kakadzhun

4 ms·

1.
by kakadzhun 4y ago
If Reinforcement Learning is anything to go by, then a naive implementation of learning from past models will overfit to the previous model and start performing worse than even earlier models. There was a paper by someone @ Microsoft who tr
2.
by kakadzhun 4y ago
Try this resource instead. Although it is 100 pages, the introductory part is already useful in and of itself! https://zenodo.org/record/1146014/files/GNU_Parallel_2018.pd...
3.
by kakadzhun 4y ago
Her emotional response was definitely not wrong nor an issue. Criticism is hard to take for anyone, but the tweets weren't even criticism, just plain mocking. What do you believe was her mistake? Sobbing about it for an emotional relea
4.
by kakadzhun 4y ago
import numpy as np print(np.__version__) a = np.array([1,2,3]) b = np.array([1]) print(a in b) print(b in a) print(np.isin(a, b)) print(np.isin(b, a)) print(np.isin(a.astype(float), b)) print(np.i
5.
by kakadzhun 5y ago
I've always wondered whether there is an upper limit to mental endurance regarding the will to live, e.g. would you grow tired of life after a certain point? What if you were truly immortal, wouldn't you be driven mad soon enough?
6.
by kakadzhun 5y ago
Script to filter out records of a .jsonl file based on some property (e.g. only want records w/ a timestamp greater than 163623...): import sys import orjson import argparse if __name__ == '__main__':
7.
by kakadzhun 5y ago
Even if Reinforcement Learning is 'enough', it will be held back by whatever methods used to implement it (e.g. Deep Neural Networks). As the algorithms get more advanced, from some point onwards, to build the general AI, you firs
8.
by kakadzhun 5y ago
Reduce, reuse, recycle. According to a HN user (don't remember who) this was THE slogan before making everything "green". "Green" makes it easier to sell you the feeling of being environmentally friendly, while redu
9.
by kakadzhun 5y ago
https://github.com/joosephook/sqlite3-numpy Minimal working example of storing numpy vectors in sqlite3 using Python: import sqlite3 from scipy.spatial.distance import cdist import numpy as np class
10.
by kakadzhun 6y ago
Could you elaborate on what you believe you'd get for free in a static language compared to type-annotated and -checked Python? I'm curious as I'm in a similar situation, but not in a company, but as a researcher.
11.
by kakadzhun 6y ago
Consider this: In [1]: np.argmax(np.array([[1,2,3],[1,9,3],[1,2,3]]).flat) Out[2]: 4
12.
by kakadzhun 6y ago
If you want to play around with Spinning Up in a Docker container, then make sure you git clone the repository, then pip install -e repository. For whatever reason, if you try to directly install it with pip, it doesn't work, at least
13.
by kakadzhun 6y ago
I had the same problem on Ubuntu 18.04, terrible lag spikes and similar messages regarding i915. After installing the least new kernel (5.3.1), it solved the issue: not one hang since.