I don't think I've ever skipped a single second in any of the videos and I always enjoy your in between talks about personal experiences. Really love how authentic this tutorial feels. Thanks a bunch!
On one hand I don't mind a little isolation because I have plenty of McWhortertainment to occupy my time however I do notice that some of the teens and twentysomethings that I know are having a rough time adjusting. I take for granted sometimes that I have a little wisdom and experience to help me process events like this but they don't quite yet. This time I will digress...as for this lesson I can see that this is a critical part of speeding things by lightening the processing load on the Nano. Another outstanding video...thanks again! I went down a small Charlton Heston rabbit hole and then had to find out what Soylent Green was.
You may have done this in your later videos but for the unknown images have the program crop the unknown face and save as a file for later processing. This is something I'm hoping can easily be done as I want my future robot to tell me when it finds something that is unknow to it. Doing this I can give it a hit and from there with deep learning have it train itself on all new unknown objects.
Where did this magical j=o variable come from? and why? i even rewatched the last lesson thinking i missed something... figured it would be used as an indexing value, its never used, copied the code from your website to ensure i didnt miss anything in my code, please explain. Thanks again for another great video!
Very good lesson Paul but with all the talk of pickles I need a grilled cheese with bacon sandwich. Enjoyed the lesson and I prefer being alone get more done. Have a great week.
Hi Paul,Hmay I make a suggestion for a small modification : I have tried with some photos on which no face was recognized and tis made the script stop with an error. So i made the following change : if (len(face_recognition.face_encodings(person))) > 0: encoding=face_recognition.face_encodings(person)[0] Encodings.append(encoding) Names.append(name) else : print("no person found in image " + path) I hope this may help someone else as well
Hello Paul, Really great video lessons on the jetson nano !!!! I have been following your lessons all along but haven't purchased a jetson nano yet. I was looking for the link for lesson 1 on the gear you use and where to buy it. I found lesson 1 but no link to the gear? Also should I go for a jetson xavier ns as shown in your review or just purchase the original jetson nano. There is also a B01 version of the jetson nano that has two camera ports like the xavier. Will you be switching to the xavier any time soon? Will all your lessons and future lessons be compatible with the xavier nx as it may have a different version of jetpack than you are running? The jetson xavier seems like a better way to go for someone that is looking to buy hardware for the first time to follow your lesson series? Your thoughts are much appreciated and again, very awesome series of lessons !!!!!! Rick
Hello Paul, For those that are buying jetson gear for the first time, would you recommend the jetson nano or the jetson xavier nx? Thanks much Paul....Rick
If you can afford it, I would get the Xavier NX. See links in my Xavier NX review. All the lessons in this Jetson Nano tutorial would be what you need to learn the Xavier. It is more expensive, but in my mind worth it.
Actually I love all fermented foods. Real pickles, kefir, sauerkraut, fermented meats, yogurt, kambuchu and so forth. My kitchen is a veritable fermentation chamber.
Hello Paul, I checked the version of jetson nano on amazon that you are using (V02) and it is now $140.00. The newer BO1 two camera version is $108.00. I guess this is to try and get you to buy the newer version? For your current jetson series and future jetson series is it best to go with the current version of the jetson nano (A02) that you are using or go with the jetson xavier nx?
The newer version is almost identical. However, the camSet string in all my lessons and programs will need an extra parameter to make it work. If you did in and figure it out, then the new board will work fine. If you want to follow through, and have exactly what I did work, then get the board I have. If you get the new board, you have to include a parameter in the camSet string to show which camera you want to use.
Will the SD card and all the contents created so far on a Nano work in a Xavier straight off or would one have to do all the installs stuff and then copy the created files?
You need to flash a new SD card with the jetpack for Xavier NX. You have to then go in and install code-oss and face recognition library and that sort of stuff. Then copy the programs over on a thumb drive. They should all work, just way faster.
