Intel Neural Compute Stick 2 / Movidius AI Test
** ** * ***** ***** ** video ************ **** ************* ****** *********** attention *******'* ******** ****** *****. ******,******** *** ********* **** ***** **** generation ** ** *********** *** *** ****** *.
*** *** **** ** ***** ****** X ***** *******?
**** *** **** *** ***** ******* of *** ****** * **** (*** the ****) *** ****** ******* ****** this ******.
Intel / ******** / ******* ******* ***** ********
******* **** ******** **** *** ***** / ******** ***** ****** **** ** deep ******** / ********** ************. ** the **** ** ** ***** ** a ****** * ****. ***** **** be **** **** ** **-***** ****, the **** **** *** **** ******** models *** ** **** ****** *************** of ******, *******, *****, ***.
**** * **** ***,***** ************ ****** ********** ** '****** ******* ******' (***). ***** *** ***** *** had *** ******** / ***** ******** Myriad * ****, **** *** *** has *** *********** *******, ***** ***** ****** "******** *** the *********** ******** ** ******** ***********".
**** ****** *** **** *** ****** it.
Areas ********
****** **** ******, ** ***** ******* across:
- **** ********* ***** *****'* ****-*********-******-**** *****
- ******* ****** ***** **, ****** * and ****** *
- ***, *** *** *** *********** ********* across *******
- ******** ********** ** *********
- **** ******* *** ******
Key ********
*** *** ******** **** *** ******* are:
- Myriad * ~= *.** ****** * ***: The NCS2 ran roughly 2.7x frames-per-second more than the NCS1 in their face detection model we tested. Intel says that different optimizations and measurement changes would increase this significantly (discussed in caveats below).
- Faster **** *** (**): The NCS2 / Myriad X outperformed the i7-7700K (a top-of-the-line in 2017) chip tested here, a significant advance given the Myriad X being significantly lower cost and lower power consumption.
- Little/no ******** ****: Our tests noted no differences in output video accuracy in line's feedback that OpenVINO's model-optimization step should be considered "lossless." They pointed out that allowing lossy model compression (when converting a model to be run on Myriad; eg FP32->FP16 data-type downsizing) would cause issues with vendors training a model with certain expectations, and having a different (lossy) model for inferring.
- Model ************* ** ******: While Intel has mentioned YOLO support for OpenVino, the current YOLOv3 is not yet compatible on Myriad chips due to a particular network layer. ***** **************** ***** *** ******* ********** ****** to ** *** ** ***, *** YOLOv3 ** *** *** ********** ** Myriad ******* *** ** * ********** network *****. **** ** & **: Intel ******* ***** *** ** *** on ******, *** *** ***** *******.**********+***: The ***** ** **** ****** **** variety, ************ ***** ****-*********-******-**** (********** ***** with * ******-**** ********).*********+***: *** ***** pending. ** **** ******* ****** ****** from ****** ** *****'* *** ******.
*** ****** ***** ******** *********** *******:
Caveat ** *** / *****
***********, *****'* *** ******* *** **** higher **** *** ***, *** **** reasons. **, *****'* ****-*********-******-**** ***** ******* 250 *** ** ***, **.* ** Myriad2 (**** ***** ******** ********** ******* numbers ** **** ******).
* *** ****** ** *** ***** Intel ******* ***: (*) ** *** Python, **** ******* **** ***** *++ code; (*) *** ***-******* ******** ******* code (** ****** *******/******* ******), *** isolating ***** ****; (*) ** *** not ***** ******* ************ ******** ** video ***** *******, ***** ** ***** into ********.
Video **********
***** ** * ***** ********** **** reviews *** *******:
Questions / ******** / **** ** **** ****?
** **** ******* ******* ** *** NCS2 / ****** * ***** **** other ********* *** **** **** ****** over *** **** *** *****. **** would *** **** ** *** ******?
*** **** ****** ** **** * USB **** **** *** ***** *** device. ** * **********, ***** ****** Compute ****** ****'* ****** *** ********** applications, *** *** **********/******* **** ****** network / ******* ** *** ****** chip (** *** *** ** ********, ensure *** **********, ***). ***'** ** using ** ********** ****** **** ** the **** *****.
