help Tom ! Week 2 # GRADED FUNCTION: update_parameters_with_adam
is this the same thing with a bad parentheis somewhere - and when you do the final step do u use “W” on both lines? parameters[“b” + str(l)] = parameters[“W” + str(l)] - (learning_rate * ( (v_corrected[“db” + str(l)] / (np.sqrt(s_corrected… more to it clearly - adding epsilon and closing parenthetical placement just trying to make the line that i think might be having trouble obvious - without posting too much
do u know of a good 1 to 1 online tutor ? zoom
Yes, it looks like you have a missing close paren or bracket on the previous line. If you can’t find a syntax error in python, it most likely means you are looking in the wrong place.
The notebook editor is syntax aware: just click on a paren or bracket and it will highlight the matching one. Or not!
Also note that once you fix that error, you will then get the wrong answer but it will look close. It’s not a rounding error: please check the math formula and note that \epsilon is in the denominator but not under the square root, right?
I recommend you NOT address your questions to a specific mentor.
Exactly. You’re better off just addressing the question to the community. There are lots of people who could potentially answer and no guarantee that any given person will be available at a particular point in time.
There is another “meta” point here: this is not the first time you have hit this mismatching parentheses issue or at least that’s my guess from looking at your recent posting history. It’s a mistake to blow off the programming part of this. Sure the ML/DL part is more interesting, but the programming is a fundamental part of it as well. If you don’t absorb the lessons as they go by, you can end up wasting a lot more time than you need to in order to get through everything here. As the old saying goes “Fool me once, shame on you. Fool me twice, shame on me!”
man im putting on a clinic of etiquette mistakes sorry about that - im having a rough day not even class related - ill think next time i vibed your were part of the class i shouldnt have assumed that and i should be open to help from anyone willing to weigh in - point taken
i concentrated on python. first over the beginning 6 months completed 8 class certs one capstone python one specialization for everybody then april to august trying things like gitlab linux the ibm ai specialization over august - - i appreciate what u r saying though - u need to speak the language to implement the ideas to then take it to a creative space even - i guess some of it hasn’t been translating - i love ML -i found the area i’m really excited about -i hope the programming part comes back i dont think im completely lost - hoping in time the ability to spot things like missing parenthesis or brackets or curly brackets will improve there is no dearth of alacrity for learning this - yet its a bit of a lonely road studying alone for a year plus now - i was thinking a tutor might not be a bad notion as opposed to erudition via video
Thanks for giving us more background on your learning situation. But in all that verbiage, I just want to make sure that you got the point I was making about using the notebook editor to check your syntax. Did you try the technique of just clicking on a parenthesis and watching what happens?
It may well be that many students would benefit from a more direct relationship with a tutor, even if it’s over a video conferencing setup. I don’t personally know of any such services, but it’s worth doing some web surfing to see if you can find someone who offers that. Of course that’s more of a time commitment to you than what you get here on Discourse, so it may well not be free. (Did I mention that the mentors here are volunteers: we do not get paid to do this.) Or perhaps you have friends or colleagues who are also in this kind of tech space and you could talk with them even in person.
wow i didnt know that - thats very generous to give of yourselves so freely - So --if no idea for a tutor - i’d fully expect to pay for that type of service -if i found a fit - Things are tight but im all in here so if i found a good teacher id have to find a way to compensate him/her - can you recommend another course perhaps take concurrently- that might tie in some of the programming esoterica specificallyl in Ml - Andrew Ng is a wonderful instructor he’s laying out the concepts here so thoroughly -the meat and potatoes are digestible but maybe slight disconnect to the programming aspects - its almost like ML programming has a bit or its own dialect - that is probably not accurate but the coding here is not an exact translation to what i’ve been taught thus far - i did use your punctuation technique and its great and handy so thanks for that - apologies for the long winded responses- ill truncate my messages from here - just thought id sort of introduce myself somewhat
right - thanks got it - wow these errors always seem so rudimentary in the rear view but while they are confounding you well its a challenge to put it euphemistically
Well, it’s a truism that “Hindsight is 20-20”. But exactly as you say, when you’re in the middle of the battle, it doesn’t seem so clear.
i cant be that terrible just made it through week 2 of course 2 - i know i have to take it to another level to be job ready but isn’t this supposed to be graduate level material -i saw a review that rated this specialization as such -
Congrats on finishing Week 2! Week 3 is also pretty interesting and contains your first introduction to TensorFlow which is a big deal.
It is true that graduate CS courses cover similar material to this, but I believe they present the material in a more advanced way. E.g. here Prof Ng soft pedals the math, in order to make the courses accessible to people who don’t even know univariate calculus, let alone the kind of matrix calculus you really need here. Stanford is actually very generous in putting materials from the graduate CS courses that are relevant to ML up on the web. You can take a look at CS 229 for example, which is the Stanford CS Grad Student course that Prof Ng has taught for quite a few years and which is the grad student version of the original Stanford Machine Learning course on Coursera. By comparing the two, you can get a sense for the differences in the level of the presentations.
Here’s a link to some notes from Stanford CS 231n, which covers Convolutional Networks as in DLS Course 4.
thanks for the help Paul - great name btw / i took Ap calculus so many moons ago i decided to take the math specialization with Lewis Serrano as well i’m putting in 3 hours a night 7 days a week - i’m furtively hopeful to get a job in this field now that i’ve settled on ML - but dare i say i yearn to be able to invent down the road and your point about comprehending the why behind the matrices and functions seems key to one day realizing this dream - it reminds of when i took on learning the guitar until i had mastery of the neck i had difficulty improvising and while i don’t really feel like i’m conscious of the notes i play when i take a solo it was not until i knew the notes so thoroughly that this ability availed itself to me (so much for abbreviated my messages but i wanted to thank you for your succor and for your philanthropy in general it inspires me and i’m certain countless others to redouble their efforts to be part of something bigger