samyak
@Samyak1729
Followers
2,361
Following
588
Media
219
Statuses
1,693
Explore trending content on Musk Viewer
Merchan
• 139192 Tweets
Dick Cheney
• 99259 Tweets
Eagles
• 89568 Tweets
#CristinaMostraElTitulo
• 82309 Tweets
Franco Escamilla
• 58132 Tweets
Barcola
• 45751 Tweets
Fraternal Order of Police
• 43303 Tweets
Packers
• 42040 Tweets
Peacock
• 36316 Tweets
Barış Alper
• 36073 Tweets
#FRAITA
• 29351 Tweets
Galler - Türkiye
• 24718 Tweets
SixTONES冠番組
• 24123 Tweets
End of 1
• 22494 Tweets
Sergio Mendes
• 20340 Tweets
#الجوهره_للاهلي_ممثل_الوطن
• 19524 Tweets
Go Birds
• 18688 Tweets
Tonali
• 18187 Tweets
فرنسا
• 17470 Tweets
Deschamps
• 16286 Tweets
Montella
• 16196 Tweets
ايطاليا
• 14878 Tweets
Iker
• 14262 Tweets
Alan Dershowitz
• 14057 Tweets
Dembele
• 10760 Tweets
goal is to get financially stable then start programming for fun
this will be the first thing i touch
44
320
6K
this guy is pretty cool if you want to revisit anything in calculus
slow paced and intuitive
26
269
4K
video tutorials really give you an illusion that you are learning.
13
37
661
if you are just starting out. your goal should be being consistent and not fast progress.
fast progress will come as a byproduct of showing up daily
7
60
497
day 11 ML progress:
> 30% through the ML course by fcc
> learnt about ML basics, Classification/Regression KNN, Naive Bayes
> will implement the same in code today
10
8
311
planning on becoming cracked enough for an internship, if not more
what level of expertise and skillset would you desire for an ml intern?
i have to get better
8
8
289
> get on twitter
> get to hang around with cool ppl
> be nice
> learn from everyone
> life is good
9
21
278
day 24 ML progress
> not satisfied
> revisited linear regression and naive bayes
> overview of logistic regression, bias and variance
> bored. want to start with a mini project now
7
7
261
day 28,29 ML progress report:
> 60% through the micrograd lecture
> understood manual back-propagation and backward function parts
> will implement the same today
5
4
265
week 7
crazy how much my progress slowed down since uni started.
comeback arc from tomorrow
no chill from now on
2
0
28
7
10
233
if you ever feel like a nerd, remember i talked about karpathy, tpot and agi on my first date
13
3
233
day 7 ML progress report
>completed the pandas tutorial
>some practice with the same
10
5
215
day 22 ML progress report:
> pretty much stuck on implementation of backprop
> tried to learn about gradients (as in linear algebra)
> looked more into gradient descent, revisited 3b1b vids, more maths
> not satisfied actually, but we keep doing the bit
15
9
194
day 59 ML progress report
> learnt about linear classifiers and logistic regression in depth as I had skimmed through it when I studied it first time
> would spend some more time on it tomorrow
1
6
171
day 52 ML progress report
> tried to wrap my head around the matmul in neural net I'm building.
> my brain too small but eventually in an hour or so i figured out
> will maximize focus intensity as I should have got done more.
9
9
171
life cant get better
> learning what I am passionate about
> working on a business idea of myself by night
> have a super supporting gf by my side
> friends and family I love with all of my heart
cant tell how grateful I am
9
0
162
day 51 of ML progress report
> made a grave mistake of attending lectures (7 hours gone)
> spent rest of the time on MNIST classifier /w samson zhang vid
> def not satisfied with today
3
3
153
day 31:
> finally implemented the backward propagation part of micrograd. i feel so happy
> also understood how backprop works finally!
> would need to revisit top sort again tho
> just 2 more days I will be done with this
4
3
147
day 14 ML progress report
> reviewed oop concepts in python
> started with scikit learn tutorial as it was heavily used in the ML course
> 25-30% done with it (learnt about bag of words, splitting data)
3
1
140
day 38 ML progress report:
> continued writing code for MNIST classifier from scratch
> never appreciated it but pandas is a pretty library
> stuck on error right before I started training
> speed i need speed!!!
3
5
130
day 27 ML progress report
> learn about computation graphs and backward differentiation
> not much progress on micrograd
> was able to get it more than 2 hours despite tight schedule!
