vald-onnx-ingress-filter is one of the official ingress filter components provided by Vald.
Its custom logic requires the input of the ONNX model as a request and outputs the result from the ONNX model as the request of the Vald Agent.
Using this component lets users vectorize various data such as text and images using the ONNX model only inside the Vald cluster without external APIs.
git clone https://github.com/vdaas/vald-onnx-ingress-filter.git
kubectl apply -f vald-onnx-ingress-filter/k8s
NOTE: The example manifest files use ResNet50-v2 from ONNX Model Zoo as the ONNX model.
You can change the model by editing k8s/deployment.yaml.
...
- |
curl -L "https://github.com/onnx/models/raw/main/vision/classification/resnet/model/resnet50-v2-7.onnx" -o /model/sample.onnx #FIXME
Please edit the example/helm/values.yaml in the Vald repository to make vald-filter-gateway available and use it for deployment.
git clone https://github.com/vdaas/vald.git
cd vald
vim example/helm/values.yaml
---
...
gateway:
...
filter:
enabled: true
...
agent:
ngt:
dimension: 1000
After editing, let’s try to deploy the Vald cluster by the helm install command.
helm install vald vald/vald --values example/helm/values.yaml
Sample code with vald-client-python
We use a random vector as an example of the insert object.
import grpc
import numpy as np
from vald.v1.payload import payload_pb2
from vald.v1.vald import (
filter_pb2_grpc,
search_pb2_grpc,
)
channel = grpc.insecure_channel("localhost:8081")
// Insert
stub = filter_pb2_grpc.FilterStub(channel)
sample = np.random.random((1, 3, 224, 224)).astype(np.float32)
resize_vector = payload_pb2.Object.ReshapeVector(
object=sample.tobytes(),
shape=[1, 3, 224, 224],
)
resize_vector = resize_vector.SerializeToString()
req = payload_pb2.Insert.ObjectRequest(
object=payload_pb2.Object.Blob(
id="0",
object=resize_vector
),
config=payload_pb2.Insert.Config(skip_strict_exist_check=False),
vectorizer=payload_pb2.Filter.Target(
host="vald-onnx-ingress-filter",
port=8081,
)
)
stub.InsertObject(req)
// Search
sstub = search_pb2_grpc.SearchStub(channel)
scfg = payload_pb2.Search.Config(
num=10, radius=-1.0, epsilon=0.01, timeout=3000000000
)
sstub.Search(
payload_pb2.Search.Request(
vector=np.zeros((1000)),
config=scfg
)
)