DepthaiExperiments gen2-facemesh

hdnzngr

There seems to be an issue with the luxonis artifactory.

Running the gen2 facemesh sample

luxonis/depthai-experimentsblob/master/gen2-facemesh/main.py

throws an exception:

404 Client Error: Not Found for url: https://artifacts.luxonis.com/artifactory/blobconverter-backup/blobs/facemesh_192x192_openvino_None_6shave.blob

requests.exceptions.HTTPError: 400 Client Error: BAD REQUEST for url: https://blobconverter.luxonis.com/compile?version=2021.4&no_cache=False During handling of the above exception, another exception occurred: File "D:\source\repos\depthai-experiments\gen2-facemesh\main.py", line 61, in <module> facemesh_nn = oak.create_nn('facemesh_192x192', input=det_nn) requests.exceptions.HTTPError: 404 Client Error: Not Found for url: https://artifacts.luxonis.com/artifactory/blobconverter-backup/blobs/facemesh_192x192_openvino_None_6shave.blob

The model can also not be found using manual search on artifactory.

jakaskerl

Hi hdnzngr
Here is the model:
luxonis/depthai-model-zootree/main/models/facemesh_192x192

Also, it looks like it's working on the SDK version:

/Users/jaka/Desktop/depthai-experiments/gen2-fac
emesh/main.py
Downloading /Users/jaka/.cache/blobconverter/facemesh_192x192_openvino_2022.1_6shave.blob...
[==================================================]
Done

Thanks,
Jaka

hdnzngr

jakaskerl

This was no model, it was only the model definitions.

I have

depthai.version == 2.22.2.0
depthai_sdk.__version == 1.12.0 (because the version in requirements.sdk cannot be pip installed, due to
ERROR: No matching distribution found for depthai-sdk==1.10.2)

This is what I get when running

import blobconverter
model_path = blobconverter.from_zoo(name="facemesh_192x192",
zoo_type="depthai",
shaves=6)
print(model_path)

-->

Downloading C:\Users\xxxx\.cache\blobconverter\facemesh_192x192_openvino_2021.4_6shave.blob...

{

"exit_code": 1,

"message": "Command failed with exit code 1, command: /opt/intel/openvino2021_4/deployment_tools/inference_engine/lib/intel64/myriad_compile -m /tmp/blobconverter/8ed514d614034f1bbd6b4341d38cd66e/facemesh_192x192/FP16/facemesh_192x192.xml -o /tmp/blobconverter/8ed514d614034f1bbd6b4341d38cd66e/facemesh_192x192/FP16/facemesh_192x192.blob -c /tmp/blobconverter/8ed514d614034f1bbd6b4341d38cd66e/myriad_compile_config.txt -ip U8",

"stderr": "**Unknown model format! Cannot find reader for model format: xml** and read the model: /tmp/blobconverter/8ed514d614034f1bbd6b4341d38cd66e/facemesh_192x192/FP16/facemesh_192x192.xml. Please check that reader library exists in your PATH.\\n",

"stdout": "Inference Engine: \\n\\tIE version ......... 2021.4.2\\n\\tBuild ........... 2021.4.2-3974-e2a469a3450-releases/2021/4\\n"

}

Conversion failed due to 400 Client Error: BAD REQUEST for url: https://blobconverter.luxonis.com/compile?version=2021.4&no_cache=False

Trying to find backup... (model="facemesh_192x192", shaves="6", version="2021.4")

Unable to fetch model from backup server due to: 404 Client Error: Not Found for url: https://artifacts.luxonis.com/artifactory/blobconverter-backup/blobs/facemesh_192x192_openvino_2021.4_6shave.blob

jakaskerl

Hi hdnzngr
I'm using blobconverter==1.4.2. The code you pasted..

hdnzngr import blobconverter
model_path = blobconverter.from_zoo(name="facemesh_192x192",
zoo_type="depthai",
shaves=6)
print(model_path)

..works on my side.

facemesh_nn = oak.create_nn('facemesh_192x192', input=det_nn) this should automatically find the model and download it.

nn_component.py - DepthaiSDK:

models = getSupportedModels(printModels=False)
            zoo_models = blobconverter.zoo_list()
            print(zoo_models)
            if str(model) in models:
                model = models[str(model)] / 'config.json'
                self._parse_config(model)
            elif str(model) in zoo_models:
                logging.warning(
                    'Models from the OpenVINO Model Zoo do not carry any metadata'
                    ' (e.g., label map, decoding logic). Please keep this in mind when using models from Zoo.'
                )
                self._blob = dai.OpenVINO.Blob(blobconverter.from_zoo(str(model), shaves=6))
                self._forced_version = self._blob.version
            else:
                raise ValueError(f"Specified model '{str(model)}' is not supported by DepthAI SDK.\n"
                                 "Check SDK documentation page to see which models are supported.")

Thanks,
Jaka