Issues running gpt-2 with provided Docker image

After following the instructions to build the gpt-2 Docker image here, I started up a bash shell into the container:

docker runĀ  -it gpt-2 bash

And then ran:

python3 src/generate_unconditional_samples.py | tee /tmp/samples

This failed with this error:

AttributeError: module 'tensorflow' has no attribute 'sort'

A post here says to upgrade to Tensorflow 1.14.0 in the container, so running:

pip install tensorflow==1.14.0

And then retrying, text generated! Now to start playing and see what this provided models will generate!

Is the quality of new questions on StackOverflow declining and/or is StackOverflow becoming less useful?

Now and then I browse StackOverflow hoping to be able to pick up a few easy questions and help some new developers. My activity on StackOverflow is sporadic, I browse a few times over a few days then a few months go by before I take another look. Why does my activity have gaps of sometimes months between activity? I haven’t thought about it before but I think I get disheartened by the experience. You invest time coming up with an appropriate answer and then even on a question where your answer is the only answer, the original poster doesn’t even bother to vote up your answer let alone select it as the best answer. More often that not, a new user with a rep of 1 presumably gets the answer they were looking for then disappears.

My motivation is more to provide help rather than earn more rep, but still, the rep system is the only tangible reward you get for participation, sometimes it just doesn’t feel worthwhile when you get nothing in return.

Here’s the other thing I’ve noticed recently: the majority of new questions asked rarely meet the requirements for acceptable questions you can ask, or are asked in a way that doesn’t meet the criteria of a good question. As a result, most new questions are downvoted and closed. That’s sad. Taking a quick look at the 10 most recent new questions tagged ‘Java’ right now and their current votes:

  • -1, no answers
  • -1 : closed, no answers
  • -4 : closed, no answers
  • -4, no answers
  • 0 with 1 answer
  • -1 : no answers
  • -3 : closed, no answers
  • 0 : 1 answer
  • 0 : no answers
  • 0 : no answers

Out of these 10:

  • 3 were already closed as not meeting the guidelines,
  • 3 have downvotes and will likely be closed unless they can be edited to meet guidelines,
  • 2 have no votes and 1 answer,
  • 2 no votes no answers.

This is pretty typical most days that I take a quick browse. Since I have Review Queue privs on new posts, out of maybe 3 out of 5 new questions I review I add comments to refer to the ‘what can I ask‘ and ‘how to ask‘ faqs, because most new questions are most obviously not following the guidelines.

Which brings me to my other observation which is pretty surprising when you think about it:

Of over 15 million registered users, users that actually respond to questions either by posting answers, comments, voting on questions or basically any interaction with the site that keeps it running in any useful way is a tiny percent of the total number of users. Look at the current leaderboard stats for all time:

You get a rep of 1 just for creating an account. This doesn’t even include users that interact with the site, searching for answer to questions etc when not even logged on or those not registered. Registered users at the lowest level of rep on the site between 1 and 199 are about 88% of the total registered users. Users with rep of 200 and above are about 12%.

This means the users that are actually providing answers to questions, editing questions/answers and asking clarifying questions are only 12% of the community’s users. That’s surprisingly low when you think the whole purpose of the StackOverflow site is to ask a question and (hopefully) get a useful answer.

Don’t get me wrong. I love StackOverflow. It’s a frequently used tool in my daily workflow as a developer and I’ve used it for years, as have many if not all developers everywhere. It’s just curious when you look at the numbers that the success of the site relies on the volunteer community of such a few users prepared to give back, where the largest percentage of users are those coming to the site with questions.

Microsoft Flight Simulator: UK cities with photogrammetry

At launch in September 2020, there were only two UK cities in Microsoft Flight Simulator that had photogrammetry (source) :

  • Portsmouth
  • Southampton

After the UK update in February 2021, an additional 5 cities were added (source) :

  • Birmingham
  • Bristol
  • Cambridge
  • London
  • Oxford

I’ve been checking out each of these looking for odd scenery glitches in the photogrammetry imagery, and have found some interesting examples.

In most locations even at a low altitude, the scenery is almost indistinguishable from real life:

However, where it doesn’t work, it really doesn’t work with weird and comical results. Here’s some examples from Southampton:

Sunken boats along the River Itchen:

This is what happens if you don’t move your car when the city resurfaces the road:

Weird plants turned to stone:

B&Q having a meltdown:

I’ve posted a few other examples before, and I’m sure there’s plenty more weirdness to be found:

Butter verses Potato: Elite Dangerous vs Microsoft Flight Simulator using a Quest 2 VR

I bought a Quest 2 specifically to play with Microsoft Flight Simulator using a Link cable. I had a GearVR previously, but this has been my first experience with PC based VR. It’s been a mixed experience. Does it work? Sort of. But you have to put in a ton of time either following settings recommendations or spend time yourself tweaking settings to find what balance works for you. It’s currently a compromise between graphics quality and framerates – unless you drop a few thousand dollars on the fastest hardware and current top of the range Nvidia 3090 (which alone will cost you around $3000) you have to make a compromise.

I’ve followed a lot of online guides for settings recommendations and came up with a combination that worked ok for me. However, I’ve come to accept that this is really two different games with completely different experiences:

  • playing with Ultra everything on desktop and enjoying the incredible graphics (which honestly is incredible)
  • playing in VR with seriously dialed back settings to get it to run smooth, but being amazed at the emersiveness and believable feeling of being in a plane cockpit and actually flying

These are not the same two things for me, at least with the hardware I have. It’s two different games with two different experiences.

Knowing that I’d probably get a much better VR experience with native Quest 2 games, I’ve tried out a few had a great experience (Star Wars Pinball, Job Simulator). so I know the Quest can give a good experience. Standalone it works great.

I’ve played a lot of hours previously with Elite Dangerous, and I remember it was one of the earlier games to get PC VR support. I’d been so focused on getting the settings just right for MS Flight Simulator that I hadn’t even tried any other PC games. So I gave it a go.

I was surprised. So very, very surprised. Compared to MS Flight Simulator it’s a night and day experience. I didn’t bother bringing up the Oculus tray and tweaking any settings, I was just initially curious to see what it would be like. I started up the game, turned on VR headset support in game (didn’t even start it from Steam, and since I’d already started Link it just went straight in. The experience was smooth, it was fluid, no lags, no stutters, it was really an incredible experience. No tweaking settings, no lowering graphics features, it just worked.

The incredible thing was the sense of massiveness of the environment that I’d never experienced playing the game sitting in front of a monitor. The insides of the ship are massive, the ships themselves are massive when viewed from your SRV outside. Flying across the vastness of space feels like you’re there. Approaching a planet you get the sense of how massive the planets are as you approach.

In summary, MS Flight Simulator has not been optimized enough to be comfortably playable yet. Elite Dangerous on the other hand is incredible in VR. The Quest 2 with Link cable can work flawlessly out of the box. I hope MS Flight Simulator will get there but it’s hit or miss whether you can get the settings tuned enough to be playable on your hardware.

Admittedly I have’t tried any other PC games at this point, but if you want a showcase experience of what’s possible, try Elite Dangerous with the Quest 2 over Link. It’s incredible.