I have a question related to this one:
TensorFlow in production for real time predictions in high traffic app - how to use?
I want to setup TensorFlow Serving to do inference as a service for our other application. I see how TensorFlow Serving helps me to do that. Additionally, it mentions a continuous training pipeline, which probably is related to the possibility that TensorFlow Serving can serve with multiple versions of a trained model. But what I am not sure is how to retrain your model as you get new data. The other post mentions the idea to run retraining with cron jobs. However, I am not sure if automatic retraining is a good idea. What architecture would you propose for a continuous retraining pipeline with a system continuously facing new, labelled data?
Edit: It is a supervised learning case. The question is would you automatically retrain your model after n new datapoints came in or would you retrain during the downtime of the customer automatically or just retrain manually?
You probably want to use some kind of semi-supervised training. There's fairly extensive research in that area.
A crude, but expedient way, which works well, is to use the current best models that you have to label the new, incoming data. Models are typically able to produce a score (hopefully a logprob). You can use that score to only train on the data that fits well.
That is an approach that we have used in speech recognition and is an excellent baseline.
Related
I'm looking into training an object detection network using Tensorflow, and I had a look at the TF2 Model Zoo. I noticed that there are noticeably less models there than in the directory /models/research/models/, including the MobileDet with SSDLite developed for the jetson xavier.
To clarify, the readme says that there is a MobileDet GPU with SSDLite, and that the model and checkpoints trained on COCO are provided, yet I couldn't find them anywhere in the repo.
How is one supposed to use those models?
I already have a custom-trained MobileDetv3 for image classification, and I was hoping to see a way to turn the network into an object detection network, in accordance with the MobileDetv3 paper. If this is not straightforward, training one network from scratch could be ok too, I just need to know where to even start from.
If you plan to use the object detection API, you can't use your existing model. You have to choose from a list of models here for v2 and here for v1
The documentation is very well maintained and the steps to train or validate or run inference (test) on custom data is very well explained here by the TensorFlow team. The link is meant for TensorFlow version v2. However, if you wish to use v1, the process is fairly similar and there are numerous blogs/videos explaining how to go about it
I'm trying to train a model for a sentence classification task. The input is a sentence (a vector of integers) and the output is a label (0 or 1). I've seen some articles here and there about using Bert and GPT2 for text classification tasks. However, I'm not sure which one should I pick to start with. Which of these recent models in NLP such as original Transformer model, Bert, GPT2, XLNet would you use to start with? And why? I'd rather to implement in Tensorflow, but I'm flexible to go for PyTorch too.
Thanks!
It highly depends on your dataset and is part of the data scientist's job to find which model is more suitable for a particular task in terms of selected performance metric, training cost, model complexity etc.
When you work on the problem you will probably test all of the above models and compare them. Which one of them to choose first? Andrew Ng in "Machine Learning Yearning" suggest starting with simple model so you can quickly iterate and test your idea, data preprocessing pipeline etc.
Don’t start off trying to design and build the perfect system.
Instead, build and train a basic system quickly—perhaps in just a few
days
According to this suggestion, you can start with a simpler model such as ULMFiT as a baseline, verify your ideas and then move on to more complex models and see how they can improve your results.
Note that modern NLP models contain a large number of parameters and it is difficult to train them from scratch without a large dataset. That's why you may want to use transfer learning: you can download pre-trained model and use it as a basis and fine-tune it to your task-specific dataset to achieve better performance and reduce training time.
I agree with Max's answer, but if the constraint is to use a state of the art large pretrained model, there is a really easy way to do this. The library by HuggingFace called pytorch-transformers. Whether you chose BERT, XLNet, or whatever, they're easy to swap out. Here is a detailed tutorial on using that library for text classification.
EDIT: I just came across this repo, pytorch-transformers-classification (Apache 2.0 license), which is a tool for doing exactly what you want.
Well like others mentioned, it depends on the dataset and multiple models should be tried and best one must be chosen.
However, sharing my experience, XLNet beats all other models so far by a good margin. Hence if learning is not the objective, i would simple start with XLNET and then try a few more down the line and conclude. It just saves time in exploring.
