# Google Coral Edge TPU
**URL:** https://community.openconversational.ai/t/google-coral-edge-tpu/8130
**Category:** Archive - Machine Learning
**Created:** [February 19, 2020, 11:08am UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130 "2020-02-19T11:08:15Z")
**Posts on this page:** 16
**Page:** 2
### Author:  [@StuartIanNaylor](https://community.openconversational.ai/u/StuartIanNaylor)
#### Post date: [May 24, 2020, 7:45pm UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/22 "2020-05-24T19:45:33Z")
@Dominik actually there is practically no difference to the parts that process models in a TPU to NPU.
As currently the defacto processing unit is Tensorcore the choice of interface has much to do if the model or current training step is loaded but both are batch to unit ram processes.
The USB sticks in comparison due to cost have a smaller compatibillity subset and register types mainly due to cost.
Google cloud TPUs are NPU based not GPU if you can accept the approximation of terminology as its the tensor cores of a GPU without the GPU stuff it doesn’t need.
---
### Author:  [@hobbyist](https://community.openconversational.ai/u/hobbyist)
#### Post date: [May 24, 2020, 8:15pm UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/23 "2020-05-24T20:15:59Z")
But TPU is an ASIC right? Another implementation of Digital Design instead of FPGA, right?
---
### Author:  [@Dominik](https://community.openconversational.ai/u/Dominik)
#### Post date: [May 24, 2020, 10:34pm UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/24 "2020-05-24T22:34:09Z")
> [@StuartIanNaylor](#):
>
> there is practically no difference to the parts that process models in a TPU to NPU.
> [@hobbyist](#):
>
> But TPU is an ASIC right? Another implementation of Digital Design instead of FPGA, right?
Yes, there is a lot of marketing speech involved here. To my understanding the Rockchip-NPU is FPGA and supports “reprogramming” while Coral Edge-TPU is “hardcoded” ASIC.
---
### Author:  [@hobbyist](https://community.openconversational.ai/u/hobbyist)
#### Post date: [May 24, 2020, 11:15pm UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/25 "2020-05-24T23:15:44Z")
So, that’s why Google Coral Dev is faster…I got it!
---
### Author:  [@StuartIanNaylor](https://community.openconversational.ai/u/StuartIanNaylor)
#### Post date: [May 24, 2020, 11:20pm UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/26 "2020-05-24T23:20:23Z")
> [@Dominik](#):
>
> Rockchip-NPU
Dunno where you get FPGA for the rk1808 but they are all much the same apart from the GPU part of a GPU 🙂
[http://opensource.rock-chips.com/images/4/43/Rockchip\_RK1808\_Datasheet\_V1.2\_20190527.pdf](http://opensource.rock-chips.com/images/4/43/Rockchip_RK1808_Datasheet_V1.2_20190527.pdf)
They are all just accelerators with GPUs (nvidia) having the most compatibility due to nvidia’s head start with cuda.
---
### Author:  [@StuartIanNaylor](https://community.openconversational.ai/u/StuartIanNaylor)
#### Post date: [May 24, 2020, 11:26pm UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/27 "2020-05-24T23:26:54Z")
I think the rockchip is somewhere between the 2.
You can hire a Cloud TPU v2 for $4.50 / TPU hour $1.35 / TPU hour on preempt (ie kicked if someone wants to pay $4.50 until spare)
Which is 11.5 petaflops and for training it doesn’t seem to make sense to purchase what you can get for hardware.
Most common voice models run happily at \> relatime on a cpu but training can be an almighty endurance chore.
But you can hire server space and have it returned in minutes what will take hours with even some condiserable hardware.
But the little sticks are quite impressive seen as a RTX2080 is about 12Tflops and the vision models are pretty excellent but thankfully voice streams are much slower than visual ones.
