tmp

15 minute read

Jupyterlab

jupyter lab 실행시 no kernel현상

(my-torch-env3.6) jhseo@jhseo-Inspiron-7559:~/anaconda3/envs/my-torch-env3.6$ jupyter lab
[W 2021-03-26 15:32:34.434 ServerApp] The module 'jupyter_kite' could not be found. Are you sure the extension is installed?
[I 2021-03-26 15:32:34.444 ServerApp] jupyterlab | extension was successfully linked.
[W 2021-03-26 15:32:34.460 ServerApp] The module 'ipyparallel.nbextension' could not be found. Are you sure the extension is installed?
[I 2021-03-26 15:32:34.460 ServerApp] nbclassic | extension was successfully linked.
[I 2021-03-26 15:32:34.483 ServerApp] The port 8888 is already in use, trying another port.
[I 2021-03-26 15:32:34.488 LabApp] JupyterLab extension loaded from /home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/jupyterlab
[I 2021-03-26 15:32:34.488 LabApp] JupyterLab application directory is /home/jhseo/anaconda3/envs/my-torch-env3.6/share/jupyter/lab
[I 2021-03-26 15:32:34.491 ServerApp] jupyterlab | extension was successfully loaded.
[I 2021-03-26 15:32:34.494 ServerApp] nbclassic | extension was successfully loaded.
[I 2021-03-26 15:32:34.495 ServerApp] Serving notebooks from local directory: /home/jhseo/anaconda3/envs/my-torch-env3.6
[I 2021-03-26 15:32:34.495 ServerApp] Jupyter Server 1.4.1 is running at:
[I 2021-03-26 15:32:34.495 ServerApp] http://localhost:8889/lab?token=ce3eb05e0eff2d69b831eb6bd752327a111e3a4ad8cd4809
[I 2021-03-26 15:32:34.495 ServerApp]  or http://127.0.0.1:8889/lab?token=ce3eb05e0eff2d69b831eb6bd752327a111e3a4ad8cd4809
[I 2021-03-26 15:32:34.495 ServerApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
[C 2021-03-26 15:32:34.501 ServerApp] 
    
    To access the server, open this file in a browser:
        file:///home/jhseo/.local/share/jupyter/runtime/jpserver-23664-open.html
    Or copy and paste one of these URLs:
        http://localhost:8889/lab?token=ce3eb05e0eff2d69b831eb6bd752327a111e3a4ad8cd4809
     or http://127.0.0.1:8889/lab?token=ce3eb05e0eff2d69b831eb6bd752327a111e3a4ad8cd4809
Opening in existing browser session.
[I 2021-03-26 15:32:39.442 LabApp] Build is up to date
TermSocket.open: 1
[I 2021-03-26 15:32:39.803 ServerApp] New terminal with specified name: 1
TermSocket.open: Opened 1
Websocket closed
[W 2021-03-26 15:32:40.188 ServerApp] Notebook MolTrans_Study.ipynb is not trusted
[I 2021-03-26 15:32:40.708 ServerApp] Kernel started: 37b30e6c-144e-47d8-8c7c-ec32ec2edded
Traceback (most recent call last):
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel_launcher.py", line 15, in <module>
    from ipykernel import kernelapp as app
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/__init__.py", line 2, in <module>
    from .connect import *
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/connect.py", line 11, in <module>
    from IPython.core.profiledir import ProfileDir
ModuleNotFoundError: No module named 'IPython.core.profiledir'
[I 2021-03-26 15:32:43.709 ServerApp] AsyncIOLoopKernelRestarter: restarting kernel (1/5), new random ports
Traceback (most recent call last):
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel_launcher.py", line 15, in <module>
    from ipykernel import kernelapp as app
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/__init__.py", line 2, in <module>
    from .connect import *
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/connect.py", line 11, in <module>
    from IPython.core.profiledir import ProfileDir
ModuleNotFoundError: No module named 'IPython.core.profiledir'
-[I 2021-03-26 15:32:46.716 ServerApp] AsyncIOLoopKernelRestarter: restarting kernel (2/5), new random ports
Traceback (most recent call last):
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel_launcher.py", line 15, in <module>
    from ipykernel import kernelapp as app
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/__init__.py", line 2, in <module>
    from .connect import *
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/connect.py", line 11, in <module>
    from IPython.core.profiledir import ProfileDir
ModuleNotFoundError: No module named 'IPython.core.profiledir'
[I 2021-03-26 15:32:49.723 ServerApp] AsyncIOLoopKernelRestarter: restarting kernel (3/5), new random ports
Traceback (most recent call last):
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel_launcher.py", line 15, in <module>
    from ipykernel import kernelapp as app
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/__init__.py", line 2, in <module>
    from .connect import *
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/connect.py", line 11, in <module>
    from IPython.core.profiledir import ProfileDir
ModuleNotFoundError: No module named 'IPython.core.profiledir'
[I 2021-03-26 15:32:52.735 ServerApp] AsyncIOLoopKernelRestarter: restarting kernel (4/5), new random ports
Traceback (most recent call last):
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel_launcher.py", line 15, in <module>
    from ipykernel import kernelapp as app
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/__init__.py", line 2, in <module>
    from .connect import *
  File "/home/jhseo/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/ipykernel/connect.py", line 11, in <module>
    from IPython.core.profiledir import ProfileDir
ModuleNotFoundError: No module named 'IPython.core.profiledir'
[W 2021-03-26 15:32:55.750 ServerApp] AsyncIOLoopKernelRestarter: restart failed
[W 2021-03-26 15:32:55.750 ServerApp] Kernel 37b30e6c-144e-47d8-8c7c-ec32ec2edded died, removing from map.

