JohanAR

// Fast coefficient reference for rowmajor array with fixed number of columns
template <class Scalar_, int Rows_, int Cols_, int Flags_, int MaxRows_, int MaxCols_,
          std::enable_if_t<(Flags_ & Eigen::RowMajor) && (Cols_ != Eigen::Dynamic), bool> = true>
auto CO(Eigen::Array<Scalar_, Rows_, Cols_, Flags_, MaxRows_, MaxCols_> &arr,
               Eigen::Index row, Eigen::Index col) -> Scalar_ & {
  return *(arr.data() + Cols_ * row + col);
}

// Fast coefficient const reference for rowmajor array with fixed number of columns
template <class Scalar_, int Rows_, int Cols_, int Flags_, int MaxRows_, int MaxCols_,

JohanAR

FROM nvidia/cuda:11.8.0-devel-ubuntu22.04

# https://developer.nvidia.com/cuda-gpus
# for a rtx 2060: ARG TORCH_CUDA_ARCH_LIST="7.5"
ARG TORCH_CUDA_ARCH_LIST="${TORCH_CUDA_ARCH_LIST:-3.5;5.0;6.0;6.1;7.0;7.5;8.0;8.6+PTX}"

LABEL maintainer="Your Name <your.email@example.com>"
LABEL description="Docker image for GPTQ-for-LLaMa and Text Generation WebUI"

RUN --mount=type=cache,target=/var/cache/apt,sharing=locked,rw apt-get update && \

JohanAR

Usage

I've tried to make the script as robust as possible, but use at your own risk, no warranties given, etc.

1. Set the environment variable LLAMA_CPP_DIR to wherever you've checked out https://github.com/ggerganov/llama.cpp
2. Make sure it contains the convert-llama-ggml-to-gguf.py script, and that you've installed all necessary requirements to run it.
3. Locate the original model repo on huggingface.co and copy its URL, i.e. if you downloaded a model quantized by TheBloke, check the model card for the link to the original model (unless it's one of TheBloke's own models.. The important thing is that the repo URL you use contains a config.json with more than one line of text).
4. Run the script with the GGML files you want to convert as first argument, and the URL as second. If everything is set up right it will create a GGUF file next to the input file.

Example

JohanAR

#!/bin/sh

# Set persistence mode on
nvidia-smi -i 0 -pm 1

# Set 200W power limit
nvidia-smi -i 0 -pl 200

# Limit max clock speed to 1850 MHz
nvidia-smi -i 0 -lgc 210,1850

JohanAR

#!/usr/bin/env bash

VLC='/usr/bin/vlc --sout-all --sout #display'

joinargs () {
  local IFS="#"
  echo "$*"
}

multivlc () {

JohanAR

# Put the exported csv from filmtipset.se next to this script and rename it to film.csv

import re

IMPORT=re.compile(r"(\d{4}-\d{2}-\d{2}),(.+);(\d+);(\d)")

def to_imdb(num):
    return 'tt{:07d}'.format(int(num))

def parse(ft):

JohanAR

from collections import namedtuple
import os
import re

FileInfo = namedtuple('FileInfo', ['toname', 'subname'])

def doit():
    directory = os.fsencode('.')

    matcher = re.compile('.*[sS](\d+)[eE]([0-9\-]+).*')

JohanAR

#ifndef _POLYNOMIAL_REGRESSION_H
#define _POLYNOMIAL_REGRESSION_H  __POLYNOMIAL_REGRESSION_H
/**
 * PURPOSE:
 *
 *  Polynomial Regression aims to fit a non-linear relationship to a set of
 *  points. It approximates this by solving a series of linear equations using 
 *  a least-squares approach.
 *
 *  We can model the expected value y as an nth degree polynomial, yielding

JohanAR

SteamVR System Report created tor jul 27 15:57:31 2017

<Report>
SteamVR Version:	1501109730
SteamVR Date:	2017-07-27
Steam:	Public
Steam Branch:	beta
Steam AppID:	250820
Tracking:	lighthouse
OS:	Linux version 4.10.0-28-generic (buildd@lgw01-11) (gcc version 6.3.0 20170406 (Ubuntu 6.3.0-12ubuntu2) ) #32-Ubuntu SMP Fri Jun 30 05:32:18 UTC 2017