reshamas / fastai_deeplearn_part1 Goto Github PK
View Code? Open in Web Editor NEWNotes for Fastai Deep Learning Course
Home Page: http://www.fast.ai
License: MIT License
Notes for Fastai Deep Learning Course
Home Page: http://www.fast.ai
License: MIT License
Hi, I followed the directions in README exactly, however when I opened lesson1.ipynb, I got the following warning:
Important: This notebook will only work with fastai-0.7.x. Do not try to run any fastai-1.x code from this path in the repository because it will load fastai-0.7.x
As a newbie, I don't know which fastai version got installed on my machine. Digging around the forums, it seems that version 1 gets installed by default. So I am not sure how to proceed at this point.
Any help would be greatly appreciated.
Since this is the primary setup link referenced in lesson 1, it would be really important to keep it accurate and simple for beginners.
Jeremy said it will be shared in Jan, 2018.
followed https://github.com/reshamas/fastai_deeplearn_part1/blob/e8fb7e9eead518432a2c032c6c22a0d7db9fd9c7/tools/aws_ami_gpu_setup.md but cannot connect when I type jupyter notebook
and enter the IP address.
the error message that is coming up is thid after doing $conda update env .How to solve this
CondaVerificationError: The package for pytorch located at /home/paperspace/anaconda3/pkgs/pytorch-0.3.1-py36_cuda9.0.176_cudnn7.0.5_2
appears to be corrupted. The path 'lib/python3.6/site-packages/torch/_C.cpython-36m-x86_64-linux-gnu.so'
specified in the package manifest cannot be found.
I have the following problem while trying to update the python library... Please help...
(fastai) paperspace@ps1dkrpha:~/fastai$ conda env update
Using Anaconda API: https://api.anaconda.org
Fetching package metadata .................
Solving package specifications: .
Warning: 2 possible package resolutions (only showing differing packages):
CondaVerificationError: The package for pytorch located at /home/paperspace/anaconda3/pkgs/pytorch-0.3.1-py36_cuda9.0.176_cudnn7.0.5_2
appears to be corrupted. The path 'lib/python3.6/site-packages/torch/_C.cpython-36m-x86_64-linux-gnu.so'
specified in the package manifest cannot be found.
CondaVerificationError: The package for pytorch located at /home/paperspace/anaconda3/pkgs/pytorch-0.3.1-py36_cuda9.0.176_cudnn7.0.5_2
appears to be corrupted. The path 'lib/python3.6/site-packages/torch/lib/libTHC.so.1'
specified in the package manifest cannot be found.
CondaVerificationError: The package for pytorch located at /home/paperspace/anaconda3/pkgs/pytorch-0.3.1-py36_cuda9.0.176_cudnn7.0.5_2
appears to be corrupted. The path 'lib/python3.6/site-packages/torch/lib/libcudnn-7b07b0f1.so.7'
specified in the package manifest cannot be found.
Seems not working
Also the cuda90 modules cause some problem in some older machines. (Would try in new Windows machines later.)
When trying to apply the promo code FASTAI3BDG I'm getting:
Warning: Could not claim promo code (code is expired)
Is there another promo code?
The promo code needs to be updated.
There is a section that I think could improve the docs in https://github.com/reshamas/fastai_deeplearn_part1/blob/master/tools/paperspace.md
Some users may need to run the line sudo ufw allow 8888
to allow traffic over the port 8888, as sometimes it is not open by default.
This was necessary for me to be able to connect with jupyter notebook
Source: http://forums.fast.ai/t/paperspace-jupyter-notebook-server-setup/6544
Your instructions are a little out of date. Selecting the fast.ai template requires prior approval.
Regards,
Jay
Hi there,
I could not find an ami for fastai p1 v2 in AWS Community AMIs.
Any help to resolve this is much appreciated.
Best regards,
Going through the setup instructions, ran into this:
This instance type has not been enabled in your account yet. For some GPU types we require that you tell us a bit more about your use case before we enable access. This is designed to reduce fraud and thus keep our prices low.
Tell us a bit more about your use case and we will prioritize your request (required)
I entered "Deep Learning for Coders Course" and clicked submit and an hour later, nothing, no response yet. I've got $15 credit on my account (using the coupon code) and have also entered my credit card billing information.
Just throwing this out there as maybe this isn't the best option for the course, or perhaps you want to note this experience on the README.
Other than that, thanks for the great setup instructions!
Hi.
When i try to create a symbolic link of a directory by using ln -s command . A link is created and is being shown as file instead of a folder shortcut. Also when i try to open it , 404 : Not Found error is displayed. Also i can't use it as a directory in jupyter notebook. Can you please tell me its solution?
In my case jupyter notebook is taking too long to respond on my local pc. I followed all the steps.
Also, can you tell me how to copy files from my local pc to cloud pc
The promo code FASTAI6GKZ is no longer valid.
this has been a fantastic resource, thanks.
I had some trouble with the part which says: "I need to replace localhost section with my IP address.
