thanks for quick answer!
shame i thought WSL Ubuntu will make use of GPU
edit: link to github discussion on adding GPU support to WSL, but for now there is no timeframe set for thisā¦ https://github.com/Microsoft/WSL/issues/829
i did try using fastai with GPU on Windows in the past but it had issues, like was very slow compared to with same GPU on linux. Probably had something do to with number of workers but also was kind of hanging in between epochs. i should give it a try now and see if things changed!
You can check this topic that I created a couple of days ago. Check out Nicolasā 'blog post that he published. I didnāt have the chance to try yet (low priority for now).
If you are successful in doing that, please post your findings.
Thanks Miwojc, I was wondering about that; being confused enough, is enough! I shall heed your advice. Thankfully I havenāt purchased it yet, despite the countless drip-feed follow-ups. I remain unsure as to the depth of coverage despite the claims.
Also, I have discovered and started reading pjh and Silvanās EXCELLENT BOOK - this is by far the best preparatory activity in anticipation of the course. As well, reading some (or ALL) of Rachaelās work is magic.
Thanks Miwojc - some of my best friends are Polish!,
Cheers, p
Salut,
I hope this post is rightly placed hereā¦
I have starfted with fastai and executed the following command in the python shell:
from fastai.vision import *
from fastai.callbacks.one_cycle import *
path = untar_data(URLs.MNIST_SAMPLE)
data = ImageDataBunch.from_folder(path)
model = simple_cnn((3,16,16,2))
learn = Learner(data, model)
learn.metrics=[accuracy]
learn.fit(1)
It works fine but when I copy it into a file test.py and call python I get an error when learn.fit(1) starts.
File āC:\Program Files\Python37\lib\multiprocessing\spawn.pyā, line 136, in _check_not_importing_main
is not going to be frozen to produce an executable.āāā)
RuntimeError:
An attempt has been made to start a new process before the
current process has finished its bootstrapping phase.
This probably means that you are not using fork to start your
child processes and you have forgotten to use the proper idiom
in the main module:
if __name__ == '__main__':
freeze_support()
...
The "freeze_support()" line can be omitted if the program
is not going to be frozen to produce an executable.
I do not unxderstand why I can run code interactively but not as a python file
Ciao, mathias
rgs)
478 # We may need to call lazy init again if we are a forked child
479 # del _CudaBase.new
ā 480 return super(_CudaBase, cls).new(cls, *args, **kwargs)
481
482
RuntimeError: CUDA error: the launch timed out and was terminated
getting this error on Window10, suspect is the dataloader num_workers issue, but I have tried to set num_workers=0, no luck yet.
Hi guys, Iām trying to run the FastAI2 pets lesson 1 after installing it in a Win10 (editable install), every time I try to load the main libraries I get an error stating that: name ālog_argsā is not defined. Any chance anyone has some insight to provide?
from fastai2.basics import *
from fastai2.callback.all import *
from fastai2.vision.all import *
from nbdev.showdoc import *
---------------------------------------------------------------------------
NameError Traceback (most recent call last)
in
----> 1 from fastai2.basics import *
2 from fastai2.callback.all import *
3 from fastai2.vision.all import *
4 from nbdev.showdoc import *
d:\fai2\fastai2\fastai2\basics.py in
----> 1 from .data.all import *
2 from .optimizer import *
3 from .callback.core import *
4 from .learner import *
5 from .metrics import *
d:\fai2\fastai2\fastai2\data\all.py in
----> 1 from ā¦torch_basics import *
2 from .core import *
3 from .load import *
4 from .external import *
5 from .transforms import *
d:\fai2\fastai2\fastai2\torch_basics.py in
3 from .torch_imports import *
4 from .torch_core import *
----> 5 from .layers import *
d:\fai2\fastai2\fastai2\layers.py in
271
272 # Cell
ā> 273 @log_args
274 @funcs_kwargs
275 class BaseLoss():
NameError: name ālog_argsā is not defined
I tried updating the libraries with git pull and tried importing functions in a more verbose way but I canāt get it to go.
If you use an editable install of fastai2, you also need an editable install of fastcore, and you have to make sure you are always updating them both together.
Seem to be getting the same error too. Running on Google Colab. Had been working a few days ago. Should be using the latest master because did a fresh install. Looks like that line was also changed 2 days ago.
NameError Traceback (most recent call last)
<ipython-input-4-97b9dd603d64> in <module>()
----> 1 from fastai2.vision.all import *
4 frames
/usr/local/lib/python3.6/dist-packages/fastai2/layers.py in <module>()
271
272 # Cell
--> 273 @log_args
274 @funcs_kwargs
275 class BaseLoss():
NameError: name 'log_args' is not defined
I downgraded pytorch to 1.2.0 as recommended. It seems working, then I update fastai2, it downgrade my fastcore and have the following error. It unstalled my fastcore, but it seems install the same version again. Any idea to fix it?
ERROR: fastai2 0.0.17 has requirement torch>=1.3.0, but youāll have torch 1.2.0 which is incompatible.
ERROR: fastai2 0.0.17 has requirement torchvision>=0.5, but youāll have torchvision 0.4.0 which is incompatible. I updated my fastcore and the nameError ālog_argsā is not definedā and now this error shown up again. Current fastcore=0.1.17.
fastai requires PyTorch 1.3.1 minimum, so you need to install this version (or wait for PyTorch to release it if itās not available already).
As for the log_args, itās because you need to do an editable install of fastcore to go with your editable install of fastai2.
I think Windows may not be appropriate for DL practice. Although I use Windows for DL, I find it better to train the methods in Cloud or Linux environment.
OK.
I just want to clarify the following side note from Jeremy is for Colab in Window and not window in general. I will try the latest pytorch. Thanks.
A side note for windows:
Currently, pytorch==1.2.0 torchvision==0.4.0 is recomended and working (1.3.1 is not available on offical pytorch site and 1.4 is most recent but not working properly.)
In
Yes, but this a point in the topic that has not been updated in a while (this is a wiki, so if some new PyTorch versions are available on Windows, fell free to update it).
We are pinned to PyTorch 1.3.1 and torchivision 0.5.0 minimum right now, as we rely on behavior introduced in those versions. PyTorch 1.4.0 used to not work with fastai v2 but this has all been fixed now, so check you can install it properly on Windows and then edit the mentioned topic .
As Sylvain mentioned the docs are out-dated. Here is my setup info if it helps:
fastai2 Version: 0.0.17
nbdev Version: nbdev not found
fastprogress Version: 0.2.3
fastpages Version: fastpages not found
python Version: 3.7.7 (default, Mar 23 2020, 23:19:08) [MSC v.1916 64 bit (AMD64)]
torchvision: 0.5.0
torch version: 1.4.0
I followed the directions on the github page and also installed pytorch as I have Cuda9.2 from the pytorch website: In my case: conda install pytorch torchvision cudatoolkit=9.2 -c pytorch -c defaults -c numba/label/dev
I added that note and at that time it was problematic. I didnāt test the very recent 1.4 version pytorch. If you use conda you can install 1.3.1 with specifed version. Still you can not find 1.3.1 on the official pytorch site. You can install multiple cuda versions with the correct path naming.
.