여러 키워드로 구글링해봄.

ImportError: cannot import name 'generator_to_async_generator'

해결
- In my case with python3.7, I uninstalled both ipython and prompt_toolkit as shown by @stas00 and just installed ipython so that it installs the compatible prompt_toolkit itself

pip uninstall -y ipython prompt_toolkit
pip install ipython

PyTorch

torch.nn.BCELoss:

Binary Cross Entropy
target - output사이의 bce측정기준 생성. 참고

Tensor:

tensors are (kind of) like np.arrays.
All tensors are immutable like Python numbers and strings:
- you can never update the contents of a tensor, only create a new one.

torch.Tensor class중 backward function은 Computes the gradient of current tensor w.r.t. graph leaves.

w.r.t.:with regard to(..에 관해서)

默认情况下是没有grad梯度属性的。检查一个tensor是否有grad属性可以通过tensor.requires_grad来查看，如果tensor中有grad属性，则返回Ture，反之返回False。

scalar: …

torch.optim은 다양한 최적화 알고리즘을 구현하는 패키지입니다. 가장 일반적으로 사용되는 방법은 이미 지원되고 있으며 인터페이스는 충분히 일반적이므로 향후 더 복잡한 방법도 쉽게 통합 할 수 있습니다.

Backward function in PyTorch

Thus, by default, backward() is called on a scalar tensor and expects no arguments. 출처

Variable API has been deprecated

How to declare a scalar as a parameter in pytorch?

x = torch.randn(1,1, requires_grad=True)

Variable API has been deprecated. official docs

CUDA out of memory

Error message

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-18-866d0e46d218> in <module>
      2 #biosnap interaction times 1e-6, flat, batch size 64, len 205, channel 3, epoch 50
      3 s = time()
----> 4 model_max, loss_history = main(1, 5e-6)
      5 e = time()
      6 print(e-s)

<ipython-input-17-36825a101074> in main(fold_n, lr)
    113         model.train()
    114         for i, (d, p, d_mask, p_mask, label) in enumerate(training_generator):
--> 115             score = model(d.long().cuda(), p.long().cuda(), d_mask.long().cuda(), p_mask.long().cuda())
    116 
    117             label = Variable(torch.from_numpy(np.array(label)).float()).cuda()

~/anaconda3/envs/my-torch-env3.6/lib/python3.6/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
    725             result = self._slow_forward(*input, **kwargs)
    726         else:
--> 727             result = self.forward(*input, **kwargs)
    728         for hook in itertools.chain(
    729                 _global_forward_hooks.values(),

~/anaconda3/envs/my-torch-env3.6/MolTrans/models.py in forward(self, d, p, d_mask, p_mask)
     90 
     91         # repeat to have the same tensor size for aggregation
---> 92         d_aug = torch.unsqueeze(d_encoded_layers, 2).repeat(1, 1, self.max_p, 1) # repeat along protein size
     93         p_aug = torch.unsqueeze(p_encoded_layers, 1).repeat(1, self.max_d, 1, 1) # repeat along drug size
     94 

RuntimeError: CUDA out of memory. Tried to allocate 640.00 MiB (GPU 0; 3.95 GiB total capacity; 2.49 GiB already allocated; 587.12 MiB free; 2.53 GiB reserved in total by PyTorch)

TORCH.UTILS.DATA

At the heart of PyTorch data loading utility is the torch.utils.data.DataLoader class. It represents a Python iterable over a dataset, with support for
- map-style and iterable-style datasets,
- customizing data loading order,
- automatic batching,
- single- and multi-process data loading,
- automatic memory pinning.

DataLoader(dataset, batch_size=1, shuffle=False, sampler=None,
           batch_sampler=None, num_workers=0, collate_fn=None,
           pin_memory=False, drop_last=False, timeout=0,
           worker_init_fn=None, *, prefetch_factor=2,
           persistent_workers=False)

CLASS torch.utils.data.Dataset

An abstract class representing a Dataset.

