Use offline models with depthai_sdk

tarocal

Hey,

I have connected an OAK device to a device that cannot have internet access. So, in order to experiment with gen2-people-tracker, I've downloaded the model into */.cache/blobconvertor/modelname_openvino2021.4_6shave.blob. I've also places the hash created and corresponding model location on the .config.json file.

But the oak.createnn( *line tries to access the internet.

Is this expected?
What is the work around?

Information:
depthai*_sdk version : 1.9.1.1
python: 3.10
OS: Windows 10

Thanks for any information/update.

jakaskerl

Hi tarocal
Update depthai_sdk version to the latest please. You can use the .json file or .blob file of the downloaded model when creating the nn. Just specify the path to it.

Thanks,
Jaka

tarocal

Thanks for the reply but when I downloaded the model, directly pointed the code at it

tracker = oak.create_nn("./mypc/modelname_openvino2021.4_6shave.blob", colorcam, tracker=True)

like above, but I get the error that tracker is not enabled.

Exception: Tracker was not enabled! Enable with cam.create_nn("[model]", tracker=True)!*
*
Updating depthai_sdk is also difficult, because the TrackerPacket module is not in the latest version on depthai_sdk.
Sorry for bothering with the experimental code, but I thought you might need to know the error I mentioned with the tracker not being enabled.

jakaskerl

Hi @tarocal

#!/usr/bin/env python3

from depthai_sdk import OakCamera
import depthai as dai
from depthai_sdk.classes import TrackerPacket
from people_tracker import PeopleTracker
from depthai_sdk.visualize.configs import TextPosition

import cv2

pt = PeopleTracker()

with OakCamera(replay='people-tracking-above-03') as oak:
    color_cam = oak.create_camera('color')
    tracker = oak.create_nn('person-detection-retail-0013', color_cam, tracker=True)
    tracker.config_nn(conf_threshold=0.6)
    tracker.config_tracker(tracker_type=dai.TrackerType.ZERO_TERM_COLOR_HISTOGRAM, track_labels=[1])

    def cb(packet: TrackerPacket):
        visualizer = packet.visualizer
        left, right, up, down = pt.calculate_tracklet_movement(packet.daiTracklets)

        visualizer.add_text(f"Up: {up}, Down: {down}", position=TextPosition.TOP_LEFT, size=1)
        visualizer.draw(packet.frame)

        cv2.imshow('People Tracker', packet.frame)

    oak.visualize(tracker.out.tracker, callback=cb)
    oak.start(blocking=True)

Thanks,
Jaka

DarshitDesai

jakaskerl Hi I am also facing the same issue, I made the yolov8 blob and when I try to import it using oak.create_nn() it gives me this error

My version of depth ai SDK

pi@luxonis:~/Roving-Comforter $ pip show depthai-sdk
Name: depthai-sdk
Version: 1.12.1
Summary: This package provides an abstraction of the DepthAI API library.
Home-page: https://github.com/luxonis/depthai/tree/main/depthai_sdk
Author: Luxonis
Author-email: support@luxonis.com
License: MIT
Location: /home/pi/.local/lib/python3.9/site-packages
Requires: blobconverter, depthai, depthai-pipeline-graph, marshmallow, numpy, opencv-contrib-python, pytube, PyTurboJPEG, sentry-sdk, xmltodict
Required-by: 
pi@luxonis:~/Roving-Comforter $

[2024-04-22 21:48:17] INFO [root.__init__:147] Setting IR laser dot projector brightness to 800mA
[2024-04-22 21:48:17] INFO [root.__exit__:408] Closing OAK camera
Traceback (most recent call last):
  File "/home/pi/Roving-Comforter/testrun.py", line 132, in <module>
    nn = oak.create_nn('/home/pi/Roving-Comforter/yolov8n_coco_640x352_openvino_2022.1_6shave.blob', color, tracker=True, spatial=True)
  File "/home/pi/.local/lib/python3.9/site-packages/depthai_sdk/oak_camera.py", line 253, in create_nn
    comp = NNComponent(self._oak.device,
  File "/home/pi/.local/lib/python3.9/site-packages/depthai_sdk/components/nn_component.py", line 215, in __init__
    self._spatial.depth.link(self.node.inputDepth)
AttributeError: 'depthai.node.NeuralNetwork' object has no attribute 'inputDepth'
Sentry is attempting to send 2 pending error messages
Waiting up to 2 seconds
Press Ctrl-C to quit

jakaskerl

https://discuss.luxonis.com/d/3354-downloading-stock-yolov8n-blob-from-luxonis-website/6