---
### Author:  [@StuartIanNaylor](https://community.openconversational.ai/u/StuartIanNaylor)
#### Post date: [May 27, 2020, 5:01am UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/28 "2020-05-27T05:01:52Z")
Arm are also in the action with the Ethos-N78 neural processing unit 1-10 tops.
> **[Arm Announces Cortex-A78 CPU, Mali-G78 GPU, Ethos-N78 NPU and Custom...](https://www.cnx-software.com/2020/05/27/arm-announces-cortex-a78-cpu-mali-g78-gpu-ethos-n78-npu-and-custom-cortex-x1-core/)**
>
> Arm has just announced its 2020 Arm Mobile IP portfolio with no less than five IP blocks including Arm Cortex-A78 CPU, Arm Mali-G78 and G68 GPUs, Arm
Guess they will be coming thick and fast and they will just be as standard as current tensor core GPU.
> **[Arm Introduces Ethos NPU Family, Mali-G57 GPU, and Mali-D37 Display Processor](https://www.cnx-software.com/2019/10/23/arm-ethos-npu-family-mali-g57-gpu-and-d37-display-processor/)**
>
> Arm has just announced four new IP solutions with the most interesting being Ethos NPU (Neural Processing Unit) family with both Ethos-N57 and Ethos-N37
---
### Author:  [@hobbyist](https://community.openconversational.ai/u/hobbyist)
#### Post date: [July 13, 2020, 3:02pm UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/29 "2020-07-13T15:02:19Z")
I try to install the following packages
```
pandas
numpy
scikit-learn
scikit-image
click tqdm
```
in Google Coral TPU, in order to work with tensorflow lite, and I get these errors:
```
File "numpy/core/setup.py", line 422, in generate_config_h
moredefs, ignored = c6wfs --no-warn-script-location --no-binary :none: --only-binary :none: -i https://pypi.org/simple -- setuptools wheel Cython>=0.29.13 "numpy==1.13.3; python_version=='3.6' and platform_system!='AIX'" "numpy==1.14.5; python_version=='3.7' and platform_system!='AIX'" "numpy==1.17.3; python_version>='3.8' and platform_system!='AIX'" "numpy==1.16.0; python_version=='3.6' and platform_system=='AIX'" "numpy==1.16.0; python_version=='3.7' and platform_system=='AIX'" "numpy==1.17.3; python_version>='3.8' and platform_system=='AIX'"" failed with error code 1 in None
```
Any ideas??
---
### Author:  [@StuartIanNaylor](https://community.openconversational.ai/u/StuartIanNaylor)
#### Post date: [July 14, 2020, 4:06am UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/30 "2020-07-14T04:06:29Z")
> [@hobbyist](#):
>
> Google Coral TPU
> **[Documentation | Coral](https://coral.ai/docs/)**
I think you have to delegate to libedgetpu.so.1 otherwise you are just running normal.
But as I said above “No as they are all very similar with different compatibility issues that you will have to research yourself.”
Don’t think many if any are using them here and likely to find better answers elsewhere.
> Add the delegate when constructing the `Interpreter` .
>
> For example, your TensorFlow Lite code will ordinarily have a line like this:
>
> interpreter = tflite.Interpreter(model\_path)
>
> So change it to this:
>
> interpreter = tflite.Interpreter(model\_path, experimental\_delegates=[tflite.load\_delegate(‘libedgetpu.so.1’)])
>
> The file passed to `load_delegate()` is the Edge TPU runtime library, and you installed it when you first set up your device. The filename you must use here depends on your host operating system, as follows:
---
### Author:  [@hobbyist](https://community.openconversational.ai/u/hobbyist)
#### Post date: [July 14, 2020, 9:08am UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/31 "2020-07-14T09:08:17Z")
You wrote at the end “as follows:” but I do not see the rest of the sentence. Do I miss something?
---
### Author:  [@hobbyist](https://community.openconversational.ai/u/hobbyist)
#### Post date: [July 14, 2020, 9:08am UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/32 "2020-07-14T09:08:52Z")
You wrote at the end “as follows:” but I do not see the rest of the sentence. Do I miss something?
