Thomas Capelle tcapelle

Linux

Setup

When setting up a fresh VM.

sudo apt update
sudo apt-get install vim tmux wget

My Personal Top Wines: by Region

This is an arbitrary good list of wines I like, don't be offended if your favorite wine is not there. I have limited tasting time, and have mostly choose little family owned producers

Everything here is Organic, most are biodynamic and some are natural.

I have visited most of the winerys myself, and now I order directly from them or by my local wine-shop that actually helps me find more wine of this quality.

Loire

Probably my fav white wine region, I love the floral and sometimes minerality of these wines. There is always a good reason to open any of these bottles. It is also a very beautiful region to visit where you find accomodation on old castles like Chteau de Beaujeu.

	from mistralai import Mistral

	import os
	import weave

	weave.init("pixtral")

	print("1. La Platforme Mistral")
	api_key = os.environ["MISTRAL_API_KEY"]
	model = "pixtral-12b-2409"

	from dataclasses import dataclass
	from pathlib import Path
	import simple_parsing


	# clone weave:
	# git clone https://github.com/wandb/weave
	WEAVE_DOCS_PATH = Path("weave/docs/docs")

	@dataclass

	name: Build and Test Locally
	on:
	workflow_dispatch: #allows repo admins to trigger this workflow from the Actions tab
	inputs:
	create_github_issue:
	type: boolean
	description: 'Create issues on failing notebooks'
	default: false
	nbs_list:
	type: choice

	import wandb
	from wandb import Api

	api = Api()
	ENTITY = "fastai"
	PROJECT = "fine_tune_timm"

	project = api.project(PROJECT, entity=ENTITY)

	sweeps = project.sweeps()

	import os
	import openai
	from rich.console import Console

	console = Console()

	openai.api_key = os.getenv("OPENAI_API_KEY")

	history = [{"role": "system", "content": "You are a helpful assistant."},]

	import wandb
	import timm
	import argparse
	from fastai.vision.all import *
	from fastai.callback.wandb import WandbCallback
	from torchvision import models

	def parse_args():
	parser = argparse.ArgumentParser()
	parser.add_argument('--batch_size', type=int, default=64)