yeldarb

35 ms·

331.
by yeldarb 6y ago
But Facebook is publicly traded. If you truly think this is the case why not just put your $1000 in their shares and let it compound at the same rate as Zuck’s?
332.
by yeldarb 6y ago
This post is in response to the discussion[1] on Hacker News this week on the post "YOLOv5: State-of-the-art object detection at 140 FPS" and the objections raised by AlexeyAB on Github[2]. The controversy surrounded around two po
333.
by yeldarb 6y ago
Yeah, they made the most popular PyTorch implementation of YOLOv3 as well so they're not entering out of the blue, though. https://github.com/ultralytics/yolov3 The author of YOLOv3 quit working on Computer Vision
334.
by yeldarb 6y ago
I was very surprised how well it did mimicking the StackOverflow archives when I trained GPT-2 on them last year: https://stackroboflow.com (Only the 345M weights were released back then; now I'm curious how much better 1.5
335.
by yeldarb 6y ago
A friend of mine created this one and hosts a half-day seminar every few months to go through how to use it (and what the numbers mean/why you may want to do different things). I found the seminar quite useful and plan to use his model
336.
by yeldarb 6y ago
GDP is per year. Analogous to a person making $100k/yr with a $110k mortgage. Edit: Although perhaps not that analogous because is this even technically "debt"?
337.
by yeldarb 7y ago
Yes, twitter is flooded with reports of it at the moment. Google Cloud has confirmed they are having a "service disruption" but have not updated their status page yet: https://twitter.com/googlecloud/status&#x
338.
Apple to Close All Retail Stores Outside of China Until March 27(macrumors.com)
2 points by yeldarb 7y ago|0 comments
339.
Pioneer Livestream (Demo Day Live at 11am Pacific)(youtube.com)
4 points by yeldarb 7y ago|0 comments
340.
by yeldarb 7y ago
I and several other founders are doing this on the Pioneer livestream next week. It's earlier stage than most YC companies but not afraid of a little transparency https://pioneer.app/livestream In fact, there are whole
341.
How to Win Pioneer(blog.roboflow.ai)
2 points by yeldarb 7y ago|0 comments
342.
by yeldarb 7y ago
Edit: I had more I wanted to add that wouldn't fit in this comment so I wrote it up as a blog post here: https://blog.roboflow.ai/how-to-win-pioneer/ I don't think that's the case. We're working on
343.
by yeldarb 7y ago
For us it’s actually the weekly updates, voting system, and leaderboard itself. It gamifies productivity. The peer group is awesome too. I’ll add the caveat that we “won” relatively recently so haven’t participated in many of the formalized
344.
by yeldarb 7y ago
Pioneer has been a great productivity hack for us. As an early stage founder there’s no benchmark for how you’re doing relative to your peers. Matching up head-to-head weekly is a great way to assess your progress. Knowing I’m going to be j
345.
by yeldarb 7y ago
It's interesting that it was a public dataset (ImageNet) that kickstarted the amazing strides in computer vision over the last decade! More public data would be awesome (if high quality).
346.
by yeldarb 7y ago
The dataset is 15,000 images. Breakdown of the number of labels per class (post fixes) is here: https://i.imgur.com/bOFkueI.png Not all of the pedestrians and cyclists were on the sidewalk, no (eg the kid on his bike in the
347.
by yeldarb 7y ago
They claim[1] they're working on building an "open source self driving car" but it looks like the project hasn't had much activity recently. [1] https://github.com/udacity/self-driving-car
348.
by yeldarb 7y ago
I've never built a self driving car but if I were to give it a go, identifying what's in a frame would be the first step I'd tackle. From there you'd add more layers to the stack to get a full understanding of the world,
349.
by yeldarb 7y ago
It is, and there are some interesting techniques published recently to help mitigate things like this. But if you don't have a good ground truth you're at the very least flying blind and at worst feeding garbage in and getting gar
350.
by yeldarb 7y ago
I hope so. I've personally tested Scale's labeling service and it was much higher quality than this dataset. But it's a pretty secretive industry so I'd bet some companies' data is better than others. It'd be
351.
by yeldarb 7y ago
It doesn't, but since the original was supposedly human generated and I've now checked them all (twice) it should be n=2. Would love for someone to triple check! If you find errors let me know and I'll be happy to correct the
352.
by yeldarb 7y ago
They claim it was ("The dataset was annotated entirely by humans using Autti" via https://github.com/udacity/self-driving-car/tree/master/anno... ). But after looking at the data I'm almost
353.
by yeldarb 7y ago
I plan to. They use a custom CSV format that my labeling tool can't work with so I converted everything to VOC XML. I need to write a script to convert back to their format to submit a PR. Not sure if they'll accept the PR though;
354.
by yeldarb 7y ago
Yes, if you don’t “punish” incorrect predictions in your loss function your neural net could just get “perfect” accuracy by putting a giant bounding box around the entire image. Technically that’s “right”; it did put a box around all the ob
355.
by yeldarb 7y ago
Hey, OP here, yeah you’re correct. The dataset doesn’t label any obstacles that small/far in the distance. I zoomed in on the region with errors for the sake of the screenshot. Here’s the original run through Google Vision AI. They ac
356.
by yeldarb 7y ago
This is really scary. I discovered this because we're working on converting and re-hosting popular datasets in many popular formats for easy use across models... I first noticed that there were a bunch of completely unlabeled images. U
357.
A popular self-driving car dataset is missing labels for hundreds of pedestrians(blog.roboflow.ai)
412 points by yeldarb 7y ago|191 comments
358.
by yeldarb 7y ago
I would love if it could also let you search by the App or website that was active when the screenshot was taken. Probably a bit harder but maybe scraping app store screenshots and the top 5000 Alexa websites and training a classifier on th
359.
by yeldarb 7y ago
Associated Colab Notebook: http://bit.ly/yolov3-colab Chess Dataset Download: https://public.roboflow.ai/object-detection/chess-full
360.
Training a YOLOv3 Object Detection Model with a Custom Dataset(blog.roboflow.ai)
4 points by yeldarb 7y ago|1 comments

More ›