starred/ProxmoxVE

Fork 0

mirror of https://github.com/community-scripts/ProxmoxVE.git synced 2026-04-25 17:26:03 +03:00

[GH-ISSUE #457] Frigate installer failing during install #96

New issue

Closed

opened 2026-02-26 12:44:23 +03:00 by kerem · 8 comments

kerem commented

2026-02-26 12:44:23 +03:00

Owner

Originally created by @jappish84 on GitHub (Nov 23, 2024).
Original GitHub issue: https://github.com/community-scripts/ProxmoxVE/issues/457

Originally assigned to: @remz1337 on GitHub.

Please verify that you have read and understood the guidelines.

yes

A clear and concise description of the issue.

Tried installing a few times and keep getting errors during install on the same place. Seems to be related to openvino and ssdlite_mobilenet_v2

I am running an old server without AVX support, although SSE 4.2 should be supported. CPU info

8 x Intel(R) Xeon(R) CPU E5620 @ 2.40GHz (1 Socket)
--
Kernel Version Linux 6.8.12-4-pve (2024-11-06T15:04Z)
Boot Mode Legacy BIOS
Manager Version pve-manager/8.2.8/a577cfa684c7476d

My plan is to run frigate with USB Google Coral - no real need for iGPU support in my case. I have also tried installing with default settings.

I found a related issue but don't understand how it was solved

What settings are you currently utilizing?

Default Settings
Advanced Settings

Which Linux distribution are you employing?

Debian 11

If relevant, including screenshots or a code block can be helpful in clarifying the issue.

Beginning of script:

    ______     _             __
   / ____/____(_)___ _____ _/ /____
  / /_  / ___/ / __ `/ __ `/ __/ _ \
 / __/ / /  / / /_/ / /_/ / /_/  __/
/_/   /_/  /_/\__, /\__,_/\__/\___/
             /____/

Using Advanced Settings
Using Distribution: debian
Using debian Version: 11
Using Container Type: 0
Using Root Password: ********
Container ID: 102
Using Hostname: frigate
Using Disk Size: 20
Allocated Cores: 4
Allocated RAM: 8096
Using Bridge: vmbr0
Using IP Address: dhcp
Using Gateway IP Address: Default
Using APT-Cacher IP Address: Default
Disable IPv6: yes
Using Interface MTU Size: Default
Using DNS Search Domain: Host
Using DNS Server IP Address: Host
Using Vlan: Default
Enable Root SSH Access: yes
Enable Verbose Mode: yes
Creating a Frigate LXC using the above advanced settings
 ✓ Using local for Template Storage.
 ✓ Using local-lvm for Container Storage.
 ✓ Updated LXC Template List
 / Creating LXC Container

Last couple of rows with the error:

-dtypes, h5py, grpcio, google-pasta, gast, defusedxml, astunparse, absl-py, openvino-dev, markdown, keras, tensorboard, tensorflow
 /  Attempting uninstall: openvino
    Found existing installation: openvino 2024.1.0
 -    Uninstalling openvino-2024.1.0:
      Successfully uninstalled openvino-2024.1.0
 /Successfully installed absl-py-2.1.0 astunparse-1.6.3 defusedxml-0.7.1 gast-0.6.0 google-pasta-0.2.0 grpcio-1.68.0 h5py-3.12.1 keras-3.6.0 libclang-18.1.1 markdown-3.7 ml-dtypes-0.4.1 namex-0.0.8 networkx-3.1 openvino-2024.5.0 openvino-dev-2024.5.0 opt-einsum-3.4.0 optree-0.13.1 tensorboard-2.18.0 tensorboard-data-server-0.7.2 tensorflow-2.18.0 tensorflow-io-gcs-filesystem-0.37.1 termcolor-2.5.0
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager, possibly rendering your system unusable.It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv. Use the --root-user-action option if you know what you are doing and want to suppress this warning.
 \################|| Downloading ssdlite_mobilenet_v2 ||################

========== Downloading /opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09.tar.gz
... 100%, 49829 KB, 33511 KB/s, 1 seconds passedd
 \
========== Unpacking /opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09.tar.gz
 |
 |========== Converting ssdlite_mobilenet_v2 to IR (FP16)
Conversion command: /usr/bin/python3 -- /usr/local/bin/mo --framework=tf --output_dir=/opt/frigate/models/public/ssdlite_mobilenet_v2/FP16 --model_name=ssdlite_mobilenet_v2 --input=image_tensor --reverse_input_channels --output=detection_scores,detection_boxes,num_detections --transformations_config=/usr/local/lib/python3.9/dist-packages/openvino/tools/mo/front/tf/ssd_v2_support.json --tensorflow_object_detection_api_pipeline_config=/opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09/pipeline.config --input_model=/opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09/frozen_inference_graph.pb '--layout=image_tensor(NHWC)' '--input_shape=[1, 300, 300, 3]' --compress_to_fp16=True

