt it
encrypted_inputs = fhemodel_client.quantize_encrypt_serialize(X_test[i].reshape(1, -1))
# print(f"Size of encrypted input {sys.getsizeof(encrypted_inputs)}")
# print(f"Size of keys {sys.getsizeof(evaluation_keys)}")
# Prepare the payload, including the evaluation keys which are needed server side
payload = {
"inputs": "fake",
# "encrypted_inputs": to_json(encrypted_inputs),
# "evaluation_keys": to_json(evaluation_keys),
"encrypted_inputs": encrypted_inputs,
"evaluation_keys": evaluation_keys,
}
print(f"{payload=}")
# Run the inference on HF servers
duration -= time.time()
print(f"Starting at {time.time()}")
encrypted_prediction = query(payload)
print(f"Ending at {time.time()}")
duration += time.time()
print(f"{encrypted_prediction=}")
encrypted_prediction = encrypted_prediction
if is_first:
is_first = False
print(f"Size of the payload: {sys.getsizeof(payload)} bytes")
# Decrypt the result and dequantize
prediction_proba = fhemodel_client.deserialize_decrypt_dequantize(encrypted_prediction)[0]
prediction = np.argmax(prediction_proba)
if verbose or True:
print(f"for {i}-th input, {prediction=} with expected {y_test[i]}")
# Measure accuracy
nb_good += y_test[i] == prediction
print(f"Accuracy on {nb_samples} samples is {nb_good * 1. / nb_samples}")
print(f"Total time: {time.time() - time_start} seconds")
print(f"Duration in inferences: {duration} seconds")
print(f"Duration per inference: {duration / nb_samples} seconds")
|