Below repo is excellent to do all this quickly. Kudos to them.
https://github.com/microsoft/nlp-recipes
It uses hugging face transformers and makes them dead simple. 😃
I have used XLNet, BERT, and GPT2 for summarization tasks (English only). Based on my experience, GPT2 works the best among all 3 on short paragraph-size notes, while BERT performs better for longer texts (up to 2-3 pages). You can use XLNet as a benchmark.
I am re-training the SSD MobileNet with 900 images from the Berkeley Deep Drive dataset, and eval towards 100 images from that dataset.
The problem is that after about 24 hours of training, the totalloss seems unable to go below 2.0:
And the corresponding mAP score is quite unstable:
In fact, I have actually tried to train for about 48 hours, and the TotoalLoss just cannot go below 2.0, something ranging from 2.5~3.0. And during that time, mAP is even lower..
So here is my question, given my situation (I really don't need any "high-precision" model, as you can see, I pick 900 images for training and would like to simply do a PoC model training/predication and that's it), when should I stop the training and obtain a reasonably performed model?
indeed for detection you need to finetune the network, since you are using SSD, there are already some sources out there:
https://gluon-cv.mxnet.io/build/examples_detection/finetune_detection.html (This one specifically for an SSD Model, uses mxnet but you can use the same with TF)
You can watch a very nice finetuning intro here
This repo has a nice fine tuning option enabled as long as you write your dataloader, check it out here
In general your error can be attributed to many factors, the learning rate you are using, the characteristics of the images themselves (are they normalized?) If the ssd network you are using was trained with normalized data and you don't normalize to retrain then you'll get stuck while learning. Also what learning rate are they using?
From the model zoo I can see that for SSD there are models trained on COCO
And models trained on Open Images:
If for example you are using ssd_inception_v2_coco, there is a truncated_normal_initializer in the input layers, so take that into consideration, also make sure the input sizes are the same that the ones you provide to the model.
You can get very good detections even with little data if you also include many augmentations and take into account the rest of the things I mentioned, more details on your code would help to see where the problem lies.
Is there any equivalent/alternate library to fastai in tensorfow for easier training and debugging deep learning models including analysis on results of trained model in Tensorflow.
Fastai is built on top of pytorch looking for similar one in tensorflow.
The obvious choice would be to use tf.keras.
It is bundled with tensorflow and is becoming its official "high-level" API -- to the point where in TF 2 you would probably need to go out of your way not using it at all.
It is clearly the source of inspiration for fastai to easy the use of pytorch as Keras does for tensorflow, as mentionned by the authors time and again:
Unfortunately, Pytorch was a long way from being a good option for part one of the course, which is designed to be accessible to people with no machine learning background. It did not have anything like the clear simple API of Keras for training models. Every project required dozens of lines of code just to implement the basics of training a neural network. Unlike Keras, where the defaults are thoughtfully chosen to be as useful as possible, Pytorch required everything to be specified in detail. However, we also realised that Keras could be even better. We noticed that we kept on making the same mistakes in Keras, such as failing to shuffle our data when we needed to, or vice versa. Also, many recent best practices were not being incorporated into Keras, particularly in the rapidly developing field of natural language processing. We wondered if we could build something that could be even better than Keras for rapidly training world-class deep learning models.
The tensorflow text summarization model as described here https://github.com/tensorflow/models/tree/master/textsum requires a multi GPU architecture in order to train. My repeated attempts at training the model has resulted in memory exceptions, machine crashing for various reasons. Is the trained summarisation model available so can make use of the summarization model without the need for training? The summarization model is trained using the not free Gigaword dataset, if the trained model is not available from Google is this a factor in reason why ?
So as far as I can tell, no one has put the trained model out there that is referenced. I too was originally running into memory issues on my macbook pro and eventually ended up using my gaming laptop which had a much better GPU.
The other option of course is to take advantage of AWS and use something like their g2.2xlarge instance. They also have their P2 instances as well, but I have not checked that out yet.
With regards to the Gigaword dataset, it simply comes down to licensing. It is not a free license from LDC and often many of the academics working on this have the dataset provided to them via their Universities or companies. I have not had luck finding it, however LDC did get back to me and advised that they do have other article datasets that have a pricetag of around $300 which is much more reasonable for those of use just trying to learn TF. That said, if you didn't want to buy anything, you can always write your own page scraper and format the data for the textsum model. https://github.com/tensorflow/models/pull/379/files
Hope this helps some. Good luck!