 -[ INFO ] MO command line tool is considered as the legacy conversion API as of OpenVINO 2023.2 release.
In 2025.0 MO command line tool and openvino.tools.mo.convert_model() will be removed. Please use OpenVINO Model Converter (OVC) or openvino.convert_model(). OVC represents a lightweight alternative of MO and provides simplified model conversion API. 
Find more information about transition from MO to OVC at https://docs.openvino.ai/2023.2/openvino_docs_OV_Converter_UG_prepare_model_convert_model_MO_OVC_transition.html
 \FAILED:
ssdlite_mobilenet_v2
 -
[ERROR] in line 114: exit code 0: while executing command $STD /usr/local/bin/omz_converter --name ssdlite_mobilenet_v2 --precision FP16 --mo /usr/local/bin/mo
root@proxmox:~#

Please provide detailed steps to reproduce the issue.

running root@proxmox:~# bash -c "$(wget -qLO - https://github.com/community-scripts/ProxmoxVE/raw/main/ct/frigate.sh)" in the proxmox shell - error occurs with both default and advanced settings

Originally created by @jappish84 on GitHub (Nov 23, 2024). Original GitHub issue: https://github.com/community-scripts/ProxmoxVE/issues/457 Originally assigned to: @remz1337 on GitHub. ### Please verify that you have read and understood the guidelines. yes ### A clear and concise description of the issue. Tried installing a few times and keep getting errors during install on the same place. Seems to be related to `openvino` and `ssdlite_mobilenet_v2` I am running an old server without AVX support, although SSE 4.2 should be supported. [CPU info](https://www.techpowerup.com/cpu-specs/xeon-e5620.c933) ``` 8 x Intel(R) Xeon(R) CPU E5620 @ 2.40GHz (1 Socket) -- Kernel Version Linux 6.8.12-4-pve (2024-11-06T15:04Z) Boot Mode Legacy BIOS Manager Version pve-manager/8.2.8/a577cfa684c7476d ``` My plan is to run frigate with USB Google Coral - no real need for iGPU support in my case. I have also tried installing with default settings. I found a [related ](https://github.com/community-scripts/ProxmoxVE/issues/294) issue but don't understand how it was solved ### What settings are you currently utilizing? - [ ] Default Settings - [X] Advanced Settings ### Which Linux distribution are you employing? Debian 11 ### If relevant, including screenshots or a code block can be helpful in clarifying the issue. Beginning of script: ``` ______ _ __ / ____/____(_)___ _____ _/ /____ / /_ / ___/ / __ `/ __ `/ __/ _ \ / __/ / / / / /_/ / /_/ / /_/ __/ /_/ /_/ /_/\__, /\__,_/\__/\___/ /____/ Using Advanced Settings Using Distribution: debian Using debian Version: 11 Using Container Type: 0 Using Root Password: ******** Container ID: 102 Using Hostname: frigate Using Disk Size: 20 Allocated Cores: 4 Allocated RAM: 8096 Using Bridge: vmbr0 Using IP Address: dhcp Using Gateway IP Address: Default Using APT-Cacher IP Address: Default Disable IPv6: yes Using Interface MTU Size: Default Using DNS Search Domain: Host Using DNS Server IP Address: Host Using Vlan: Default Enable Root SSH Access: yes Enable Verbose Mode: yes Creating a Frigate LXC using the above advanced settings ✓ Using local for Template Storage. ✓ Using local-lvm for Container Storage. ✓ Updated LXC Template List / Creating LXC Container ``` Last couple of rows with the error: ``` -dtypes, h5py, grpcio, google-pasta, gast, defusedxml, astunparse, absl-py, openvino-dev, markdown, keras, tensorboard, tensorflow / Attempting uninstall: openvino Found existing installation: openvino 2024.1.0 - Uninstalling openvino-2024.1.0: Successfully uninstalled openvino-2024.1.0 /Successfully installed absl-py-2.1.0 astunparse-1.6.3 defusedxml-0.7.1 gast-0.6.0 google-pasta-0.2.0 grpcio-1.68.0 h5py-3.12.1 keras-3.6.0 libclang-18.1.1 markdown-3.7 ml-dtypes-0.4.1 namex-0.0.8 networkx-3.1 openvino-2024.5.0 openvino-dev-2024.5.0 opt-einsum-3.4.0 optree-0.13.1 tensorboard-2.18.0 tensorboard-data-server-0.7.2 tensorflow-2.18.0 tensorflow-io-gcs-filesystem-0.37.1 termcolor-2.5.0 WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager, possibly rendering your system unusable.It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv. Use the --root-user-action option if you know what you are doing and want to suppress this warning. \################|| Downloading ssdlite_mobilenet_v2 ||################ ========== Downloading /opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09.tar.gz ... 100%, 49829 KB, 33511 KB/s, 1 seconds passedd \ ========== Unpacking /opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09.tar.gz | |========== Converting ssdlite_mobilenet_v2 to IR (FP16) Conversion command: /usr/bin/python3 -- /usr/local/bin/mo --framework=tf --output_dir=/opt/frigate/models/public/ssdlite_mobilenet_v2/FP16 --model_name=ssdlite_mobilenet_v2 --input=image_tensor --reverse_input_channels --output=detection_scores,detection_boxes,num_detections --transformations_config=/usr/local/lib/python3.9/dist-packages/openvino/tools/mo/front/tf/ssd_v2_support.json --tensorflow_object_detection_api_pipeline_config=/opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09/pipeline.config --input_model=/opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09/frozen_inference_graph.pb '--layout=image_tensor(NHWC)' '--input_shape=[1, 300, 300, 3]' --compress_to_fp16=True -[ INFO ] MO command line tool is considered as the legacy conversion API as of OpenVINO 2023.2 release. In 2025.0 MO command line tool and openvino.tools.mo.convert_model() will be removed. Please use OpenVINO Model Converter (OVC) or openvino.convert_model(). OVC represents a lightweight alternative of MO and provides simplified model conversion API. Find more information about transition from MO to OVC at https://docs.openvino.ai/2023.2/openvino_docs_OV_Converter_UG_prepare_model_convert_model_MO_OVC_transition.html \FAILED: ssdlite_mobilenet_v2 - [ERROR] in line 114: exit code 0: while executing command $STD /usr/local/bin/omz_converter --name ssdlite_mobilenet_v2 --precision FP16 --mo /usr/local/bin/mo root@proxmox:~# ``` ### Please provide detailed steps to reproduce the issue. running `root@proxmox:~# bash -c "$(wget -qLO - https://github.com/community-scripts/ProxmoxVE/raw/main/ct/frigate.sh)"` in the proxmox shell - error occurs with both default and advanced settings

