I recently trained a stable-baselines PPO model for a couple of days and it is performing well on test environments. Essentially, I am trying to iterate on this model. I was wondering if it is possible to use this model as a new baseline for future model training. So, instead of starting with some naive policy for my environment, it could use this model as a starting spot and potentially learn a better approach to solving the environment.
The answer is yes. You basically need to do following things to achieve this:
Save your PPO model after the training in any environment you used from stable-baseline repository.
Load the saved PPO model for your new environment training rather than create a new PPO model. In this case, the starting spot will be the trained policy and the policy will evolve from there for better.
One thing you might be interested in is Transfer Learning, which basically systematically does what you want to do (train a model in one environment and use that as a baseline for another new environment to save training time for the second environment). The key to make it work is make sure there is high similarity between the two environments. If they are totally different, the pre-trained policy from the first environment may not help much.
In addition, the network architecture of the PPO is the same when you do this. In reality, the optimal network architecture for different environments are different.
Related
I am trying to make a PPO model using the stable-baselines3 library. I want to use a policy network with an LSTM layer in it. However, I can't find such a possibility on the library's website although it exists on the previous version of stable-baselines here https://stable-baselines.readthedocs.io/en/master/modules/policies.html#stable_baselines.common.policies.MlpLstmPolicy.
Does this possibility exist in stable-baselines3 (not stable-baselines)? if not, is there any other possibility I can do this? Thanx.
From the migration doc.
https://stable-baselines3.readthedocs.io/en/master/guide/migration.html
Breaking Changes¶
LSTM policies (MlpLstmPolicy, CnnLstmPolicy) are not supported for
the time being (see PR #53 for a recurrent PPO implementation)
Currently this functionality does not exist on stable-baselines3.
However, on their contributions repo (stable-baselines3-contrib) they have an experimental version of PPO with LSTM policy. I have not tried it myself, but according to this pull request it works.
You can find it on the feat/ppo-lstm branch, which may get merged onto master soon.
Even though I am quite familiar with the concepts of Machine Learning & Deep Learning, I never needed to create my own dataset before.
Now, for my thesis, I have to create my own dataset with images of an object that there are no datasets available on the internet(just assume that this is ground-truth).
I have limited computational power so I want to use YOLO, SSD or efficientdet.
Do I need to go over every single image I have in my dataset by my human eyes and create bounding box center coordinates and dimensions to log them with their labels?
Thanks
Yes, you will need to do that.
At the same time, though the task is niche, you could benefit from the concept of transfer learning. That is, you can use a pre-trained backbone in order to help your model to learn faster/achieve better results/need fewer annotations example, but you will still need to annotate the new dataset on your own.
You can use software such as LabelBox, as a starting point, it is very good since it allows you to output the format in Pascal(VOC) format, YOLO and COCO format, so it is a matter of choice/what is more suitable for you.
I've used three parallel sentence files to train my custom translator model. No dictionary files and no tuning files too. After training is finished and I've checked test results, I want to make some adjustments in the model. And here are several questions:
Is it possible to tune the model after training? Am I right that the model can't be changed and the only way is to train a new model?
The best approach to adjusting the model is to use tune files. Is it correct?
There is no way to see an autogenerated tune file, so I have to provide my own tuning file for a more manageable tuning process. Is it so?
Could you please describe how the tuning file is generated, when I have 3 sentence files with different amount of sentences, which is: 55k, 24k and 58k lines. Are all tuning sentences is from the first file or from all three files proportionally to their size? Which logic is used?
I wish there were more authoritative answers on this, I'll share what I know as a fellow user.
What Microsoft Custom Translator calls "tuning data" is what is normally known as a validation set. It's just a way to avoid overfitting.
Is it possible to tune the model after training? Am I right that the model can't be changed and the only way is to train a new model?
Yes, with Microsoft Custom Translator you can only train a model based on the generic category you have selected for the project.
(With Google AutoML technically you can choose to train a new model based on one of your previous custom models. However, it's also not usable without some trial and error.)
The best approach to adjusting the model is to use tune files. Is it correct?
It's hard to make a definitive statement on this. The training set also has an effect. A good validation set on top of a bad training set won't get us good results.
There is no way to see an autogenerated tune file, so I have to provide my own tuning file for a more manageable tuning process. Is it so?
Yes, it seems to me that if you let it decide how to split the training set into the training set, tuning set and test set, you can only download the training set and the test set.
Maybe neither includes the tuning set, so theoretically you can diff them. But that doesn't solve the problem of the split being different between different models.
... Which logic is used?
Good question.
I need to find a deep learning based prediction model, where can I find it?
You can use Pytorch and Tensorflow pretrained models.
https://pytorch.org/docs/stable/torchvision/models.html
They can be automatically downloaded. There are some sample codes, that you can try:
https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html#sphx-glr-beginner-blitz-cifar10-tutorial-py
If you are interested in deep learning, I suggest you review the basics of it in cs231n stanford. Your question is a bit odd, because you first need to define your task specifically. Prediction is not a good description. You could look for models for classification, segmentation, object detection, sequence2sequence(like translation), and so on...
Then you need to know how to search through projects on github, and then you need to know python (in most cases), and then use a pretrained model or use your own dataset to train or fine-tune the model for that task. Then you could pray that you have found a good model for your task, after that you need to validate the results on a test set. However, implementation of a model for real-life scenarios is another thing that you need to consider many other things, and you usually need some online-learning strategy, like Federated Learning. I hope that I could help you.
Is there a simple way (e.g. without modifying caffe code) to load wights from multiple pretrained networks into one network? The network contains some layers with same dimensions and names as both pretrained networks.
I am trying to achieve this using NVidia DIGITS and Caffe.
EDIT: I thought it wouldn't be possible to do it directly from DIGITS, as confirmed by answers. Can anyone suggest a simple way to modify the DIGITS code to be able to select multiple pretrained networks? I checked the code a bit, and thought the training script would be a good place to start, but I don't have in-depth knowledge of Caffe, so I'm not sure what the best/quickest way to achieve this would be.
As Shai suggested, there was no way of doing this, so I decided to clone the official repository and make the appropriate changes. I changed the code so that multiple pretrained networks can be loaded by using a colon as separator.
I created a pull request on the official repository and my changes were then merged with the main branch of DIGITS, meaning it is now possible to use this functionality in DIGITS.
AFAIK there is no straight forward way of doing so.
However, you can use net surgery to load the pretrained models and manually assign their weights to the target net. Once you have a single net with all the weights initialized according to the various pretrained models, you can save it and use it as a single pretrained model for the rest of your work.