1
3
113
day 25 of ML progress report
> was busy. couldn't do much
> goal was to learn and implement logistic regression and svm but only covered some theory on log regression
> if time allows I will give it an overview but for now I start with micrograd
5
2
108
day 23 ML progress report:
> finally starting to wrap my head around backprop
> did quite some maths around the same (thank god for pen and paper)
> implemented the cost function in the neural_net
> I understood what I learnt yesterday but still left with so many whys!!
6
1
107
never realized but I had been copying code for so much time in the name of "coding along'
idk man if you do the same this is a gentle reminder skill issues can be fixed. fix them even if it takes time but do it now
7
0
94
50 days of machine learning
> it has been very fun. but a lot of things I could do better
> will be sure to make the next 50 days worth it
> thanks to everyone who followed. wont disappoint you
5
4
87
day 26 ML progress report
> started out with
@karpathy
's micrograd video
> not much done. did the part for core value object
> studied a little about correlation and regression as a revision to details for classical models
3
1
87
day 53 ml progress report:
> more progress on MNIST classifier
> it would be complete by tomorrow
> mega chore day, did lot of things and tired now
2
0
85
finally done with engine part of micrograd
tried not to exactly copy the code so yeah there might be some hidden bugs i haven't encountered yet
but hey it feels so great being unstuck
1
0
73
i doubt if my ml progress reports do any value addition at this point but i'll keep doing it until I get an internship
might work as a reference for someone starting out later
after that it'll be only posts with high value information
yours truly :)
5
0
57
wanted to look up some old notes and seems notion is down rn
this shit just started a chain of thoughts of how a company has so much control over "my" data and has all the capabilities to cut me off from it
i understand open course folks now
11
3
57
turns out i don't know much of python
everyday we learn something new
4
0
53
day 65 ml progress report
>spent some time trying to understand matmul in NNs
>studied about cross validation
>super slow progress on MNIST classifier. can't believe I'm still on it
>reviewed previous code
1
0
51
grateful that people are carrying the work grant started. one person can only do so much.
imagine the rich resources we'll accumulate as the time goes
2
3
49
day 48 ML progress report
> couldn't complete today's goals
> continued coding micrograd. an error again. no, claude inst helping
> good start but dipped later. will try to fix that tomorrow
2
0
48
day 37 ML progress report:
> building a MNIST classifier from scratch in numpy
> revisited underlying maths and did basic setup (including backprop function)
> given the time I had I could have completed this. more focus on increasing speed from now
3
2
45
Day 61, 62
>weekly revision + covering blind spots in learning
>grasped a high level understanding of entropy
>revisited random variable and probability distribution
1
1
42
day 38 ML progress thread
> this thread will contain what I got done for every focused time block
> just an experiment. will continue this if i think it improves my productivity otherwise notebook ftw
3
0
42
man i don't remember vector spaces. I learnt that thing just last year
im kinda dumb ig
2
0
39
day 9 ML progress report
> completed matplotlib tutorial
> did some practice with the same
2
0
37
major win for all of us!
⚡️ Excited to share that I am starting an AI+Education company called Eureka Labs.
The announcement:
---
We are Eureka Labs and we are building a new kind of school that is AI native.
How can we approach an ideal experience for learning something new? For example, in the case
2K
4K
28K
2
0
35
day 39 ML progress report:
> turns out my understanding of NN is very high level thus implementing from scratch is going to take time
> will be revisiting NN basics and maths for it
> sticking to the weekly goals, started learning about CNNs
4
1
35
day 57 ML progress report
> delved into entropy and cross entropy just cuz of the loss func
> added loss function to MNIST classifer
> studying for a test that's tomorrow
1
0
35
am i the only who sometime skims the through the lecture just to get confidence on what content i'll be encountering?
4
0
33
learning ml starting today.
alongside there will
> daily logs
> blog posts
6
0
32
day 12 of ml progress
>was a slop day (bad mood couldn't get shit done)
>revisited data prep part the same course i am following
> will revise today
2
1
30
day 19 of ML progress:
> cleared my understanding of linear regression, gradient descent
> trying to implement a neural network from scratch
2
1
31
day 49
> moved forward like 10 mins in micrograd lec for all the time that I got in clg
> working on college lab work they need it tomorrow
> was at hospital for a frnd for quite some time. read nearcyan blog and went through few sections of
@elliotarledge
's cuda course
3
0
30
crazy how much my progress slowed down since uni started.
comeback arc from tomorrow
no chill from now on
2
0
28
Day 60
>was outside due to some work for most of the time
>spent rest of the time on a leetcode question (search auto complete system)
>sadly no ML today :(
1
0
29
day 73 ml progress
> have taken up a project for a hackathon involving CNN and GAN
> updates will be back starting tomorrow
> was busy with college so couldn't keep track of what happened from day 65-72
4
1
28
got basic understanding for scikit-learn, will rely on documentation for rest
now back to the ML course.