kerem closed this issue

2026-02-26 12:44:23 +03:00

kerem commented

2026-02-26 12:44:23 +03:00

Author

Owner

@ristomatti commented on GitHub (Nov 24, 2024):

I wonder if Frigate is too much of a moving target to support since the recommended installation method is via Docker? There's also still breaking changes on pretty much each major release. I've been running Frigate on Docker within a Debian LXC well over a year (originally installed using @tteck's "Docker LXC" script). Even with the nested virtualization, I've been surpised on how good the performance is.

@ristomatti commented on GitHub (Nov 24, 2024): I wonder if Frigate is too much of a moving target to support since the recommended installation method is via Docker? There's also still breaking changes on pretty much each major release. I've been running Frigate on Docker within a Debian LXC well over a year (originally installed using @tteck's "Docker LXC" script). Even with the nested virtualization, I've been surpised on how good the performance is.

kerem commented

2026-02-26 12:44:23 +03:00

Author

Owner

@MickLesk commented on GitHub (Nov 24, 2024):

I wonder if Frigate is too much of a moving target to support since the recommended installation method is via Docker? There's also still breaking changes on pretty much each major release. I've been running Frigate on Docker within a Debian LXC well over a year (originally installed using @tteck's "Docker LXC" script). Even with the nested virtualization, I've been surpised on how good the performance is.

I've often asked myself the same question, and this is by far the script that causes the most problems. tteck had already seen it that way and wanted to remove it several times, but then many users complained.

