keuv-grvl

{
	"if __name__ == \"__main__\":": {
		"body": [
			"if __name__ == \"__main__\":\n\t$0"
		],
		"prefix": [
			"__name__",
			"__main__",
		]
	},

keuv-grvl

Performance for activation function

timeit applied to various Pytorch activation function on CPU and GPU, on a tensor of shape (20000,)

You must run this in a IPython session. Pytorch version is 2.0.0+cu117.

See run.py.

Results

keuv-grvl

import torch
import datasets


class HFDataset(torch.utils.data.Dataset):
  def __init__(self, dset: datasets.Dataset):
    self.dset = dset

  def __getitem__(self, idx):
    return self.dset[idx]

keuv-grvl

%pip install pure-python-adb imageio

from ppadb.client import Client as AdbClient

# plug your android device with USB debugging
client = AdbClient(host="127.0.0.1", port=5037)
device = client.devices()[0]

screencap = device.screencap()

keuv-grvl

Benchmark write speed

Write zeros to a device (using a mountpoint).

Command

# Choose one
FOLDER="$HOME"
FOLDER="/tmp"

keuv-grvl

gzip with tqdm progress bar

Prereq

pip install tqdm

Minimalistic

keuv-grvl

(properly) Lock and turn off your screen on Pop!_OS 22.04 LTS

Pop!_OS does not turn off your screen when you lock your screen. Where you hit tue Super+L shortcut, it seems to start some screensaver which display a black image and coincidently lock your session. But the screen itself does not go to its standby mode.

Here is my solution:

1. Nuke the <Super>l shortcut occurences:
   • List conf keys: {dconf dump / ; gsettings list-recursively } | less

• Use dconf-editor to unset them

keuv-grvl

@profile
def func1(i: float):
  import time
  time.sleep(i)


@profile
def func_mother():
  func1(1.2)
  a = 2 + 3

keuv-grvl

from fast_flask import Flask

app = Flask(__name__)

@app.get("/users/")  # equivalent to `@app.route("/users/")` from flask.Flask
def get_user_details(user_id: int = 123):
    return {"123": "bob"}


if __name__ == "__main__":

keuv-grvl

import gradio as gr
import numpy as np
import tensorflow as tf
import requests

# Load a classification model
model = tf.keras.applications.MobileNetV3Large(
    input_shape=None,
    alpha=1.0,
    minimalistic=False,