Hi,Short lesson but very useful info.Just saw your comments re jetpack 4.4. I'm running on 4.4 mainly because I wasn't able to get a back-up of the 4.3 based SD card so I created a new image on a new SD card which was based on jet pack4.4. Have you been able to copy an image from a 128Gb SD card ? I used win32DiskImager but had problems.Haven't had a problem so far with 4.4 running your tutorials …..I have just got OpenCV running on a Raspberry pi 4 B rev1.2 and have run most of your tutorials on it successfully. I haven't done a back to back speed comparison but it seems to be pretty quick.I want to evaluate the Raspi 4 for College use for basic image processing and AI use. Its cheaper (£42 for the board) and has all the WiFi stuff built in.By the way, totally agree with you re Covid isolation, theres are a lot of plus's! I'm more concerned the more we open up in the UK. Saw someone mention LIDAR Any thoughts in that direction?Thanks again for the excellent content
I have been able to back up SD card using a Linux machine, and standard dd comand. However, I was unable to create a bootable backup of Jetpack 4.4. One of the many problems I had with jetpack 4.4. The programs will run on Jetpack 4.4, but the face recognizer runs painfully slow. Really recommend backing up and using 4.3.
I was running these tutorials on a Pi4 (4gb) but gave up at face detection. Whilst the Pi did complete the task(s) it was far too slow compared to the Jetson Nano.
@@paulmcwhorter Thanks a lot, tutorial is Excellent. I'm not Sure if facedetection lib is using the shaders/CUDA cores of the Jetson, it might be the problem, as NVML shows no activity. Paul, in this tutorial you talked about the great performance of the recognition, even if we have trained a person with only a photo. How could we use more than 1 photo per person to train the model? Regards
I don't think I've ever skipped a single second in any of the videos and I always enjoy your in between talks about personal experiences. Really love how authentic this tutorial feels. Thanks a bunch!
On one hand I don't mind a little isolation because I have plenty of McWhortertainment to occupy my time however I do notice that some of the teens and twentysomethings that I know are having a rough time adjusting. I take for granted sometimes that I have a little wisdom and experience to help me process events like this but they don't quite yet. This time I will digress...as for this lesson I can see that this is a critical part of speeding things by lightening the processing load on the Nano. Another outstanding video...thanks again! I went down a small Charlton Heston rabbit hole and then had to find out what Soylent Green was.
Thank you so much Paul. I make it work at the end, it was a little bit hard with the cut and paste. Great video!
Hi Paul,
Yes, I am having fun. Looking forward to make use of the GPIO pins!
I am still going through your Arduino lesson and this is making me more excited. Thank you so much Paul.
Highly appreciated sir nicely explained, and on my way to the next video
All the best
You may have done this in your later videos but for the unknown images have the program crop the unknown face and save as a file for later processing. This is something I'm hoping can easily be done as I want my future robot to tell me when it finds something that is unknow to it. Doing this I can give it a hit and from there with deep learning have it train itself on all new unknown objects.
another amazing video thanks Paul
Where did this magical j=o variable come from? and why? i even rewatched the last lesson thinking i missed something... figured it would be used as an indexing value, its never used, copied the code from your website to ensure i didnt miss anything in my code, please explain. Thanks again for another great video!
I have the same problem. Have you been able to figure it out ?
Very good lesson Paul but with all the talk of pickles I need a grilled cheese with bacon sandwich. Enjoyed the lesson and I prefer being alone get more done. Have a great week.
Much awaited, great going Paul, count down begins :-)
Hi Paul,Hmay I make a suggestion for a small modification :
I have tried with some photos on which no face was recognized and tis made the script stop with an error. So i made the following change :
if (len(face_recognition.face_encodings(person))) > 0:
encoding=face_recognition.face_encodings(person)[0]
Encodings.append(encoding)
Names.append(name)
else :
print("no person found in image " + path)
I hope this may help someone else as well
Great! Thank you
Thank you as always!
Hello Paul,
Really great video lessons on the jetson nano !!!! I have been following your lessons all along but haven't purchased a jetson nano yet. I was looking for the link for lesson 1 on the gear you use and where to buy it. I found lesson 1 but no link to the gear? Also should I go for a jetson xavier ns as shown in your review or just purchase the original jetson nano. There is also a B01 version of the jetson nano that has two camera ports like the xavier. Will you be switching to the xavier any time soon? Will all your lessons and future lessons be compatible with the xavier nx as it may have a different version of jetpack than you are running? The jetson xavier seems like a better way to go for someone that is looking to buy hardware for the first time to follow your lesson series? Your thoughts are much appreciated and again, very awesome series of lessons !!!!!! Rick
Links are in the description of each video . . . you might have to click 'show more' to see the whole description.
