peterc

On Runpod with latest PyTorch image (2.4.0) with a GPU > 32GB VRAM (e.g. NVIDIA A100 80GB PCIe).

ssh in and:

apt update -y
apt install -y nano screen git
pip install git+https://github.com/huggingface/trl.git accelerate transformers datasets peft wandb tqdm ninja flash-attn

deepfates

import argparse
import json
import logging
import os
import re
import shutil
from concurrent.futures import ProcessPoolExecutor, as_completed
from dataclasses import dataclass
from datetime import datetime
from typing import Any, Callable, Dict, List, Literal, Optional, Tuple

cpursley

PostgreSQL is Enough

• Simplify: move code into database functions
• Just Use Postgres for Everything
• Just use Postgres
• PostgreSQL is the worlds’ best database
• Postgres is eating the database world
• Hacker News discussion

Background and Cron Jobs

yyscamper

package main

import (
	"fmt"
	"math"
	"strconv"
	"strings"
)

func ParseFloat(str string) (float64, error) {

hiwonjoon

import subprocess
import datetime
import numpy as np

THREAD_NUM=4

def get_video_info(fileloc) :
    command = ['ffprobe',
               '-v', 'fatal',
               '-show_entries', 'stream=width,height,r_frame_rate,duration',

askilondz

Adaptive Streaming has become the neccessity for streaming video and audio. Unfortantely, as of this post, there isn't a whole lot of tutorials that accumulate all of the steps to get this working. Hopefully this post achieves that. This post focuses on using Amazon Web Services (AWS) to transcode for HLS and DASH and be the Content Delivery Network (CDN) that delivers the stream to your web page. We'll be using Video.js for the HTML5 player as well as javascript support libaries to make Video.js work with HLS and DASH.

So Here's what you need:

• An AWS Account
• Video.js
• videojs-contrib-hls.min.js
• dash.all.min.js
• videojs-dash.min.js

Set up three S3 buckets

mingrammer

# -*- coding: utf-8 -*-

import hashlib

import boto3


class ETSManager:
    """
        @todo: manages and provides the ets services

mekza

from boto.cloudfront.distribution import Distribution
from cryptography.hazmat.primitives.asymmetric import padding
from cryptography.hazmat.primitives import serialization
from cryptography.hazmat.backends import default_backend
from cryptography.hazmat.primitives import hashes
import base64

class BetterThanBoto(Distribution):

    def sign_rsa(self, message):

bpholt

#!/bin/bash

cluster=default
container_instance=  # container instance guid

tasks=$(aws --region us-west-2 ecs list-tasks --container-instance $container_instance --cluster $cluster | jq -r '.taskArns | map(.[40:]) | reduce .[] as $item (""; . + $item + " ")')
for task in $tasks; do
	aws --region us-west-2 ecs stop-task --task $task --cluster $cluster
done
aws --region us-west-2 ecs deregister-container-instance --cluster $cluster --container-instance $container_instance

justinvw

# Delete the possibly existing autocomplete test index
curl -X DELETE localhost:9200/autocomplete_test

# Put the config of the autocomplete index
curl -X PUT localhost:9200/autocomplete_test -d '
{
  "settings" : {
    "index" : {
      "analysis" : {
        "analyzer" : {