---
### Author:  [@StuartIanNaylor](https://community.openconversational.ai/u/StuartIanNaylor)
#### Post date: [July 14, 2020, 2:48pm UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/33 "2020-07-14T14:48:17Z")
Yes it would seem the docs.
> **[Run inference with TensorFlow Lite in Python | Coral](https://coral.ai/docs/edgetpu/tflite-python/#load-tensorflow-lite-and-run-an-inference)**
>
> How to use the Python TensorFlow Lite API to perform inference with Coral devices
---
### Author:  [@hobbyist](https://community.openconversational.ai/u/hobbyist)
#### Post date: [July 19, 2020, 10:58am UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/34 "2020-07-19T10:58:52Z")
I have ran a CNN model on Google Coral TPU Dev Board, but I am not sure it does accelerate the model… Is there a command to use in order to see specifically the ASIC load? And more important the time duration that tensorflow lite accelerates the CNN model?
---
### Author:  [@StuartIanNaylor](https://community.openconversational.ai/u/StuartIanNaylor)
#### Post date: [July 19, 2020, 11:37am UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/35 "2020-07-19T11:37:01Z")
> [@StuartIanNaylor](#):
>
> interpreter = tflite.Interpreter(model\_path, experimental\_delegates=[tflite.load\_delegate(‘libedgetpu.so.1’)])
I dunno but at a guess if
> interpreter = tflite.Interpreter(model\_path, experimental\_delegates=[tflite.load\_delegate(‘libedgetpu.so.1’)])
Is included in your code then it should use the tpu
if
interpreter = tflite.Interpreter(model\_path)
Its just going to use cpu.
So try testing your code and see if it makes a difference.
You need to find a Google Coral TPU group or forum as really no-one is using them here.
---
### Author:  [@muema111](https://community.openconversational.ai/u/muema111)
#### Post date: [November 20, 2022, 9:21am UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/36 "2022-11-20T09:21:37Z")
I was wondering if there is open source to blur faces, license blades and company labeling on vehicles on-the-fly in video streams?
This would make the stream GDPR compliant as long as not recording private property.
Ideal would be if an additional feature would allow to black out private property if a camera is fixed mounted at one position.
If there is such software, how performant would the Coral EDGE TPU be? 1 or more streams? What resolution and frame rate?
---
### Author:  [@StuartIanNaylor](https://community.openconversational.ai/u/StuartIanNaylor)
#### Post date: [November 20, 2022, 11:10am UTC](https://community.openconversational.ai/t/google-coral-edge-tpu/8130/37 "2022-11-20T11:10:52Z")
[https://docs.frigate.video/hardware/](https://docs.frigate.video/hardware/) is prob the one to check as build around a coral edge.
> A single Coral can handle many cameras and will be sufficient for the majority of users. You can calculate the maximum performance of your Coral based on the inference speed reported by Frigate. With an inference speed of 10, your Coral will top out at `1000/10=100` , or 100 frames per second. If your detection fps is regularly getting close to that, you should first consider tuning motion masks. If those are already properly configured, a second Coral may be needed.
A pi though comes at the bottom of the list as the base board.
A Coral m.2 is half the price of with maybe the dual e key one being best value even though have never tried.
GDPR is more about using, informing and retention of data and stoping the misuse of that data and sharing.
If you comply correctly then thoses faces, licence blades and company labeling are not a problem, but yeah there are ML segmentation models to capture those so likely the same could be used to obscure.
If you do some googling and find the perfect int8 model for edge ml then performance is extremely impressive.
> **[Insights](https://www.raccoons.be/resources/insights/)**
>
> Over the years, we have experimented with new technologies and in doing so, we have gathered a lot of insights. We want to share them with you.
[Previous page](https://community.openconversational.ai/t/google-coral-edge-tpu/8130.md?page=1)