MolTrans – FCS

subword-nmt
- About
  - Unsupervised Word Segmentation for Neural Machine Translation and Text Generation
- Installation
  - pip install subword-nmt

Python

enumerate

Return an enumerate object. iterable must be a sequence. returns a tuple containing a count.

seasons = ['Spring', 'Summer', 'Fall', 'Winter']
list(enumerate(seasons))  # [(0, 'Spring'), (1, 'Summer'), (2, 'Fall'), (3, 'Winter')]

list(enumerate(seasons, start=1))  # [(1, 'Spring'), (2, 'Summer'), (3, 'Fall'), (4, 'Winter')]

What is the difference between arguments and parameters?

Parameters are defined by the names that appear in a function definition, whereas arguments are the values actually passed to a function when calling it. Parameters define what types of arguments a function can accept.

For example, given the function definition:

def func(foo, bar=None, **kwargs):
    pass

foo, bar and kwargs are parameters of func. However, when calling func, for example:

func(42, bar=314, extra=somevar)

the values 42, 314, and some var are arguments.

Role of Underscore(_) in Python

Use In Interpreter

>>> 54 + 45
99
>>> _
99
>>> 1 + _
100
>>> 2*_
200

Ignoring Values

## ignoring a value
a, _, b = (1, 2, 3) # a = 1, b = 3
print(a, b)

## ignoring multiple values
## *(variable) used to assign multiple value to a variable as list while unpacking
## it's called "Extended Unpacking", only available in Python 3.x
a, *_, b = (7, 6, 5, 4, 3, 2, 1)
print(a, b)

1 3
7 1

Use In Looping

Naming

제일 중요한거.
Underscore(_) can be used to name variables, functions and classes, etc..,

Single Pre Underscore:- _variable

Single Pre Underscore is only meant to use for the internal use.

## filename:- my_functions.py

def func():
  return "datacamp"

def _private_func():
  return 7

>>> from my_functions import *
>>> func()
'datacamp'
>>> _private_func()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name '_private_func' is not defined

Signle Post Underscore:- variable_
Double Pre Underscores:- __variable
Double Pre and Post Underscores:- __variable__

Class attribute vs. Instance attribute

Python glossary

용어정리(추천)

attribute

A value associated with an object which is referenced by name using dotted expressions. For example, if an object o has an attribute a it would be referenced as o.a.

pep뜻

PEP 0 – Index of Python Enhancement Proposals (PEPs) peps

method

A function which is defined inside a class body.

Machine Learning

AUPRC

The area under the precision-recall curve (AUPRC) is a useful performance metric for imbalanced data in a problem setting where you care a lot about finding the positive examples.

AUROC vs. AUPRC

Usually you would obtain the same conclusion based on both measures. 출처.
ROC AUC is the area under the curve where x is false positive rate (FPR) and y is true positive rate (TPR).
PR AUC is the area under the curve where x is recall and y is precision. 출처.

Confusion matrix

a confusion matrix (error matrix), is a specific table layout that allows visualization of the performance of an algorithm, typically a supervised learning one.

Basic-Confusion-matrix

헷갈릴땐 (예측을) True/False 해서 (실제값을) Positive/Negative.
$TP+FN=P$, $FP+TN=N$, $TP+FP=P’$, $FN+TN=N’$.

Accuracy: 전체중 참인 예측의 비율 $\frac{TP+TN}{P+N}$
Error-rate: 전체 관측치 중 실제값과 예측치가 다른 정도 $\frac{FP+FN}{P+N}=\frac{FP+FN}{(TP+FN)+(FP+TN)}$

Precision: 전체Positive 중 참Positive인 비율.PPV. $\frac{TP}{P'} =\frac{TP}{TP+FP}$
Recall: 실제값Positive중 참Positive의 비율 $\frac{TP}{P} = \frac{TP}{TP+FN}$
Sensitivity: 실제값이 True인 관측치 중 예측치가 적중한 정도. TPR. $\frac{TP}{P}=\frac{TP}{TP+FN}$
Specificity: 실제값이 False인 관측치 중 예측치가 적중한 정도. FPR. $\frac{TN}{N}=\frac{TN}{FP+TN}$

$\mathrm{Precision}=\mathrm{PPV} , \mathrm{PPV}\neq \mathrm{FPR}$

$\mathrm{recall} = \mathrm{TPR} = \mathrm{Sensitivity}$

F1 Score: Precision과 Recall의 조화평균 $\frac{2*\mathrm{Precision}*\mathrm{Recall}}{\mathrm{Precision}+\mathrm{Recall}} = \frac{2*TP}{2*TP+FP+FN}$

Share on

Twitter Facebook LinkedIn

peter(星期八)

tmp

Jupyterlab

PyTorch

MolTrans – FCS

Python

Python glossary

Machine Learning

Share on

You may also enjoy

Java review(1)

ICT 실무 교육

VS Code 에서 c/c++ 사용하기

Interview