Good question! I am collecting human data on how quantization affects outputs. See here for more information: ggml-org/llama.cpp#5962
In the meantime, use the largest that fully fits in your GPU. If you can comfortably fit Q4_K_S, try using a model with more parameters.
See the wiki upstream: https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix
| # Credit 🙏: I just used the example from langchain docs and it works quite well: https://python.langchain.com/en/latest/use_cases/question_answering.html | |
| # Note 2: The Arxiv -> PDF logic is a bit messy, I'm sure it can be done better | |
| # Note 3: Please install the following: | |
| # To run: | |
| # Save this in a `app.py` | |
| # pip install arxiv PyPDF2 langchain chromadb | |
| # The chat feature was shipped in H2O nightly this week, we will need to install from nightly link: |
| // Turn off "Sends data to servers when leaving pages" | |
| user_pref("beacon.enabled", false); | |
| // Prevention of some telemetry related to the newtab | |
| user_pref("browser.newtabpage.directory.ping", ""); | |
| user_pref("browser.newtabpage.directory.source", ""); | |
| user_pref("browser.newtabpage.enhanced", false); | |
| // "In the release channels the Mozilla location service is used to help in figuring out regional search defaults." | |
| // Which means sending collectable data |
| #!/usr/bin/env bash | |
| if [ "$(id -u)" != "0" ]; then | |
| echo "This script must be run as root" 1>&2 | |
| exit 1 | |
| fi | |
| KMAJ="$(uname -r | cut -d. -f1)" | |
| KMIN="$(uname -r | cut -d. -f2)" | |
| if [ "${KMAJ}" != "4" ]; then | |
| echo "This script has to be run on a 4.x kernel." 1>&2 |
Unless you are using Safari on OSX, most browsers will have some kind of free plugin that you can use to export the browser's history. So that's probably the easiest way. The harder way, which seems to be what Safari wants is a bit more hacky but it will also work for other browsers. Turns out that most of them, including Safari, have their history saved in some kind of sqlite database file somewhere in your home directory.
The OSX Finder cheats a little bit and doesn't show us all the files that actually exist on our drive. It tries to protect us from ourselves by hiding some system and application-specific files. You can work around this by either using the terminal (my preferred method) or by using the Cmd+Shft+G in Finder.
Once you locate the file containing the browser's history, copy it to make a backup just in case we screw up.
| # Hello, and welcome to makefile basics. | |
| # | |
| # You will learn why `make` is so great, and why, despite its "weird" syntax, | |
| # it is actually a highly expressive, efficient, and powerful way to build | |
| # programs. | |
| # | |
| # Once you're done here, go to | |
| # http://www.gnu.org/software/make/manual/make.html | |
| # to learn SOOOO much more. |
This is a follow-up to this discussion: Can NAT traversal be Tor's killer feature?
If torrents are P2P's killer application, and NAT traversal/"static IP" are Tor's (via hidden services), putting them together could prove to be the best incentivization scheme for growing the Tor network other than cold crypto cash.
Everybody knows you're not supposed to use torrents with tor, right?
(by @andrestaltz)
If you prefer to watch video tutorials with live-coding, then check out this series I recorded with the same contents as in this article: Egghead.io - Introduction to Reactive Programming.
| # -*- coding: utf-8 -*- | |
| """ | |
| Created on Thu Oct 3 12:24:41 2013 | |
| @author: cheesinglee | |
| """ | |
| import praw | |
| from csv import DictWriter |
(Note: This was originally from my blog post on AlexanderRobotics.com, but I brought that site down and wanted to keep this post alive since it's referenced by Stack Overflow).
The beauty of the Arduino lies in its simplicity. A look under the covers of the Arduino and its IDE, however, reveals a fascinating chipset and an open, extensive toolchain.