**** **** ** *** *** **** in *** **** ***** (*'** *** about ******* ** ****-****** ******* ** it), *** **'* **** **** **** for ***. ***'** ** ****** ****** cloning ********'* *** ****** *++ *****. You *** **** ***** (***** *** install ********) ** ~/*****/*******************/****************/****************/*******. ** ** "are **** **** ****** / ********", my ******** ***** ** "***" - **** want ** *** *** *******. *** you'll **** ** **** *** ******** license (**** ** *******) ** ~/*****/*******************/*********.
**** ****** *** ****...
******* ********, *** ** ******* $ or $$ ** $$$?
**** *** $***. ********** **********, *** ****'** *** ****-***** *********
****** **** **** **** **** **-*****. I ******* *** * **** *** and ** ** ******** ******.
**** ******** - **** ***% *** utilization ** *** **-***** **** **** each ********* *** ***** *** ***** with **** ** ** ** ** it ******** ******* **** ***** ******?
****, *** *** *** *** **** get ****** *******? ** ***** * significant ****** ** ******* ** *** stick? **** ********* **** *** ****** will ** ** *** ******* *******/****.
*** **** * ** ***, *** I'm ************* ******. * ***** **'* as *** ******, ** ********* ***% full *********** ** * **** *** of **'* * ***** (** ********** aggregate ****** ****). ** ************ *'* thinking ** **** *.**/* ***** *****.
* ***** *** ***** *** *** felt **, ** ********** ****. *'** try ** ***** ** * ****** way ** *******. **** ***** ** a * *** *****, *********** ******* of ****** * **** (** ***** rather **** *****) ******** ** **** common ****.
* ***** **'* ** *** ******, it ********* ***% **** *********** ** 1 **** *** ** **'* * cores (** ********** ********* ****** ****).
** **** ** ****** *** ****, I ***** ******* **** *** ****** need ** ** ********** ****** ********, e.g. ***%=**%.
****** **** ******** *** ***%, ***** the * *** ** *****? ** it **** ***** ******** **** **** 1?
* *** **** ******* ******* ** the ******* *******. ***** ** ** Intel *** ***** *** ********* *** TDP(Thermal ****** *****). **’* **** *** draw *** ****. * ** *** certain *** ** ******** ** *** TDP ** * ******** **** **** on ******* ** ** ******* *** industry ***** ** ******* **** **** as **** ** ********... ** *** not **** ****. * *****’* **** been **** ** **** ******* ** thermal **** ** **** ****** *****. Here ** *** ****** * ******** to ***** ****** **********:
****** * - *.* **** ***
**-***** - ** **** ***
**** ********* - *-** **** ***
****** ***** - ** ** *** watt *** is ** *** *** **** ******** uses *** ****** **** ********** ****** analytics *********
* *****’* **** **** ** **** any *******/******* **** *** ****** ********** like ********** ** *****. ***** ***** to ******* ********** **** ***. * suspect **** *** *** * **** but ******** **** ******* ***** ** thermal ***********.
***, ** *** ***** ** ** 101 / ************ ******** ** ** / **** ******** ** *** **** future ** ****.
** *** **** ******* ** ** attention ***, ********** **, **** *** not **** **** ***** / ******** described ****. ******** ********* **** ******* that.
****** ************ ******** **** **** ** understand *** ******* ******.
* ****** * **** ** ***** to **** **** **** ****** ** the ****** ****** ************ *** *** able ** **** **** **** **** the *******. ** ******* * *** openVino **** *** ***** *** *** OS **** *** ********* *** ***** at ***(*** ******). ******* ** ********** it *** ** ****** *** *** as **** ** ******.
(** *** ** * *** ****, the "***" ****** ** **** *** intel ***'* *** ******, ***.)
* ** ************* ** *++, **** getting *******, *** *** ******* **** pretty ****.
* **** **** ******** *** ******* and ******* **** ** *** ******* line ** *** **** *********** ** like. ** ** **** ** ********** the *** *** ****** **** **** other *** *** ****** **** *** on ***********. *** *** ******* ***** video *** ****** ** ********* **** is *** **** *** ******* ***** detection ***** ** *** **** *** license ***** ***********. **** **** * distinction ******* ********* *** ***********.
** ***, ****** ************ **** ***********, but ** ** ****** ***** ** even ***** ***.
*'* **** ** ********** **** **** so * *** ****. ** *** "plug" **** ***** **** * ******** and *** * ******* *** *** camera ** ***** *********** *******? *** these ******** **** ****** ** ******** for *********?