We are a 50+ company with a working, prototype robotic vertical farm that is already shipping fully chemical-free vegetables farmed by robots to major European supermarkets.
Our international team of engineers, scientists and roboticists spans Lisbon, Milan and Oxford. As a CV/ML Engineer you will propose, implement and test new deep learning algorithms that will increase the autonomy of our vertical farm operations.
For more info, e-mail: joao.henriques at planetfarms dot ag
Planet Farms | Computer Vision and Machine Learning Engineer | Portugal or REMOTE | https://www.planetfarms.ag
Fully automated vertical farms.
We are a 50+ company with a working, prototype robotic vertical farm that is already shipping fully chemical-free vegetables farmed by robots to major European supermarkets.
Our international team of engineers, scientists and roboticists spans Lisbon, Milan and Oxford. As a CV/ML Engineer you will propose, implement and test new deep learning algorithms that will increase the autonomy of our vertical farm operations.
For more info, e-mail: joao.henriques at planetfarms dot ag
Thanks, but I think that lucb1e's confusion was probably the same as mine -- given pretrained CLIP features, how is this translated to zero-shot tracking?
Are initial bounding boxes given as usual, or are objects of interest created automagically?
It uses an object detection model (in our example code[1], we used one from Roboflow Universe[2] but you should be able to use any object detection model) to get the bounding boxes and then sends a crop of each detected box to CLIP to get the feature vector that Deep SORT uses to differentiate between and track instances across frames.
This is in comparison to the original Deep SORT[3] which requires you to train a second custom "deep appearance descriptor" model for the tracker to use.
There's also Siamese Networks, they're pretty fast and simple, and people keep updating their backbone CNNs to more modern (computationally expensive) ones.
> significant amount of engineering cost to continue to deliver static sites, even with reduced fidelity and a worse user experience
A few things to consider:
- The reputational damage of solidifying the narrative that Google drops projects without a good reason from the users' point of view.
- A bare-bones deletion/flagging UI does not seem like a huge undertaking (though as an engineer I realizing it's not exciting work); working things out with a non-profit such as the Internet Archive or a museum could be an even lower-cost solution.
- 90's websites are also hideous ("worse user experience") by today's standards, yet they have charm and are a product of their times. I don't think anyone would argue that they should be deleted on account of that, just like ancient pottery often has "badly drawn" human figures yet the value is in the cultural expression. There are digital conservation efforts by museums (e.g. restoring 80's arcade games); I'm sure that consulting with a digital conservator would have arrived at a very different treatment of this data.
> The reputational damage of solidifying the narrative that Google drops projects without a good reason from the users' point of view.
That ship has sailed. Google's reputation for instability can't get any worse. At this point, it would take a complete 180° and a good 5-10 years of angelic behavior from Google for me to even consider relying on them in any capacity.
gmail is still nice. I wonder when they are going to "work" on that.
i stll giggle at my ignorance thinking they wouldnt do youtube and search.
i really mis search, it was such a nice product. Now it is so bad that internet natives who could slap together a website in 20 min dont understand when i ask them why they have no website of their own. "you mean like a resume?"
they are right of course. no one would ever find it.
Implementing flagging and deletion is easy-ish, the maintenance is hard.
There's maintenance for unit, integration, and web driver tests. Someone needs to be oncall when they fail or become flaky. Someone needs to keep them up to date as tools and infra are deprecated. What's the SLA and SLO? Someone will have to maintain that. Who responds to monitoring?
Regulatory changes require effort, tests need to be updated, legal needs to sign off, UX may need to be involved. Who will make sure the sites maintain regulatory requirements?
You're right that it may be better to partner with someone else for archiving.
The effort required to maintain a single non-changing binary, while nonzero, is far, faaar less, than the ongoing effort required to maintain user facing tools with regulatory compliance issues.
Good job, this looks extremely expressive and with fewer corner cases ("dark knowledge") than MPL/SNS, and unlike Altair/Plotly doesn't require a whole browser to display the output!
Still, I'd like to ask if you considered alternatives to MPL for the back-end. It's a venerable but ancient project with years of accumulated technical debt, and I'm sure you had to deal with lots of inconsistencies there.
For example, PyQtGraph is an alternative with a clear class hierarchy and can handle large-scale datasets without slowing down (while anything non-trivial in MPL has you wait seconds to render).
(I'd love to hear more suggestions that don't require a JS engine and don't build on MPL.)
Thanks! I'm focused on building the user-facing API, as this is where I believe I'm best suited to make improvements due to my experience teaching and writing.
I'm definitely open to looking at alternative backends in the future and will check out PyQtGraph, but am sticking to matplotlib for now.
It depends what you call the "home row" and where you keep your fingers.
If you expect to use these shortcuts (typically with the outer fingers of your hands) your hands will rest at a slightly higher position in your keyboard, with the pinkies near Esc and Del, and the thumbs near the spacebar. (Other fingers over the letters.)
