Skip to content

Instantly share code, notes, and snippets.

View agucova's full-sized avatar
👋
Hi there!

Agustín Covarrubias agucova

👋
Hi there!
View GitHub Profile
We can't make this file beautiful and searchable because it's too large.
System,Domain,Organization,Authors,Publication date,Reference,Link,Notability criteria,Notability criteria notes,Parameters,Parameters notes,Training compute (FLOP),Training compute notes,Training dataset,Training dataset notes,Training dataset size (datapoints),Dataset size notes,Abstract,Confidence,Country (from Organization),Organization categorization,Training time (hours),Training time notes,Training hardware,Model accessibility,Finetune compute notes,Hardware quantity,Hardware utilization,Base model,Citations,Compute cost notes,Finetune compute (FLOP),Epochs,Training compute cost (2023 USD),Batch size,Batch size notes
DeepSeek-Coder-V2-236B,Language,DeepSeek,"Qihao Zhu, Daya Guo, Zhihong Shao, Dejian Yang, Peiyi Wang, Runxin Xu, Y. Wu, Yukun Li, Huazuo Gao, Shirong Ma, Wangding Zeng, Xiao Bi, Zihui Gu, Hanwei Xu, Damai Dai, Kai Dong, Liyue Zhang, Yishi Piao, Zhibin Gou, Zhenda Xie, Zhewen Hao, Bingxuan Wang, Junxiao Song, Deli Chen, Xin Xie, Kang Guan, Yuxiang You, Aixin Liu, Qiushi Du, Wenjun Gao, Xuan
@agucova
agucova / to_webp.py
Created September 30, 2023 00:11
Script for converting <img> references in an Anki CSV to inline webp images.
#!/usr/bin/env python3
import typer
import pandas as pd
import requests
import base64
from bs4 import BeautifulSoup
from PIL import Image
from io import BytesIO
from typing import Optional
// Clean and simplify HTML
function cleanHTML(html: string): string {
const div = document.createElement('div');
div.innerHTML = html;
// Remove all class attributes and data-* attributes
div.querySelectorAll('*').forEach(el => {
el.removeAttribute('class');
Array.from(el.attributes).forEach(attr => {
if (attr.name.startsWith('data-')) {
set remindersOpen to application "Reminders" is running
set todoistToken to "<YOUR_TOKEN_HERE>"
set apiCall to "curl https://api.todoist.com/sync/v9/quick/add"
set listOfCommands to {}
tell application "Reminders"
@agucova
agucova / dist_refactor.py
Last active June 13, 2023 23:37
Automated libCST refactor for squigglepy
import libcst as cst
from libcst.codemod import CodemodTest, VisitorBasedCodemodCommand
from libcst.codemod.visitors import AddImportsVisitor
DISTS: dict[str, str] = {
"base": "BaseDistribution",
"complex": "ComplexDistribution",
"const": "ConstantDistribution",
"uniform": "UniformDistribution",
@agucova
agucova / test.bash
Created April 14, 2023 19:50
Script para testear proyecto completo de VHDL
#!/usr/bin/env bash
set -uo pipefail
trap 's=$?; echo "$0: Error on line "$LINENO": $BASH_COMMAND"; exit $s' ERR
IFS=$'\n\t'
# Run syntax check
echo "Checking syntax..."
nvc --std=2008 --syntax lib/**/*.vhd
nvc --std=2008 --syntax test/**/*.vhd
[flake8]
max-line-length = 88
extend-ignore = E203, W503, E501
exclude =
# No need to traverse our git directory
.git,
# There's no value in checking cache directories
__pycache__,
# This contains our built documentation
build,
+(/Users/agucova/.pyenv/bin/pyenv:23): enable -f /Users/agucova/.pyenv/bin/../libexec/pyenv-realpath.dylib realpath
+(/Users/agucova/.pyenv/bin/pyenv:30): '[' -z '' ']'
++(/Users/agucova/.pyenv/bin/pyenv:32): type -P greadlink readlink
++(/Users/agucova/.pyenv/bin/pyenv:32): head -1
+(/Users/agucova/.pyenv/bin/pyenv:32): READLINK=/opt/homebrew/bin/greadlink
+(/Users/agucova/.pyenv/bin/pyenv:33): '[' -n /opt/homebrew/bin/greadlink ']'
+(/Users/agucova/.pyenv/bin/pyenv:58): '[' -z /Users/agucova/.pyenv ']'
+(/Users/agucova/.pyenv/bin/pyenv:61): PYENV_ROOT=/Users/agucova/.pyenv
+(/Users/agucova/.pyenv/bin/pyenv:63): export PYENV_ROOT
+(/Users/agucova/.pyenv/bin/pyenv:65): '[' -z '' ']'
@agucova
agucova / requirements.md
Last active July 22, 2021 19:25
Technical requirements for Reguleque

Currently our setup for our search engine is merely an InstantSearch front-end to an open-source engine called TypeSense. We made specialized scripts to clean and load the government transparency data into the Typesense instance.

Currently, the front-end is public and has unlimited usage, but we want to limit the daily searches available for each visitor (similar to newspaper paywalls), while still allowing people to request a fair use account (for journalists, by example).

Technically this means:

  • Implement a back-end that acts as a proxy to the search requests to Typesense and checks for user authentication. This needs to implement full-fledged user management under reasonable security standards, preferabily implemented through existing authentication solutions or libraries.
  • Modifying the InstantSearch-ba
#
# .zshrc is sourced in interactive shells.
# It should contain commands to set up aliases,
# functions, options, key bindings, etc.
#
autoload -U compinit
compinit
#allow tab completion in the middle of a word