Hi, if you have an accounting degree, you can check the Tabular notebook, Rossmann notebook, Collaborative.
I think you know what tabular data. In that tabular notebook, you use the input data to predict the possible income, and it has a state of art result. In Rossman data, you will clean the data. And it can map a product region on a map based on the data set.
Hey zachery, maybe you can build a tabular and train it to make a prediction or draw a regional category map to group the interest from the users.
Hi, I am the new user for fastai. I have trained model on GPU and now planning to deploy on CPU/VM. However on CPU, utilization go up to 350% every predict request i make. And if i open multiple process using gunicorn its just unmanageable. Kindly help.
pred_class, pred_idx, outputs = fastai_learner.predict(img_fastai). ā 3.5 Sec and 350% CPU on every request.
CPU-
learner = load_learner(model_fastai_path,model_fastai)
pred_class, pred_idx, outputs = fastai_learner.predict(img_fastai) - 3.5 Sec and 350% CPU on every request.
Hey guys, Can a complete beginner in deep learning with no experience in other frameworks can get started with v4 or should he have some experience with PyTorch? (I am actually recommending this to a friend of mine who is complete beginner)
For those who are wondering - itās a standard YouTube video. So look at the options menu at the bottom right of the video and you should see an icon with CC (stands for Closed Captions) as text.
Hi I will be finishing the Practical Deep Learning for Coders, v3(this is the latest one yes?)
And I intend to go on to Part 2: Deep Learning from the Foundations at the link (https://course19.fast.ai/part2) however on the side bar of said url , there is lessons part 1 and 2 does part 1 refer to Practical Deep Learning for Coders, v3 or ā¦? Because some of the lessons have the same names as those in Practical Deep Learning for Coders, v3 but some are not the sameā¦ just wondering as I would like to come up with a study plan
Cheers and happy learning guys
Hi All and sorry if this a repeat as just bought the Deep Learning for Coders with fast.at and PyTorch and was getting started with the first project using Paperspace Gradient but all the free machines are out of capacity. Given that the bookās website says that everything in the course works with Gradient and there are some issues when using CoLab, my question is, I have a free CoLab account but is it worth it to use one of the pay machines on Paperspace Gradient rather than hassle with issues on CoLab?
You can use Colab pro with 25GB of RAM(if I remember correctly) with Tesla V100 with just 9.99 a month.
If you are just using the fastaiās notebook, it shouldnāt have any problem, and I remember the notebook also teaches you to use something like gc.collection to collect the unused allocated memory.
Same, not the best user experience to have to try and figure out why thereās no ability to post as a new user. I donāt mind the adventure, but it creates some friction.