lucataco

Here's the plan to build the local cog binary on this machine:

# Plan

The prerequisites are mostly in place (mise, Go 1.26, Docker). Here are the steps:

# Step 1: Install mise-managed tools

mise install

lucataco

name	replicate
description	Search, explore, and run ML models on Replicate (image gen, video, audio, text, etc.)
homepage	https://replicate.com
metadata	openclaw requires primaryEnv bins env curl jq REPLICATE_API_TOKEN REPLICATE_API_TOKEN

Replicate

Run state-of-the-art open-source and proprietary ML models via the Replicate cloud API.

lucataco

name

openclaw-introspect

description

Explore, understand, and reconfigure your own OpenClaw gateway, agent harness, and system prompt. Use when you need to inspect or change OpenClaw configuration (openclaw.json), understand how the system prompt is built, debug session/channel/model issues, navigate the docs or source code, or tune agent defaults (models, thinking, sandbox, tools, heartbeat, compaction, channels, skills, plugins, cron, hooks). Also use for questions about OpenClaw architecture, the agent loop, context window, or how any OpenClaw feature works internally.

OpenClaw Self-Introspection

Explore and reconfigure your own harness. This skill gives you structured knowledge about the OpenClaw internals so you can inspect, debug, and tune the running gateway.

Quick commands

lucataco

#!/bin/bash
# Update
sudo apt-get update

# Install tools
sudo apt install nvidia-cuda-toolkit

# Install cog
sudo curl -o /usr/local/bin/cog -L "https://github.com/replicate/cog/releases/latest/download/cog_$(uname -s)_$(uname -m)"
sudo chmod +x /usr/local/bin/cog

lucataco

palette = 0=#23272e
palette = 1=#f38020
palette = 2=#a8e6a3
palette = 3=#faae40
palette = 4=#4da6ff
palette = 5=#ff80ab
palette = 6=#66d9ef
palette = 7=#c0c5ce
palette = 8=#4f5b66
palette = 9=#f38020

lucataco

# Replace /etc/docker/daemon.json docker config in Brev.dev Crusoe GPUs
{
    "default-runtime": "nvidia",
    "mtu": 1500,
    "runtimes": {
        "nvidia": {
            "args": [],
            "path": "nvidia-container-runtime"
        }
    },

lucataco

#!/bin/bash

# This is local cli command that allows users to use kokoro on a Macbook Pro
# Requires you to first run the kokoro docker container:
docker run -p 8880:8880 ghcr.io/remsky/kokoro-fastapi-cpu:latest
# Then save this file to /usr/local/bin
# Finally you can test:
kokoro "The quick brown fox jumped over the lazy dog"
# Or even pipe from a stream like:
llm "tell me a joke" | kokoro

lucataco

# Setup:
# conda create -n wan python=3.10
# conda activate wan
# pip3 install torch torchvision torchaudio
# pip install git+https://github.com/huggingface/diffusers.git@3ee899fa0c0a443db371848a87582b2e2295852d 
# pip install accelerate==1.4.0
# pip install transformers==4.49.0
# pip install ftfy==6.3.1

lucataco

services:
  pihole-unbound:
    image: 'bigbeartechworld/big-bear-pihole-unbound:2024.07.0'
    environment:
      - SERVICE_FQDN_PIHOLE_8080
      - SERVICE_FQDN_PIHOLE_10443
      - 'DNS1=127.0.0.1#5353'
      - DNS2=no
      - TZ=America/Chicago
      - WEBPASSWORD=$SERVICE_PASSWORD_PIHOLE

lucataco

from optimum.quanto import freeze, qfloat8, quantize
from diffusers import FluxPipeline
import torch
import time

seed=1337
generator = torch.Generator("cuda").manual_seed(seed)

pipeline = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", torch_dtype=torch.bfloat16).to("cuda")