@@KianBrose btw your "how to make advanced recognition bots in python" video was the first python tutorial for a project that I watched. Keep making good stuff brother 👍
Hey, ML researcher here, just a heads up, show your validation metrics instead of just loss. As loss is an arbitrary measure of a network’s ability and if it’s training loss should be decreasing due to gradient descent, regardless of the network’s applicability. Instead opt to show discernible metrics from your validation set, say for a classification model you can show the percent of validation inputs that were classified correctly that epoch of your model, if no viable metrics are available you can show the validation loss if needed but default to metrics. Overall great vid though, super cool!
Thanks for the tip! The graphs I showed were of the validation loss (I didn't make it clear by the titles) but I definitely should have referenced if the AI was able to classify major constituents.
Hi, I am a Student currently in grade 11(aiming for ivy leagues) and I wanna start model rocketry. I have a rasp on microcontrollers and can program them. can you suggest that from where should I learn the PID basics and thrust-vectoring.. Thanks!
If you are just getting started, go watch BPS Space's video "How to Build a Thrust Vectored Model Rocket", he explains PID controllers intuitively and that video is what made it click for me. If you don't feel confident with Linux and want something that can debug quickly, get an Arduino and use the Arduino PID library, it'll save you a lot of time compared to a Raspberry Pi Zero W (Which is what I used) but in the long term, I recommend you make the switch to Raspberry Pi's or something similar because their functionalities are great for model rockets. To learn how to make a Thrust Vector Control Mount (Get Fusion 360's Educational Access and start learning it), get inspired by people who have already made one, you don't have to copy it, but you get a good idea of the basic design.
@@mightbeamaury7721 Thanks! I have been watching bps space since a long time but I did not know about the video you mentioned. I think I might just start with the pico if I can understand!
@@keyaankhatri8468 Pico is a great starting point, also you will need to solder all your components on a board for the final version (But initially use jumper cables and breadboards to test sensors and circuitry before soldering everything), trust me it will save your sanity. I recommend using a perforated board or a PCB (If you are confident doing that).
I take the IB curriculum, the rest I learned from Textbooks, RUclips, or a LOT of trial and error. if you mean the major I'm applying for, it's Mechanical Engineering.
Ah yes yet another video recommendation to make me feel dumb
Damn bro. Didn't know you were still alive! Since you haven't uploaded on a long time and all that.
@Omnimistic lol
@@KianBrose btw your "how to make advanced recognition bots in python" video was the first python tutorial for a project that I watched. Keep making good stuff brother 👍
@@Omnimistic Ty!
Wth did not expect you to comment, I loved your vid on aimbots and cheating keep it up!
Another, MIT maker portfolio video??? Let me watch and rethink about my life.
Who cares dude. None of these ideas make the world better. Waste of time stuff.
I saw this at 78 clicks, I ll return again at 100.000, great stuff
Really cool, good job
Fantastic video Amaury! Your very proud grandad just sent me this 🙌
dude insane stuff!!
hoping we both get in :)))
Thanks man! Good luck and fingers crossed.
Hey, ML researcher here, just a heads up, show your validation metrics instead of just loss. As loss is an arbitrary measure of a network’s ability and if it’s training loss should be decreasing due to gradient descent, regardless of the network’s applicability. Instead opt to show discernible metrics from your validation set, say for a classification model you can show the percent of validation inputs that were classified correctly that epoch of your model, if no viable metrics are available you can show the validation loss if needed but default to metrics. Overall great vid though, super cool!
Thanks for the tip! The graphs I showed were of the validation loss (I didn't make it clear by the titles) but I definitely should have referenced if the AI was able to classify major constituents.
MIT Maker Portfolio | Amaury Patesson
3
Nice
191 views is a crime
ready for an internship at spacex
Hi, I am a Student currently in grade 11(aiming for ivy leagues) and I wanna start model rocketry. I have a rasp on microcontrollers and can program them. can you suggest that from where should I learn the PID basics and thrust-vectoring.. Thanks!
If you are just getting started, go watch BPS Space's video "How to Build a Thrust Vectored Model Rocket", he explains PID controllers intuitively and that video is what made it click for me. If you don't feel confident with Linux and want something that can debug quickly, get an Arduino and use the Arduino PID library, it'll save you a lot of time compared to a Raspberry Pi Zero W (Which is what I used) but in the long term, I recommend you make the switch to Raspberry Pi's or something similar because their functionalities are great for model rockets. To learn how to make a Thrust Vector Control Mount (Get Fusion 360's Educational Access and start learning it), get inspired by people who have already made one, you don't have to copy it, but you get a good idea of the basic design.
@@mightbeamaury7721 Thanks! I have been watching bps space since a long time but I did not know about the video you mentioned. I think I might just start with the pico if I can understand!
@@keyaankhatri8468 Pico is a great starting point, also you will need to solder all your components on a board for the final version (But initially use jumper cables and breadboards to test sensors and circuitry before soldering everything), trust me it will save your sanity. I recommend using a perforated board or a PCB (If you are confident doing that).
what type of study did you do ?
I take the IB curriculum, the rest I learned from Textbooks, RUclips, or a LOT of trial and error. if you mean the major I'm applying for, it's Mechanical Engineering.
Another MIT maker portfolio to make me feel dumb?! Noice 🥲👍