I guess it depends on what you've trained yourself to do.
I don't think I've seen anyone with their fingers stretched out or diagonal like that as a 'home row' position, though. It makes much more sense to 'default' to having your hands somewhat straight and on the keys you're likely to press more regularly (the (center) letter row).
key chords, F-keys, waiting for windows to pop up, those are all extra things that take time and get in the way. My fingers need to stay near their home on the home row.
Not affiliated in any way, but you should give Signal a try.
The hardest part with any IM service is going to be convincing your friends to join though. I wish someone came up with a good way to overcome this "critical mass" problem.
Also not affiliated, just an user -- but Telegram has a lot of momentum going for it. It's possibly more of a substitute for Facebook than Whatsapp at this point, but in chat form, with a rich bot ecosystem and great UX.
My wife and I tried Signal in the past and ran into situations where messages were lost or sometimes delayed by hours, some showing up well after other messages were, making conversations occasionally very frustrating.
Messaging apps are definitely one of those situations where I expect things to work all of the time, not just most of the time.
Feature wise telegram is definitely the best challenger right now. Signal is good but not as good. I wish it was better because of privacy. I like telegram too much now.
I'm really hoping that Chat over IMAP picks up. I use Delta Chat every now and then with some journalists who have it installed, and it just makes so much sense. We already have everyone's mail ID's anyway. Or at least I thought so. In countries like India, a large number of people have skipped email and gone to phone based services. So for Chat over IMAP to succeed over here it would need to have phone number based matching cos people do not use their mail ID's at all. Everything gets done over whatsapp.
I've told the important people in my life that if they want to message me, they will need do it through Signal. It takes 2 minutes to set up, I show them how, and tell them that I won't reply to any SMS.
Many switched, including the least technical people I know. A few did not. It's really not hard; you just have to draw a line in the sand.
As a result of everyone I know using Signal, it fundamentally doesn't matter to me if the "mainstream" switches. We're already all on board.
And then you end up in that situation when they didn't actually switch, but rather made you a favor and caved in to your self-centric demands and stubborness. Like with a tantrum-prone child. They still use WhatsApp between themselves, because that's what their dentist, car mechanic and a favourite sushi takeout use. They will also more eagerly reply to IMs if they are not sent through your preferred chat system. So in the end you technically made them do your thing, but the victory is hollow.
The problem is when one friend demands Signal, another WhatsApp, another Matrix, another Telegram, another Threema. These are all closed platforms and having loads of different apps installed which perform similar functions，to satisfy various people's dogmatic views is annoying.
Matrix is not a closed platform, and in fact the only one of the projects you mentioned that even makes an attempt to work together with other chat protocols and networks. It is explicitly open, not locked in to any single service provider, and making an explicit effort to not limit you to chat with only other Matrix users (despite how much Signal or WhatsApp works towards the opposite).
An average person will have pages of apps installed, but that would actually use only what's on the first page. So they have a choice of bumping something off their precious FP to fit my special app or just keep it where it lands after installation, on the back pages. And if it ends up there, out of rotation, it won't see much voluntary use.
If you really want to make a difference please try to make people to migrate to something open. Quicksy[^1], DeltaChat[^2], something. But not another, essentially closed platform.
I agree. However email metadata isn't secure, so that's a nonstarter. I'm rooting for Matrix but thusfar Riot's UX to get an encrypted conversation going leaves a lot to be desired. No way are the non-technical people I know going to be able to do that. Signal "just works" with secure defaults, so that's the best we've got for now.
> I've told the important people in my life that if they want to message me, they will need do it through Signal. It takes 2 minutes to set up, I show them how, and tell them that I won't reply to any SMS.
I've done the opposite: there's no way I'm going to install a messaging app, and everyone I care to interact with knows they need to either use SMS or email to reach me. Everyone's been perfectly fine with that.
WhatsApp itself started as an obscure chat application. They nailed the simple SMS-like experience without high SMS fees, and that was enough for mainstream users. No reason people won't jump ship when WhatsApp loses that core experience.
It's still going to remain the same. Whatsapp is too big to fail, perhaps people in the US just don't comprehend how big of a user base Whatsapp has around the world.
> WhatsApp itself started as an obscure chat application
It actually started as a one-line status app, a bit like finger for the mobile era. Hence the name.
It was only when they discovered that users were interactively chatting by changing their status lines in response to each other that WhatsApp pivoted to full chat.
A good example of adapting to improvised user behaviour.
Fully automated vertical farms.
We are a 50+ company with a working, prototype robotic vertical farm that is already shipping fully chemical-free vegetables farmed by robots to major European supermarkets.
Our international team of engineers, scientists and roboticists spans Lisbon, Milan and Oxford. As a CV/ML Engineer you will propose, implement and test new deep learning algorithms that will increase the autonomy of our vertical farm operations.
For more info, e-mail: joao.henriques at planetfarms dot ag