Hello Paul,
For those that are buying jetson gear for the first time, would you recommend the jetson nano or the jetson xavier nx? Thanks much Paul....Rick
If you can afford it, I would get the Xavier NX. See links in my Xavier NX review. All the lessons in this Jetson Nano tutorial would be what you need to learn the Xavier. It is more expensive, but in my mind worth it.
great lesson
Fun lesson Paul. Thank you. Can we have some salami and rye bread with the pickle or do we just get red herring and bologna? :>)
Actually I love all fermented foods. Real pickles, kefir, sauerkraut, fermented meats, yogurt, kambuchu and so forth. My kitchen is a veritable fermentation chamber.
Hello Paul,
I checked the version of jetson nano on amazon that you are using (V02) and it is now $140.00. The newer BO1 two camera version is $108.00. I guess this is to try and get you to buy the newer version? For your current jetson series and future jetson series is it best to go with the current version of the jetson nano (A02) that you are using or go with the jetson xavier nx?
The newer version is almost identical. However, the camSet string in all my lessons and programs will need an extra parameter to make it work. If you did in and figure it out, then the new board will work fine. If you want to follow through, and have exactly what I did work, then get the board I have. If you get the new board, you have to include a parameter in the camSet string to show which camera you want to use.
Its the final countdown
Will the SD card and all the contents created so far on a Nano work in a Xavier straight off or would one have to do all the installs stuff and then copy the created files?
You need to flash a new SD card with the jetpack for Xavier NX. You have to then go in and install code-oss and face recognition library and that sort of stuff. Then copy the programs over on a thumb drive. They should all work, just way faster.
I worry 'bout your caffeine addiction ! (lol) but another great lesson anyway :) thanks
Thank you
Hi,Short lesson but very useful info.Just saw your comments re jetpack 4.4. I'm running on 4.4 mainly because I wasn't able to get a back-up of the 4.3 based SD card so I created a new image on a new SD card which was based on jet pack4.4. Have you been able to copy an image from a 128Gb SD card ? I used win32DiskImager but had problems.Haven't had a problem so far with 4.4 running your tutorials …..I have just got OpenCV running on a Raspberry pi 4 B rev1.2 and have run most of your tutorials on it successfully. I haven't done a back to back speed comparison but it seems to be pretty quick.I want to evaluate the Raspi 4 for College use for basic image processing and AI use. Its cheaper (£42 for the board) and has all the WiFi stuff built in.By the way, totally agree with you re Covid isolation, theres are a lot of plus's! I'm more concerned the more we open up in the UK. Saw someone mention LIDAR Any thoughts in that direction?Thanks again for the excellent content
I have been able to back up SD card using a Linux machine, and standard dd comand. However, I was unable to create a bootable backup of Jetpack 4.4. One of the many problems I had with jetpack 4.4. The programs will run on Jetpack 4.4, but the face recognizer runs painfully slow. Really recommend backing up and using 4.3.
Thanks Paul, 4.4 seemed to run Ok with lesson 41 but I still have my 4.3 based SD card just in case
I was running these tutorials on a Pi4 (4gb) but gave up at face detection. Whilst the Pi did complete the task(s) it was far too slow compared to the Jetson Nano.
@@paulmcwhorter Thanks a lot, tutorial is Excellent. I'm not Sure if facedetection lib is using the shaders/CUDA cores of the Jetson, it might be the problem, as NVML shows no activity.
Paul, in this tutorial you talked about the great performance of the recognition, even if we have trained a person with only a photo. How could we use more than 1 photo per person to train the model? Regards
@@carlosrm8091 I was wondering the same thing. Hopefully we learn about this soon.
After pickling Trump was not mis-identified as PMcW but as an unknown person. Different!
It was hard one..
waiting....
Thank you
Thank you