GiliardGodoi

DvC - Data Version Control

Controla a versão de um dataset com base no hash do arquivo. Digamos que um projeto é desenvolvido no diretório experimento. Ao iniciar o dvc na pasta com dvc init, é criado um arquivo que gerencia o projeto.

Documentação

• Get Started: Data Versioning | Data Version Control · DVC
• Get Started: Experimenting Using Pipelines | Data Version Control · DVC

GiliardGodoi

def measure_certainty(df, bins=[0.75, 0.85, 0.95], diff_max=0.10):
  
  bins=[0.75, 0.85, 0.95] # bins deve corresponder com o array de categorias
  categories = np.array(['tema não identificado', 'dúvida', 'semelhante ao tema', 'certeza'])
  columns = [c for c in df.columns if c.startswith("Proba")]
  values = -(df[columns].values) # valores das probabilidades
  idxs = np.argsort(values, axis=1) # indexes que ordena os vetores (linhas) das probabilidades
  ordered = -(np.take_along_axis(values, idxs, axis=1)) # pode ser substituido simplesmente por np.sort
  binned = np.digitize(ordered[:, 0], bins)
  

GiliardGodoi

import random
from evol import Population, Evolution

CHROMO_LENGTH = 30
CHROMO_PARTITION = 10
TARGET_VALUE = 52

POPULATION_SIZE = 100

def random_chromosome(size):

GiliardGodoi

# The commands below are a guide to remove a large file that has been
# accidentally committed to a Git repository's history. If the file is
# larger than 100 MB, GitHub will prevent you from pushing your latest
# commits. The annotated steps below should help you remove the large
# file from your commit history, even if you've made new commit since.


# Some Git users advise against rebasing. You can safely use it here
# because you haven't published your changes yet.
# So, you first need to rebase your current branch onto the point that

GiliardGodoi

GIT

Estados

• Modificado (modified) ou área de trabalho (working directory)
• Preparado (staged/index)
• Consolidado (comitted)

Ajuda

GiliardGodoi

height: 500
width: 700

GiliardGodoi

license : mit
height : 550

GiliardGodoi

license : mit
height : 960