Comments (8)
Did you try passing device /dev/apex_0 into the docker container? I think it's the --devices
flag on docker or this in docker compose. As well, most of the time it doesn't work on the first try, it seems to load some driver to the TPU the first time you try and the next time it will work. It only happens the first time you try it.
devices:
- /dev/apex_0
from hassio-addons.
Thank you for your answer.
Could you tell me how can I pass the --devices flag when launching the container as Home Assistant add-on?
It is not so clear to me where I can add the flag.
Should I uninstall doods from the add-ons and start itvia cli?
Thanks again
from hassio-addons.
I don't totally remember how to update those settings. It's been a long time since I've used that. If you can run it manually from the command line using the --devices
flag and prove that it fixes it, we can see about how to update the repo. It might take editing the addon repo.
from hassio-addons.
I just looked at this again... You should not need the --devices flag. It's included in the Add-On configuration. When starting the container it should automatically pass through the USB device. FWIW, I just pushed an update to the hass-io repo. Reload the REPOs and then install DOODS2. It's a total re-write and works way better. It shoudl solve all your issues. Re-Open this issue if it persists.
from hassio-addons.
Hi @snowzach, first off thank you for doods2!
Hope I'm doing the right thing re-opening rather than creating a new ticket?
I'm currently running doods2 with a usb tpu on an old NUC with HA Core and your docker container.
It works perfectly.
I'm now trying to migrate to a newer NUC11 running HaOS on bare metal. I'm having problems getting doods2 to recognise a M.2 tpu.
I have tested both a M.2 Accelerator A+E key and a M.2 Accelerator with Dual Edge. Using DOODS2-amd64 add-on.
Both of them raise this error:
doods.doods - ERROR - Could not create detector tflite/edgeTPU: Could not load EdgeTPU
Config:
- name: edgeTPU
type: tflite
modelFile: /share/models/efficientdet_lite3_512_ptq_edgetpu.tflite
labelFile: /share/models/coco_labels_lite3.txt
hwAccel: true
Both tpu's seem to be recognised in HAOS:
[core-ssh dev]$ ls /dev/apex_0
/dev/apex_0
[core-ssh dev]$ lsmod | grep apex
apex 20480 0
gasket 110592 1 apex,[permanent]
I have not been able to verify tpu's with lspci cmd as HAOS does not allow it out of the box.
Are /dev/apex_0 (M.2) devices truly supported by this HAOS add-on?
Or is this a HA Core issue with wrong Apex/Gasket drivers?
Or is it simply that my NUC does not support these M.2 E-key tpu's (but then I would not expect /dev/apex_0 to show up).
Any hints/help would be apprciated!
Regards, Lars
from hassio-addons.
Okay... I see the issue.. We need to pass the apex device through to DOODS... I made a change to the repo... You may need to remove and re-add the doods repo or hit refresh for it to work.
from hassio-addons.
Apologies for the delay, travel is getting in the way. No luck I'm afraid. Addon will not start when I add my detector:
- name: edgeTPU
type: tflite
modelFile: /share/models/efficientdet_lite3_512_ptq_edgetpu.tflite
labelFile: /share/models/coco_labels_lite3.txt
hwAccel: true
Log is completely empty (debug). Sorry...
from hassio-addons.
Try it now...
from hassio-addons.
Related Issues (18)
- Icon missin on Supervisor Dashboard HOT 9
- "Repostitory structure for master is not compliant" Error HOT 6
- You should be able to set the config in there with HOT 9
- passing USB Coral to container via Hassio
- Add on needs way to pass dev/bus/usb for EdgeTPU HOT 11
- Unable to start the add-on HOT 1
- Exception: 'TensorflowLite' object has no attribute 'input_data' HOT 2
- 404 stream not found , but working with your proxy HOT 4
- Missing repository.yaml HOT 2
- Share folder mapping problems HOT 2
- custom model HOT 3
- DOODS does not start on boot HOT 3
- Memory used by the addon
- Error while setting up doods platform for image_processing HOT 4
- Port change to 8181 HOT 2
- Please add a cuda add-on ;) HOT 4
- Deprecated format after upgrade to supervisor 2021.02.5 HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from hassio-addons.