Question: Remove default models or not?
MrSiO opened this issue · comments
This is a question concerning documentation. In one of your screenshots, it shows a list of models removed from the originals.
I found how to add your custom models reading this procedure: https://docs.deepstack.cc/custom-models/deployment/ (I'm using docker).
But this procedure seems to be adding custom models to the default ones and not removing the original ones... Am I right?
I wish to remove the original ones and only use yours.
Thx for any info!
This is a question concerning documentation. In one of your screenshots, it shows a list of models removed from the originals.
I found how to add your custom models reading this procedure: https://docs.deepstack.cc/custom-models/deployment/ (I'm using docker).
But this procedure seems to be adding custom models to the default ones and not removing the original ones... Am I right?
I wish to remove the original ones and only use yours.
Thx for any info!
What software are you using to send the image to deepstack
This is a question concerning documentation. In one of your screenshots, it shows a list of models removed from the originals.
I found how to add your custom models reading this procedure: https://docs.deepstack.cc/custom-models/deployment/ (I'm using docker).
But this procedure seems to be adding custom models to the default ones and not removing the original ones... Am I right?
I wish to remove the original ones and only use yours.
Thx for any info!
What software are you using to send the image to deepstack
I'm using iSpy-AgentDVR.
I'm using iSpy-AgentDVR.
I was looking at iSpy-AgentDVR instructions and it looks like you need to change the endpoint to "/v1/vision/custom/combined" if you what to use the combined model or "/v1/vision/custom/general" if you what to use the general model.
I'm using iSpy-AgentDVR.
I was looking at iSpy-AgentDVR instructions and it looks like you need to change the endpoint to "/v1/vision/custom/combined" if you what to use the combined model or "/v1/vision/custom/general" if you what to use the general model.
Great!! Will try that. Thx!!
I'm using iSpy-AgentDVR.
I was looking at iSpy-AgentDVR instructions and it looks like you need to change the endpoint to "/v1/vision/custom/combined" if you what to use the combined model or "/v1/vision/custom/general" if you what to use the general model.
Great!! Will try that. Thx!!
Also make sure you change the Object Classes to match the model you are using.
Combined Object Classes:
person,bicycle,car,motorcycle,bus,truck,bird,cat,dog,horse,sheep,cow,bear,deer,rabbit,raccoon,fox,skunk,squirrel,pig
General Object Classes:
person,vehicle
I'm using iSpy-AgentDVR.
I was looking at iSpy-AgentDVR instructions and it looks like you need to change the endpoint to "/v1/vision/custom/combined" if you what to use the combined model or "/v1/vision/custom/general" if you what to use the general model.
Great!! Will try that. Thx!!
Hi again!
I just want to validate something with you.
The endpoint in my iSpyDVR is in fact /v1/vision/detection
But I do not have that folder path on my deepstack docker host.
I have a mounted volume of:
volumes: - /home/user1/deepstack_data/Data/pytorch-models:/modelstore/detection
Is the /v1/vision/detection binded to /modelstore/detection ?
If so, I guess I should create a folder /modelstore/detection/custom/combined where I will put the combiend.pt file.
Then in iSpy, the endopoint would be /va/vision/????
I'm kinda confused about the iSpy endpoint config VS my volume on container.... :(
Thx for your lights.
I'm using iSpy-AgentDVR.
I was looking at iSpy-AgentDVR instructions and it looks like you need to change the endpoint to "/v1/vision/custom/combined" if you what to use the combined model or "/v1/vision/custom/general" if you what to use the general model.
Great!! Will try that. Thx!!
Hi again!
I just want to validate something with you.The endpoint in my iSpyDVR is in fact /v1/vision/detection
But I do not have that folder path on my deepstack docker host.
I have a mounted volume of:
volumes: - /home/user1/deepstack_data/Data/pytorch-models:/modelstore/detection
Is the /v1/vision/detection binded to /modelstore/detection ?If so, I guess I should create a folder /modelstore/detection/custom/combined where I will put the combiend.pt file.
Then in iSpy, the endopoint would be /va/vision/????I'm kinda confused about the iSpy endpoint config VS my volume on container.... :(
Thx for your lights.
Anyone able to clear this up for me?
Thx.
You might want to ask your question in the iSpy repo. I do not use this software. I am using Blue Iris so I am not 100% sure on what is need to use custom models with iSpy, sorry I could not help.
The developer is working on integrating with CodeProject.AI which has all my custom models preinstalled and is more up to date then DeepStack.
https://github.com/ispysoftware/iSpy
https://www.codeproject.com/Articles/5322557/CodeProject-AI-Server-AI-the-easy-way
https://www.codeproject.com/Articles/5352147/How-to-Setup-a-Custom-AI-Model-in-Agent-DVR-Using
So true. Thx.
I'll get back here with the infos for other users. Thx
So true. Thx. I'll get back here with the infos for other users. Thx
Thanks
Follow iSpy answers here for those interested: doitandbedone/ispyagentdvr-docker#545