@MickLesk commented on GitHub (Nov 24, 2024): > I wonder if Frigate is too much of a moving target to support since the recommended installation method is via Docker? There's also still breaking changes on pretty much each major release. I've been running Frigate on Docker within a Debian LXC well over a year (originally installed using @tteck's "Docker LXC" script). Even with the nested virtualization, I've been surpised on how good the performance is. I've often asked myself the same question, and this is by far the script that causes the most problems. tteck had already seen it that way and wanted to remove it several times, but then many users complained.

kerem commented

2026-02-26 12:44:23 +03:00

Author

Owner

@jappish84 commented on GitHub (Nov 24, 2024):

I actually managed to get past this by forking the repo and commenting out the Openvino install part in frigate-install.sh
Suggestion to the maintainers is to make optional installs deselectable if possible without to much hassle

@jappish84 commented on GitHub (Nov 24, 2024): I actually managed to get past this by forking the repo and commenting out the Openvino install part in `frigate-install.sh` Suggestion to the maintainers is to make optional installs deselectable if possible without to much hassle

kerem commented

2026-02-26 12:44:23 +03:00

Author

Owner

@remz1337 commented on GitHub (Nov 24, 2024):

Give me a few days, I'll look into it

@remz1337 commented on GitHub (Nov 24, 2024): Give me a few days, I'll look into it

kerem commented

2026-02-26 12:44:23 +03:00

Author

Owner

@remz1337 commented on GitHub (Dec 16, 2024):

I'm suspecting it's the lack of AVX support. Unfortunately, I can't test further. I have a recent CPU and the default installation ran fine.

@jappish84 Can you log into the container and run this? please paste the detailed logs

cd /opt/frigate/models && /usr/local/bin/omz_converter --name ssdlite_mobilenet_v2 --precision FP16 --mo /usr/local/bin/mo

@remz1337 commented on GitHub (Dec 16, 2024): I'm suspecting it's the lack of AVX support. Unfortunately, I can't test further. I have a recent CPU and the default installation ran fine. @jappish84 Can you log into the container and run this? please paste the detailed logs ``` cd /opt/frigate/models && /usr/local/bin/omz_converter --name ssdlite_mobilenet_v2 --precision FP16 --mo /usr/local/bin/mo ```

kerem commented

2026-02-26 12:44:23 +03:00

Author

Owner

@Dino0811 commented on GitHub (Dec 18, 2024):

i got the same Error, CPU is Intel Celeron J4105 with integrated Intel® UHD Graphics 600 (Generation 9)

result of your code:

========== Converting ssdlite_mobilenet_v2 to IR (FP16)
Conversion command: /usr/bin/python3 -- /usr/local/bin/mo --framework=tf --output_dir=/opt/frigate/models/public/ssdlite_mobilenet_v2/FP16 --model_name=ssdlite_mobilenet_v2 --input=image_tensor --reverse_input_channels --output=detection_scores,detection_boxes,num_detections --transformations_config=/usr/local/lib/python3.9/dist-packages/openvino/tools/mo/front/tf/ssd_v2_support.json --tensorflow_object_detection_api_pipeline_config=/opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09/pipeline.config --input_model=/opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09/frozen_inference_graph.pb '--layout=image_tensor(NHWC)' '--input_shape=[1, 300, 300, 3]' --compress_to_fp16=True

[ INFO ] MO command line tool is considered as the legacy conversion API as of OpenVINO 2023.2 release.
In 2025.0 MO command line tool and openvino.tools.mo.convert_model() will be removed. Please use OpenVINO Model Converter (OVC) or openvino.convert_model(). OVC represents a lightweight alternative of MO and provides simplified model conversion API. 
Find more information about transition from MO to OVC at https://docs.openvino.ai/2023.2/openvino_docs_OV_Converter_UG_prepare_model_convert_model_MO_OVC_transition.html
FAILED:
ssdlite_mobilenet_v2

according to: https://www.techsavvy.de/prozessoren/intel_celeron_j4105/

Features
Turbo-Modus ✔
Hyperthreading / SMT ✘
Befehlssatz 64-bit
MMX ✔
SSE ✔
SSE2 ✔
SSE3 ✔
SSSE3 ✔
SSE4.1 ✔
SSE4.2 ✔
AVX ✘
AVX2 ✘
FMA3 ✔
FMA4 ✘
AES Unterstützung ✔
Virtualisierungstechnik ✔

