mattharrison

What Works

Install Docker image https://github.com/eugr/spark-vllm-docker/

git clone https://github.com/eugr/spark-vllm-docker.git
cd spark-vllm-docker
./build-and-copy.sh --use-wheels

mattharrison

mattharrison

import pandas as pd

url = 'http://atmenv.envi.osakafu-u.ac.jp/omu-content/uploads/sites/1215/2015/10/KyotoFullFlower7.xls'
cherry_raw = pd.read_excel(url, skiprows=25, dtype_backend='pyarrow')

def tweak_cherry(df):
    return (df
            #.query('~`Full-flowering date (DOY)`.isna()')
            .rename(columns={'AD': 'year',
                             'Full-flowering date (DOY)': 'flowering_doy',

mattharrison

from sklearn.metrics import roc_auc_score
from hyperopt import hp, Trials, fmin, tpe

from hyperopt import fmin, tpe, hp, STATUS_OK, Trials
from sklearn.metrics import accuracy_score, roc_auc_score  

from typing import Any, Dict, Union
import xgboost as xgb

def hyperparameter_tuning(space: Dict[str, Union[float, int]], 

mattharrison

Hardware

• Fans!
  • M2 is quiet
  • P1 is noisy (used TPFanControl to help)

Software

• Emacs keybindings everywhere
  • P1 - AutoHotKeys with emacseverywhere
  • M2 - Hammerspoon with https://github.com/justintanner/universal-emacs-keybindings
• Zoom

mattharrison

mattharrison

mattharrison

mattharrison

mattharrison

%matplotliblib inline
import pandas as pd
url = 'https://github.com/COVID19Tracking/covid-tracking-data/raw/master/data/states_daily_4pm_et.csv'
df = pd.read_csv(url, parse_dates=['date', 'dateChecked'])

(df
 [df.state == 'NC']
 .set_index('date')
 [['positive', 'hospitalized']]
 .fillna(0)