smellslikeml smellslikeml

Findings from running remyxai-cli explore across 6 production repos

Recent AI research lands in existing codebases through specific extension points — modules, callbacks, or data-structure fields where a new method can plug in. Which extension points a repo provides determines which methods can be tried against it without a rewrite. We ran an agentic method-search loop that dispatches recent arxiv papers as draft integrations against 6 production repos; the by-product across 36 cycles was a per-repo inventory of the specific extension points those repos are missing.

The dispatch mode

Packaged as a CLI subcommand in remyxai-cli #46:

remyxai outrider explore --repo owner/name \

GLM Tries, Opus Triages: Behavioral Differences in Research-to-Code Agents

A controlled comparison across 19 paired runs spanning 19 repository forks — 38 individual workflow executions total — running an identical paper-implementation pipeline (remyxai/outrider — Claude Code under the hood, with glm-5.2 routed at z.ai's Coding Plan endpoint vs default Opus). The pipeline ran in two modes that probe different parts of the workflow:

Selection-pass mode (n=9): no pin; each provider freely selects its own paper from the candidate pool. Exercises the full pipeline including selection + verification gates.
Pin-method mode (n=10): same paper pinned on each fork, both providers run their full chain on identical input. Isolates implementation-side behavior on a forced pick.

The aggregate verdict comes from the n=19 union; the two mode-specific breakdowns below show where the difference comes from.

Reproducing this

Running GitHub Copilot VSCode extension against local Code Llama model

Tested on NVIDIA RTX 4090, but these instructions also cover AMD and Mac in case you wanna try those.
This guide assumes you are running Linux (I ran this on Ubuntu).

Before you get excited:

Setting up a MSI laptop with GPU (gtx1060)

Installing Ubuntu 18.04, CUDA, CDNN, Pytorch and TensorFlow

Installing Ubuntu 18.04

Get Image

https://tutorials.ubuntu.com/tutorial/tutorial-create-a-usb-stick-on-ubuntu#0

Install (solving issues)

(Unofficial) Discord server rules suggestions list

Author's Note

I'll start off with letting you know this is a fork from someone else. However, for some bizarre reason, this is the one everyone finds, so I better get round to updating this. Credit to Cristiano#2233 for the original idea.

Also, I've had a lot of people saying the rules are to strict. If you pick all the rules here, you're right, it would be very strict. However the rules below are guidelines! They are there for you to pick the ones you desire, you can ignore ones you don't want. Hopefully they might help with rules you wouldn't have thought of otherwise.

This guide is kept up-to-date as Discord and available resources change!
A basic server template is available here

Creating and Growing a Discord Server

Introduction

Hello! I'm jagrosh#4824! I'm writing this guide to try to help new server owners set up and grow their servers, which is a commonly-requested topic. It's very easy to go about this the wrong way, so it's best to be prepared and make smart decisions so that your community can flourish!

	#!/bin/bash

	### steps ####
	# verify the system has a cuda-capable gpu
	# download and install the nvidia cuda toolkit and cudnn
	# setup environmental variables
	# verify the installation
	###

	### to verify your gpu is cuda enable check

	import logging
	import os
	import sys
	import traceback
	from contextlib import contextmanager

	import diart.operators as dops
	import numpy as np
	import rich
	import rx.operators as ops

	# This file contains a collection of workarounds for missing TFLite support from:
	# https://github.com/tensorflow/magenta/tree/master/magenta/music
	# as posted in https://github.com/tensorflow/tensorflow/issues/27303
	# Thanks a lot to github.com/rryan for his support!

	# The function for testing MFCC computation given PCM input is:
	# - test_mfcc_tflite
	# Please not that the output has not yet been compared to the one produced by the respective TF functions.

	# This file also contains test code for other problems in the context of audio processing with TF and TFLite:

	#!/usr/bin/python
	import cv2
	import numpy as np
	from primesense import openni2
	from primesense import _openni2 as c_api
	openni2.initialize("<PATH TO OPENNI2 REDIST FOLDER>")
	dev = openni2.Device.open_any()
	depth_stream = dev.create_depth_stream()
	depth_stream.start()
	depth_stream.set_video_mode(c_api.OniVideoMode(pixelFormat = c_api.OniPixelFormat.ONI_PIXEL_FORMAT_DEPTH_100_UM, resolutionX = 640, resolutionY = 480, fps = 30))