@Dino0811 commented on GitHub (Dec 18, 2024): i got the same Error, CPU is Intel Celeron J4105 with integrated Intel® UHD Graphics 600 (Generation 9) result of your code: ```shell ========== Converting ssdlite_mobilenet_v2 to IR (FP16) Conversion command: /usr/bin/python3 -- /usr/local/bin/mo --framework=tf --output_dir=/opt/frigate/models/public/ssdlite_mobilenet_v2/FP16 --model_name=ssdlite_mobilenet_v2 --input=image_tensor --reverse_input_channels --output=detection_scores,detection_boxes,num_detections --transformations_config=/usr/local/lib/python3.9/dist-packages/openvino/tools/mo/front/tf/ssd_v2_support.json --tensorflow_object_detection_api_pipeline_config=/opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09/pipeline.config --input_model=/opt/frigate/models/public/ssdlite_mobilenet_v2/ssdlite_mobilenet_v2_coco_2018_05_09/frozen_inference_graph.pb '--layout=image_tensor(NHWC)' '--input_shape=[1, 300, 300, 3]' --compress_to_fp16=True [ INFO ] MO command line tool is considered as the legacy conversion API as of OpenVINO 2023.2 release. In 2025.0 MO command line tool and openvino.tools.mo.convert_model() will be removed. Please use OpenVINO Model Converter (OVC) or openvino.convert_model(). OVC represents a lightweight alternative of MO and provides simplified model conversion API. Find more information about transition from MO to OVC at https://docs.openvino.ai/2023.2/openvino_docs_OV_Converter_UG_prepare_model_convert_model_MO_OVC_transition.html FAILED: ssdlite_mobilenet_v2 ``` according to: https://www.techsavvy.de/prozessoren/intel_celeron_j4105/ Features [Turbo-Modus](https://www.techsavvy.de/prozessoren/technik/#cpu_turbotakt) ✔ [Hyperthreading / SMT](https://www.techsavvy.de/prozessoren/technik/#smt) ✘ [Befehlssatz](https://www.techsavvy.de/prozessoren/technik/#befehlssatz) 64-bit [MMX](https://www.techsavvy.de/prozessoren/technik/#mmx) ✔ [SSE](https://www.techsavvy.de/prozessoren/technik/#sse) ✔ [SSE2](https://www.techsavvy.de/prozessoren/technik/#sse2) ✔ [SSE3](https://www.techsavvy.de/prozessoren/technik/#sse3) ✔ [SSSE3](https://www.techsavvy.de/prozessoren/technik/#ssse3) ✔ [SSE4.1](https://www.techsavvy.de/prozessoren/technik/#sse41) ✔ [SSE4.2](https://www.techsavvy.de/prozessoren/technik/#sse42) ✔ [AVX](https://www.techsavvy.de/prozessoren/technik/#avx) ✘ [AVX2](https://www.techsavvy.de/prozessoren/technik/#avx2) ✘ [FMA3](https://www.techsavvy.de/prozessoren/technik/#fma3) ✔ [FMA4](https://www.techsavvy.de/prozessoren/technik/#fma4) ✘ [AES Unterstützung](https://www.techsavvy.de/prozessoren/technik/#aes) ✔ [Virtualisierungstechnik](https://www.techsavvy.de/prozessoren/technik/#virtualisierung) ✔

kerem commented

2026-02-26 12:44:23 +03:00

Author

Owner

@jappish84 commented on GitHub (Dec 18, 2024):

As mentioned, I forked the repo and excluded openvino - frigate then installed without errors.

root@frigate:~# cd /opt/frigate/models && /usr/local/bin/
omz_converter --name ssdlite_mobilenet_v2 --precision FP1
6 --mo /usr/local/bin/mo
-bash: /usr/local/bin/omz_converter: No such file or dire
ctory
root@frigate:/opt/frigate/models#

@Dino0811 you could use my fork if your cpu doesn't support AVX.

@remz1337 I could try to modify the script to prompt the user to choose whether or not to install openvino instead of just skipping it and create a PR, let me know if that's how you'd like to go around the issue.

Another approach is to better check CPUs compability maybe?

@jappish84 commented on GitHub (Dec 18, 2024): As mentioned, I [forked](https://github.com/jappish84/ProxmoxVE) the repo and excluded openvino - frigate then installed without errors. ``` root@frigate:~# cd /opt/frigate/models && /usr/local/bin/ omz_converter --name ssdlite_mobilenet_v2 --precision FP1 6 --mo /usr/local/bin/mo -bash: /usr/local/bin/omz_converter: No such file or dire ctory root@frigate:/opt/frigate/models# ``` @Dino0811 you could use my fork if your cpu doesn't support AVX. @remz1337 I could try to modify the script to prompt the user to choose whether or not to install openvino instead of just skipping it and create a PR, let me know if that's how you'd like to go around the issue. Another approach is to better check CPUs compability maybe?

kerem commented

2026-02-26 12:44:23 +03:00

Author

Owner

@remz1337 commented on GitHub (Dec 18, 2024):

I've removed the SSE 4.2 support. Should fix this issue.

@remz1337 commented on GitHub (Dec 18, 2024): I've removed the SSE 4.2 support. Should fix this issue.

kerem referenced this issue

2026-02-26 13:30:43 +03:00

[PR #96] [MERGED] Fix links #2590

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

starred/ProxmoxVE#96

No description provided.

Rows
Columns