How to ask for help

Before asking for help, please search the forums. Click the magnifying glass in the top right. Try a few different searches if you don’t find anything right away.

When you ask for help, you have a lot of context about your problem that the person you’re asking doesn’t know (and the answer depends on that context). By following these guidelines, others will be able to more quickly and effectively help you. Be sure to include:

  • Tell us what steps you’ve tried to fix the problem yourself, what you expected in each case, and what actually happened
  • Tell us you hypotheses about what might be going wrong, and your ideas about what approaches might be able to fix it
  • What computer you’re running the analysis on: eg. AWS (p2 or t2?), Paperspace, your own computer (which OS?), …
    ** Be sure to mention if you’re not using our standard AMIs/scripts, or if you’ve made some changes
    ** Note that what computer you’re actually typing at is generally not what we need to know, unless your problem is with awscli or some similar PC problem; since the actual analysis is generally running on a remote machine (e.g. AWS or Paperspace) it’s the details of that which we need
  • What you’ve typed, particularly if you’ve done anything differently to the setup and notebooks we’ve given you
  • Exactly what error message you received
  • If relevant, screenshots of problems that occurred
  • Whether the results of any earlier steps or configuration processes looked different in any way to what we did in class
  • What you’ve already tried to fix your problem (for instance, if you tried restarting the kernel on your notebook, or you read a related wiki page that didn’t cover your use case)
  • If relevant, if there’s something similar that did work for you (for example, you were able to get a command to work on t2, but not on p2).
15 Likes

Here is another article on “how to ask a data scientist for help”.

1 Like

I’ve read all the FAQs and I still can find where in interface I can create a forum post and ask question?

2 Likes

For real this is probably the best deep learning course content on the net, but also has the worst website design and functionality. Spent 10 mins looking and can’t even figure out how to make a forum post.

3 Likes

I agree and I genuinely need help with the course

Has anyone done both the fast.ai course and the Microsoft Certification for AI ? Just wondering the pros and cons between them.

Windows 7 Professional with fresh install of the fast.ai tools for CPU.

Errors during pytest.

Online searches including recommendations I’ve read on fast.ai aren’t helping. Am I required to have a GPU and/or Windows 10? Is there some form of workaround?

Disabled non-student, interested party, reading the how-to’s of Deep Learning for years, and just discovered fast.ai the other night using up my free hour on Crestle loving what I saw. A video catering to that first hour might be in order to help recruit more interested parties. Really sell it.

My first cats and dogs project was going to be differentiating Nissl bodies from Futurama brain slugs since I’ve been using Twitter to promote deep learning biology to help the BRAIN Initiative. I also admin a few groups on Facebook.

I’m praying I’m not at an impasse when it comes to my fast.ai predicament.

Jupyter errors with imports as well.

Regarding a workaround, when I coded a gaming calculator in JavaScript, I made a demo that used fixed data to demonstrate its functionality, something that could be readily demonstrated without actually playing the adventure.

Otherwise, as I’m in need to upgrade (while using my iPhone to create music with the Fusion Guitar it plugs directly into for effects), what should a deep learning enthusiast shoot for in the way of a new system?

Have you created a symlink to fastai folder where all the .py files are located? See step 8 of the following: Howto: installation on Windows

Dang! That didn’t fix it. I still get the same errors with torch. Must be lack of NVIDIA on my system. I was hoping to get something to work on this dinosaur. If you do look to port backward, much of the Department of Defense still runs on Windows XP.

Note to editors: On my secondary admin account I have to run Anaconda Prompt as Administrator for mklink to work.

Good news, though! I can budget Alienware. Just chatted with a Dell representative. So until then, adieu, unless, of course, there’s a place I can report the typos I keep finding in the Notebooks. I’m an essayist by demand, and spend too much time editing.

You just don’t know how long I’ve been waiting for an SDK like this. My interest in cybernetic technology dates back to 1975 at age 5, and in 1996 when I left school, neural networks were just a glimmer in the eye of professors. I’m so giddy: the computer can’t arrive soon enough.

I want to know how can I post questions in this forum?

2 Likes

Can you tell me how to post a problem on forum?

If you found out how please share the knowledge :wink:

Dear Jeremy
Thank you for all useful fastai lessons and its Github repo
I have a big problem that took three day of my time. I hope you or any of guys here can help me.
I am using lesson 3-camvid.ipynb.
I have my own dataset perfectly similar to camvid dataset. (size/dimensionality/data type everything is the same). But when I run it on google colab when I am running this line of code:
lr_find(learn)
learn.recorder.plot()
It returns an error:
RuntimeError: CUDA error: device-side assert triggered

I restart and reset the Runtime several times but I still see this error.

I really appreciate your help.

Hi Everyone,
I am working on the Leaf Classification Kaggle Problem!!
I am using the ImageDataBunch_from_df library to make the data but I am unable to process the test set using the same.
The data set has 3 folders.
Image Folder - It contains all the images with its name as {id}.jpg
Train.csv - It has an id column, the target column and many more features which I dont want to use for now.
Test.csv - Similiar as the train csv (Excluding the target column)

Since I am not using other features from the train csv, I made a dataframe with id and target
column and used ImageDataBunch_from_df.

But it is unable to read the test set for kaggle submission.

Can someone help me with this?

It comes in the home page only after you become “BASIC”

How does one become “basic”? Thanks.

2 Likes

Hello,

I ve got a question :slight_smile:
I made a fastai nlp model but i want to execute it on a new dataframe
So i tried the following
data_lm = (TextList.from_df(df1_test, Path(“model”), cols=‘Commentaire’)
.split_by_rand_pct(0.1, seed=42)
.label_for_lm()
.databunch(bs=128, num_workers=1, backwards=True))

then loading the model
classifier = load_learner(’/content/drive/My Drive/’, ‘verbatmodel2.pkl’)
Message /usr/local/lib/python3.6/dist-packages/torch/serialization.py:657: SourceChangeWarning: source code of class ‘fastai.text.models.awd_lstm.AWD_LSTM’ has changed. you can retrieve the original source code by accessing the object’s source attribute or set torch.nn.Module.dump_patches = True and use the patch tool to revert the changes.
warnings.warn(msg, SourceChangeWarning)

then
classifier.data.add_test(df1_test)

then
prob_preds = classifier.get_preds(ds_type=DatasetType.Test, ordered=True)

gives an error : AttributeError: ‘WeightDropout’ object has no attribute ‘idxs’