ECE 495Autonomous Vehicles

67% liked it36% easy68% useful

From 9 ratings and 7 comments on UW Flow. The most helpful comments, quoted as written, without names. Read them all on UW Flow.

• The course expects us to use google colab for the machine learning assignments, however they do not provide any form of support for additional compute resources. Once your free GPU credits run out that's it. Its ridiculous that a course that I'm paying thousands in tuition for cannot provide the minimum resources for an effective learning experience. It's unreal that I'm expected to keep some notebook open for days because I ran out of GPU credits and am forced to train my network using CPU.
  The bare minimum they can do is provide a version of the assignment that can be completed from the schools ubuntu servers so that I can at least use GPU compute.
  Beyond that, half the assignments are…

  Feb 2023 · didn’t like it · 5 found this helpfulOn UW FlowRemove this

• This is the most cobbled together course I have ever taken in my 5 years at uw.
  I've only done two assignments so far and both were just straight up copied from other sources (udacity, Stanford cs231). This itself would've been fine, but what irks me is that there was absolutely no effort put in to even change the slightest thing about the assignments. A2 in this class is A1 in the Stanford cs course, but the prof did not even bother to change the folder name from "assignment 1" from "assignment 2". And all the other useless files were left in the folder (since we only needed part of the Stanford assignment).
  And the main challenge from A2 comes from getting a working environment, not…

  Jan 2023 · didn’t like it · 4 found this helpfulOn UW FlowRemove this

• Very fun and interesting course. It is mostly delivered in a flipped classroom style, with occasional guest lectures that showcase the latest developments in the AV world.
  There is definitely a lot of material to get through, and the professor himself said he would've preferred this to be split into two separate courses.
  The good thing is, if you are not interested, you do not really have to go through all the content/can skim it to get a decent grade, as the assignments and assessments are very clear in what parts they cover.
  Overall fun course

  May 2025 · liked itOn UW FlowRemove this

• Pretty mid course. Lecture style is flipped classroom, where you watch/read the videos/slides yourself. Weekly assignments that are pretty easy worth 50%. Midterm and final were almost identical to practice versions. There is a lot of course content! 3 hour videos a week. But, as long as you keep up with content, given the easy exams and assignments, it's not hard to pull an 80+ in the course with little effort. I would summarize the course as breadth not depth. You cover many topics from classical image processing, to CNNs, to state models. If you're okay with learning all the content yourself, it is a fairly interesting course. Personally would have preferred a non-flipped classroom style…

  May 2023 · liked itOn UW FlowRemove this

• Course with a very broad range of topics in robotics and a very good starting point for variety of machine learning topics. If you have an interest in working in the field of autonomous vehicles or robotics, the course will certainly be valuable to you in terms of course content.
  In terms of the workload however, it is quite substantial especially given the variety of topics, which require some self-learning in order to fully grasp and appreciate (e.g. classical image processing, CNN's, Bayesian Filtering, etc.). If you're a fan of robotics and autonomous navigation I wouldn't let this aspect dissuade you from taking it, as the assignments taught me a significant number of transferable…

  Apr 2023 · liked itOn UW FlowRemove this