What would have been clearer to me: "I need to replace localhost section with the public IP address of my paperspace machine"
Trying to follow instructions here:
I can get through part 2, but when I try to use the command jupyter notebook --generate-config
I get the following error:
Traceback (most recent call last):
File "/home/ubuntu/src/anaconda3/envs/fastai/lib/python3.6/site-packages/notebook/services/sessions/sessionmanager.py", line 10, in <module>
import sqlite3
File "/home/ubuntu/src/anaconda3/envs/fastai/lib/python3.6/sqlite3/__init__.py", line 23, in <module>
from sqlite3.dbapi2 import *
File "/home/ubuntu/src/anaconda3/envs/fastai/lib/python3.6/sqlite3/dbapi2.py", line 27, in <module>
from _sqlite3 import *
ImportError: libncurses.so.6: cannot open shared object file: No such file or directory
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/ubuntu/src/anaconda3/envs/fastai/bin/jupyter-notebook", line 7, in <module>
from notebook.notebookapp import main
File "/home/ubuntu/src/anaconda3/envs/fastai/lib/python3.6/site-packages/notebook/notebookapp.py", line 85, in <module>
from .services.sessions.sessionmanager import SessionManager
File "/home/ubuntu/src/anaconda3/envs/fastai/lib/python3.6/site-packages/notebook/services/sessions/sessionmanager.py", line 13, in <module>
from pysqlite2 import dbapi2 as sqlite3
ModuleNotFoundError: No module named 'pysqlite2'
After this I tried to instal pysqlite2 but it isn't available to install with the following output:
PackageNotFoundError: Packages missing in current channels:
- pysqlite2
We have searched for the packages in the following channels:
- https://conda.anaconda.org/soumith/linux-64
- https://conda.anaconda.org/soumith/noarch
- https://repo.continuum.io/pkgs/main/linux-64
- https://repo.continuum.io/pkgs/main/noarch
- https://repo.continuum.io/pkgs/free/linux-64
- https://repo.continuum.io/pkgs/free/noarch
- https://repo.continuum.io/pkgs/r/linux-64
- https://repo.continuum.io/pkgs/r/noarch
- https://repo.continuum.io/pkgs/pro/linux-64
- https://repo.continuum.io/pkgs/pro/noarch
From here I tried to install sqlite3 and got the same error.
I decided to test the environment as well. However, I wasn't able to get all the tests to pass when I ran pytest tests
and received the result of 49 tests passing, 1 failing, and 1 being skipped. The following message was displayed for the one I think failed:
_________________________________________________________ test_T __________________________________________________________
HalfTensorMock = <MagicMock name='HalfTensor' id='139635496833544'>
@mock.patch("fastai.core.torch.cuda.HalfTensor")
def test_T(HalfTensorMock):
tensor = torch.ones([1, 2])
> assert core.T(tensor) is tensor
E assert \n 1 1\n[torch.cuda.FloatTensor of size 1x2 (GPU 0)]\n is \n 1 1\n[torch.FloatTensor of size 1x2]\n
E + where \n 1 1\n[torch.cuda.FloatTensor of size 1x2 (GPU 0)]\n = <function T at 0x7eff810ad1e0>(\n 1 1\n[torch.FloatTensor of size 1x2]\n)
E + where <function T at 0x7eff810ad1e0> = core.T
tests/test_core.py:38: AssertionError
==================================================== warnings summary =====================================================
tests/test_core.py::test_T
/home/ubuntu/anaconda3/envs/fastai/lib/python3.6/site-packages/torch/cuda/__init__.py:97: UserWarning:
Found GPU0 GRID K520 which is of cuda capability 3.0.
PyTorch no longer supports this GPU because it is too old.
warnings.warn(old_gpu_warn % (d, name, major, capability[1]))
-- Docs: http://doc.pytest.org/en/latest/warnings.html
=============================== 1 failed, 49 passed, 1 skipped, 1 warnings in 5.39 seconds ================================
I'm not sure how to proceed, this probably isn't a fast.ai error but it might need to be addressed in your code or instructions.
here is my code and I want to insert a bar for each epoch and also want to generate an accuracy graph for each epoch please help me out
`learn.lr_find()
learn.recorder.plot()
lr = 0.01
learn.fit_one_cycle(9, slice(lr))
learn.save('classif_restnet50')`
lesson1-rf.ipynb needs to be updated: first cell in the section "Reducing over-fitting" - "Subsampling", proc_df i called.
df_trn, y_trn = proc_df(df_raw,'SalePrice')
and throws
ValueError: too many values to unpack (expected 2)
proc_df returns 3 variables rather than two. From the documentation:
[x, y, nas, mapper(optional)]:
So it could be df_trn, y_trn, foo = proc_df(df_raw,'SalePrice')
When I try to access the fast.ai template under public section, it asked me for more information before giving me access.
I put donw for fast ai study
and it has been couple days now but I have not yet heard back from Paperspace, is there anything I should put specifically?
Thanks
I just got a P4000 on paperspace so this might not be true anymore.
I am getting a warning regarding one of the external libraries from fast.ai.
I am running most current versions of Python and Jupyter Notebook.
# This file contains all the main external libs we'll use from fastai.imports import * /home/paperspace/anaconda3/envs/fastai/lib/python3.6/site-packages/sklearn/ensemble/weight_boosting.py:29: DeprecationWarning: numpy.core.umath_tests is an internal NumPy module and should not be imported. It will be removed in a future NumPy release. from numpy.core.umath_tests import inner1d In [ ]:
Do I just ignore this line and exclude it? What would you recommend? I have made sure that all of my files are updated via Paperspace as well.
Did paperspace change the template process? I cannot assign the fastai template during the creation of the machine. Later it does provide the option to assign a template, but I still cannot access public templates.
I am on a windows machine and �the European Server. I have asked paperspace support.
Please provide a brief instruction on how to do brew install ssh-copy-id
for windows.
Thanks!
Just thought this should be added to the tutorial, you know how to state the reason why you want to use a public template and then await access to them.
This happens when choosing the server in Europe.
can I setup fastai learning env in my Centos7.1 with cuda8.0, I don't want to change my cuda8.0 to cuda90~~
The following command will incorrectly count the number of files (will overestimate by one), because the first line of output of ls -l
on some versions of Linux as well as Mac OS X will give the total number of files:
!ls -l {PATH}train/cats | grep ^[^d] | wc -l
I suggest correcting it to the following version which will disregard the first line of ls -l
:
!ls -l {PATH}train/cats | grep ^[^dt] | wc -l
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.