0
0
27
it takes no time before a person starts blaming his situations for his skill issues
too many people blaming their college for not giving them enough time and I definitely don't wanna be one of them
work hard anon and don't stop. stopping is death
1
1
28
trying to build an intuition for matmul for NNs, like how to place features and weights to get what we want
and it actually gets complicated for large examples
below is my attempt to teach myself through hit and trial
if it helps i'll try to make a vid explaining the same
0
0
29
day 47 ml progress report
> facing errors on micrograd so much I decided to re-write everything
> one major thing to improve, check twice or thrice to know you got everything right. don't move forward without that. you'll face problems later
2
1
28
day 49 of ML progress report
> busy day barely got any time today
> getting an error on the loss func part. fked around. also asked angkul. apparently he had faced the same problem.
>expecting tomorrow to be good
1
0
27
day 31 ML progress report
> implemented backward function that need to to called manually on individual nodes in micrograd
> will be speeding up things from today
4
0
26
day 1
> watched numpy tut
time left to become cracked : [▓▓▓▓▓░░░░░░░░░░░░░░] 24.68% completed
1
0
27
one thing i've noticed is non cs people don't care if it's a gpt-wrapper
they'll use it if it gets their job done
2
0
26
in micrograd vid at 1:25:21. sensei talks about a bug that we might have introduced. and how we have to accumulate the gradients to avoid overwriting them
then he directed us to multivariable case of chain rule. which I skipped back then
1
0
26
day 35 ML report
> wasn't feeling well
> will try to get back today
4
0
24
day 43,44:
> i am sick again ;-;
> couldn't do much. spent most of the time trying to fix errors in micrograd. i must be really dumb.
> fixed the error that i was stuck on since a week (it was just one line of code)
> still the computational graph isn't rendering completely
3
0
25
day 15 of ML progress report
> reviewed a scikit-learn tutorial
> imp points include fit, transform, splitting data, vectorizer
> i'll code the same today
> will be spending rest of the day on buildspace project
2
0
25
i had actually skimmed over classical ML models.
now will spend next 2 days solidifying my understanding about the same + plus implementation
1
0
25
day 13 ML progress
> revised what I had learnt (it was a good day)
> good idea to keep one day purely for revision
> gave the amazon ML summer school test, didn't go too well
> looking forward to next fellowships/programme
2
1
25
going to build the whole site for fashaa. maybe broken but it will be up by this week.
man i'd really be speed running MERN in a week
2
1
24
i couldn't get anything done in last 3 days here's what i'll be fixing
> stop setting daily goals so hard that you don't even start
> stay curious else you wont wanna sit down and study
> go out and touch grass every once in a while
back on track now lfg🔥
0
0
24
gm!
last few days were quite slow in terms of learning
will try to put in deliberate efforts to improve speed
note: no distraction from phone or anything. i just had slow speed for some reason
2
0
24
sorry to keep you guys off the loop
had taken a break from social media overall
back to postmaxxing. will try to build something instead of "study study"
3
1
23
day 8 ML progress report
>halfway through the matplotlib course
> my terminal broke for some reason, will spend some time fixing it today
2
0
23
day 18 ML progress report!
> started learning about neural nets
> watched 3b1b vids, did some reading from michael nielsen's book
> learnt about perceptron, sigmoid neuron and gradient descent
excited to learn more!
2
0
20
@arpitingle
the goodbye scene gets me everytime
also the music is too good
still surprised abt how tf did a white guy produce music that sounds so authentically indian
1
0
20
day 33 ml progress thread
> will be updated with every study session
> lets see how it goes
1
0
20
day 20 ML progress report
> still building my neural net
> not satisfied with today's progress, college work has started sadly
> will try to make up for it today!!
3
1
20
day 21 ML progress report:
> watched welch labs neural_nets playlist
> would spend more time doing things on paper then implementing it in code
> not satisfied with the progress.
2
0
18
day 63,64 of ml progress
>been busy with one thing or other lately
>whatever time I got today was taken away by a college assignment
>will be back on track from thursday :(
0
0
19
will come back to micrograd implementation again in a few days with a fresh mind
or will give it one hour every day
feeling stuck right now
3
0
18