Updated 4/11/2018
Here's my experience of installing the NVIDIA CUDA kit 9.0 on a fresh install of Ubuntu Desktop 16.04.4 LTS.
""" | |
Python script to find the memory allocations | |
i. over a module/function that is running | |
ii. between the loops | |
It is a script so that it can be torn apart for your own use-cases. | |
Based on this nice article on trying to find memory leaks in python: | |
https://www.fugue.co/blog/diagnosing-and-fixing-memory-leaks-in-python.html | |
But the code is non-existent and isn't reproducible as of 2023. |
//=================================================================== | |
// File: circular_buffer.cpp | |
// | |
// Desc: A Circular Buffer implementation in C++. | |
// | |
// Copyright © 2019 Edwin Cloud. All rights reserved. | |
// | |
//=================================================================== | |
//------------------------------------------------------------------- |
//=================================================================== | |
// File: circular_buffer.cpp | |
// | |
// Desc: A Circular Buffer implementation in C++. | |
// | |
// Copyright © 2019 Edwin Cloud. All rights reserved. | |
// | |
//=================================================================== | |
//------------------------------------------------------------------- |
அ | |
ஆ | |
இ | |
ஈ | |
உ | |
ஊ | |
எ | |
ஏ | |
ஐ | |
ஒ |
""" | |
Want to generate sorted random numbers optimally? Here you go. | |
Traditionally if you wanted a list of sorted random numbers, | |
you generate `n` random numbers and then sort them. which at worst | |
case has: | |
Time Complexity : O(n log n) # for generating & sorting. | |
Space Complexity: O(n) # for storing all of them in array/set. |
""" | |
This gist comes out of frustration that I couldn't have a | |
random-int-without-replacement-generator-given-a-range. | |
random.sample, and np.random.choice(replace=False) both fail on really large numbers. | |
Python crashes saying OOM & segfaults. | |
Problem was for smaller `n` (<5 million) they optimized for linear/sub-linear | |
times and linear storage (set, pool-tracking list). |
""" | |
Trying to find a way to handle channels-like situations (producer->consumer) in Python | |
using asyncio.Queue. | |
borrowed code actually from: | |
https://web.archive.org/web/20210507031446/https://asyncio.readthedocs.io/en/latest/producer_consumer.html | |
""" | |
import asyncio | |
import random |
from keras.preprocessing.text import Tokenizer | |
from keras.preprocessing.sequence import pad_sequences | |
from keras.utils.np_utils import to_categorical | |
from keras.models import Sequential | |
from keras.layers.embeddings import Embedding | |
from keras.layers.core import Dense, Dropout | |
from keras.layers import LSTM | |
from keras.regularizers import l2 | |
from keras.optimizers import Adam |
Updated 4/11/2018
Here's my experience of installing the NVIDIA CUDA kit 9.0 on a fresh install of Ubuntu Desktop 16.04.4 LTS.