Séb Arnold seba-1511

Google DeepMind
Mountain View
http://sebarnold.net

Recently created

Least recently created

Recently updated

Least recently updated

seba-1511 / jax2torch.py

Created November 21, 2022 17:12 — forked from mattjj/jax2torch.py

	import torch
	import torch.utils.dlpack
	import jax
	import jax.dlpack

	# A generic mechanism for turning a JAX function into a PyTorch function.

	def j2t(x_jax):
	x_torch = torch.utils.dlpack.from_dlpack(jax.dlpack.to_dlpack(x_jax))
	return x_torch

seba-1511 / Blind75LeetCodeQuestions.md

Last active February 27, 2023 00:39 — forked from krishnadey30/LeetCodeQuestions.md

Curated List of Top 75 LeetCode

Array

seba-1511 / pytorch-diffopt-proposal.py

Last active August 20, 2020 23:24

	inner_model = nnModuleModel(input_size=input_size, output_size=num_classes)
	metamodel = nnModuleModel(input_size=input_size, output_size=1)

	step = 0

	metaopt = Optim(metamodel.parameters(), with_no_grad=True) # conventional optimizer with "with no_grad"

	for epoch in num_epochs:

	# Normal training loop

seba-1511 / learn2learn-reptile-fix#132.py

Created May 15, 2020 19:28

Working implementation of Reptile

	#!/usr/bin/env python3

	"""
	TODO:
	* Remove dependency on ppt and randopt.
	* Replace get_problems with mini-imagenet dataset.
	"""

	import os
	import random

seba-1511 / hello_world.md

Created April 4, 2020 19:31 — forked from lava/hello_world.md

Hello, world: Deep analysis of a shallow program.

Hello, world!

Please explain in detail what will happen if the following program is executed:

#include <iostream>

int main() {
    std::cout << "Hello, world!" << std::endl;
}

seba-1511 / cython_tricks.md

Created February 24, 2020 02:59 — forked from ctokheim/cython_tricks.md

cython tricks

Cython

Cython has two major benefits:

Making python code faster, particularly things that can't be done in scipy/numpy
Wrapping/interfacing with C/C++ code

Cython gains most of it's benefit from statically typing arguments. However, statically typing is not required, in fact, regular python code is valid cython (but don't expect much of a speed up). By incrementally adding more type information, the code can speed up by several factors. This gist just provides a very basic usage of cython.

seba-1511 / config_dict.py

Created February 21, 2019 18:40 — forked from JACKHAHA363/config_dict.py

A principled way to have config dictionary that can be saved/restored, support pre-defined type and default values, and support cmd parsing

	import yaml
	import argparse

	class _ConfigDict(dict):
	""" A subclass of dict that supports required args
	"""
	# Change this to customize it
	fields = []

	@classmethod

seba-1511 / ram_image_example.py

Created November 19, 2018 04:52 — forked from alexlee-gk/ram_image_example.py

Get color and depth image as numpy array using Panda3d 1.10.0 and python 3.5

	import numpy as np
	import time
	import cv2
	from direct.showbase.ShowBase import ShowBase
	from panda3d.core import FrameBufferProperties, WindowProperties
	from panda3d.core import GraphicsPipe, GraphicsOutput
	from panda3d.core import Texture
	from panda3d.core import loadPrcFileData

	loadPrcFileData('', 'show-frame-rate-meter true')

seba-1511 / example

Created September 15, 2017 19:02 — forked from takluyver/example

Debug on interrupt

	$ python3 interrupt-debug.py
	0
	1
	2
	3
	^C--Return--
	> /home/takluyver/scratch/interrupt-debug.py(6)handle_sigint()->None
	-> pdb.set_trace()
	(Pdb) p a
	4

seba-1511 / dtrpo-bench

Created November 17, 2016 10:04

This file will contain the explanations to run dtrpo sequentially, and how to reproduce the results of algorithm "dtrpo-bench".