Does MxNet as a DL framework still have a place given Pytorch/tensorflow pretty much dominated all use cases?Amazon/AWS still “officially” supported it but given its product driven culture it could replace it with whatever framework that move faster and is more demanded by customers. Vendor Lock-in in this case probably won’t work as well since Amazon is not quite a leader in this case
Dive into Deep Learning
71–80 of 94 posts
Re: Dive into Deep Learning
#72As an engineer I find myself in this type of situation quite often - if anyone can point me to some good resources or has any advice, I'd be quite grateful: - Some non-technical stakeholder comes to me and says "can we solve this problem with Machine Learning?" usually it's something like "there need to be two supervisors on the factory floor at all times, and I want an email alert everytime there are less than 2 sup…
I would do the following:
- Manually scan through a couple of hours of data and setup a human baseline.
- Run standard algorithms and find their accuracy.
- Find errors in the model and analyze why the errors are happening. Is the model classifying some other object as a supervisor? Is the model not classifying the supervisor in certain lighting conditions or scenarios.
- Retrain the model with the failure scenarios so that it learns.
Re: Dive into Deep Learning
#73Earlier quoted context omitted.
This solution would change the work place culture - and I 100% bet would lead to a lot of good (for MY definition of good!) supervisors leaving. Imagine where you worked suddenly introduced this: "Yes, previously everyone could wear whatever they wanted - but from today, just the senior programmers must code while wearing a high-vis jacket around the office so we can track when they at their desks". The supervisors h…
I really don't understand your comment. 1. Supervisors are already by definition "superior" than their subordinates. 2. Supervisors on factories already wear distinctive clothing - especially in fully automated factories. Finally, you have yet to propose a solution to the problem yourself that would be highly accurate and easy to train. You vastly underestimate the difficulty to create a bespoke solution from scratch…
I don't have a deep-learning solution to the problem (I know nothing about it, that's why I clicked on it!). Seems really hard to me. I'd certainly go with an obvious "clock in and out" or "rfid" approaches... but I've worked in factories - and if you make someone wear some special clothes (or do some special tasks) when they didn't have to before - you're asking for trouble. People really hate change. That's presumably why the OP was looking for an answer for a tough problem.
A nerd analogy would be making a programmer change OSs (or even text editors) against their will: They could do it, but they won't be very happy about it.
Re: Dive into Deep Learning
#74Earlier quoted context omitted.
There are a load of questions here. > Where do I go from here? Keep trying different models? > ...after [the labeling is] all done is this even going to work? > [How to label] > If we spend all this time and money labelling footage, how well is this going to work? How much footage do we need to label? Generally, you're discussing the space of model improvement and refinement. This is the costliest and most dangerous…
Thanks - this validates many of the assumptions I had about this part of the process. It has been challenging communicating many of these realities to non-technical folks, who seem to be quite misguided about implementing these types of systems as opposed to "non-ML" systems where there is a less imperfect and more predictable idea of what's possible, how well it will work, and how much effort is required to pull it…
Yeah IME expectations with ML are just the worst. Somehow, non-ML-educated stakeholders expect it to be predictable, like they pretend traditional software engineering is... but also to be magical in scope.
Re: Dive into Deep Learning
#75Does MxNet as a DL framework still have a place given Pytorch/tensorflow pretty much dominated all use cases?Amazon/AWS still “officially” supported it but given its product driven culture it could replace it with whatever framework that move faster and is more demanded by customers. Vendor Lock-in in this case probably won’t work as well since Amazon is not quite a leader in this case
MXNet existed before AWS picked it, and it has a lot of strengths. I’d use it (especially with Gluon) over TF any day. But that said, PyTorch is usually easy to use on AWS... the preference for MXNet seems weak
Re: Dive into Deep Learning
#76Re: Dive into Deep Learning
#77Does MxNet as a DL framework still have a place given Pytorch/tensorflow pretty much dominated all use cases?Amazon/AWS still “officially” supported it but given its product driven culture it could replace it with whatever framework that move faster and is more demanded by customers. Vendor Lock-in in this case probably won’t work as well since Amazon is not quite a leader in this case
MXNet existed before AWS picked it, and it has a lot of strengths. I’d use it (especially with Gluon) over TF any day. But that said, PyTorch is usually easy to use on AWS... the preference for MXNet seems weak
It existed at CMU, but it seems like even CMU has moved over to PyTorch. I think Amazon just doesn't want to seem like an "also ran" by conceding to one of its competitor's frameworks.
Re: Dive into Deep Learning
#78As an engineer I find myself in this type of situation quite often - if anyone can point me to some good resources or has any advice, I'd be quite grateful: - Some non-technical stakeholder comes to me and says "can we solve this problem with Machine Learning?" usually it's something like "there need to be two supervisors on the factory floor at all times, and I want an email alert everytime there are less than 2 sup…
I'm somewhat surprised at the responses for this. I believe your issue can be easily solved - have supervisors wear a distinctive color from a non-supervisor. For example let's say it's yellow. OK so now you have yellow wearing supervisors and everyone else. To resolve the issue you have described acquire a month or so of footage, with labels per minute describing how many yellow wearing supervisors and how many peop…
This is a pragmatic and valid approach. No matter what anyone else says.
Imagine, you told a 10 YO child to do this task. Even the child would ask the same question - how do I know who is a supervisor and who is not.
Not only is face recognition hard, it is almost impossible to accomplish in a factory floor like setting. Not totally impossible but it is really really hard. Face detection is still possible but face recognition is far more computationally expensive. You'll need a shit ton of data and you'll need access to the employee database. You'll need a whole new engineering pipeline to make this happen and of course a team.
Compared to that expense and time, you are way better off getting the company to approve special vests for supes.
Re: Dive into Deep Learning
#79As an engineer I find myself in this type of situation quite often - if anyone can point me to some good resources or has any advice, I'd be quite grateful: - Some non-technical stakeholder comes to me and says "can we solve this problem with Machine Learning?" usually it's something like "there need to be two supervisors on the factory floor at all times, and I want an email alert everytime there are less than 2 sup…
At this stage the sibling comment to use GCP is pretty solid recommendation.
You can use Google for labelling (Mechancial Turk style), and AutoML Vision to train your model. It's going to be a bit pricey, but cheaper than your time to do the equivalent and will give you an educated guess at how much work it'll be to beat it. It costs about $100 to train a cloud vision model, I think (not including labelling)? You can also try the API for free to see how well Google does at finding people, they have better off the shelf models than you can get publicly.
https://cloud.google.com/vision/automl/docs/
You can try exploiting other things. Is your scene static? Try using frame differences as a feature. If it's a fixed environment then you should get a boost when fine tuning a model, versus some general person detector. COCO pretrained models should be quite good at finding people out of the box.
I wrote my own labelling tool specifically for Yolo which you may find useful (ie you label your data and export to a train-ready format): https://github.com/jveitchmichaelis/deeplabel
People who are not experienced are usually terrible at tagging images. They're not consistent, they miss objects and they don't understand why it's an issue. It will be faster to pay an "expert" service like mechanical turk, or do it yourself.
Basically a lot of your questions are open research problems. How much data do you need? Not a clue. It depends how your model is failing, which is always worth checking anyway. Figure out what the model is bad at and try and improve it, it should be doable to figure out where that 25% is going.
You should do better with a model like Faster-RCNN or its ilk. AutoML will do something like this, and you can try Facebook’s Detectron2 toolkit, or the Tensorflow Object Detection API.
Detecting unique people is a hard problem, by the way (eg two people versus the same person detected twice). You're better off just using an established method like RFID tags for presence/absence.
Another sibling made a great point. Don't detect people, train a model to output the number of people in the frame. This is how ML is applied to camera trap data with animals. In your case you can reduce this to a binary classification problem - >= 2 people, positive output.
Re: Dive into Deep Learning
#80As an engineer I find myself in this type of situation quite often - if anyone can point me to some good resources or has any advice, I'd be quite grateful: - Some non-technical stakeholder comes to me and says "can we solve this problem with Machine Learning?" usually it's something like "there need to be two supervisors on the factory floor at all times, and I want an email alert everytime there are less than 2 sup…
Honestly, I'd recommend trying Google's AutoML. I'm not a shill or employee at all, I've just had really great luck with it identifying my cats (each by name) with only a small amount (couple hundred, low thousands) labeled frames.
In my case it probably used transfer learning on like a resnet-150 or inception or something. Regardless, it approaches the limits of what an expert in machine learning can accomplish, so you'll know very quickly whether you need higher quality video / yellow vests.