dmarx

import numpy as np
from openai import OpenAI
import plotly
import plotly.graph_objs as go
import umap


url = "http://localhost:80"

client = OpenAI(

dmarx

https://services.math.duke.edu/~rtd/EOSP/EOSP2021.pdf - stochastic processes
https://arxiv.org/abs/2106.10165 - principles of deep learning theory
https://arxiv.org/abs/2104.13478 - geometric deep learning
https://hastie.su.domains/ElemStatLearn/printings/ESLII_print12_toc.pdf
https://ia601508.us.archive.org/35/items/collection-of-mathematics-probability-and-statistics-books/A%20First%20Course%20in%20Probability%2C%20Global%20Edition%20%28Sheldon%20Ross%29.pdf
https://www.cs.uoi.gr/~arly/courses/ml/tmp/Bishop_book.pdf - PRML, Chris Bishop
https://github.com/probml/pml2-book/releases/latest/download/book2.pdf
https://maurice-weiler.gitlab.io/cnn_book/EquivariantAndCoordinateIndependentCNNs.pdf
https://arxiv.org/abs/1504.03001 - Chaos on the interval (225pp)
https://www.ma.imperial.ac.uk/~dturaev/kuznetsov.pdf - Elements of Applied Bifurcation Theory (614pp)

dmarx

# NB: ar5iv endpoint is lagged relative main arxiv
import subprocess


def convert_arxiv_url_to_markdown(arxiv_url, target_format='markdown_github'):
    parts = arxiv_url.split('/')
    article_id = parts[-1] if parts[-1] else parts[-2]  # Account for trailing slash
    article_id = article_id.replace('.pdf', '')
    html_url = f'https://ar5iv.labs.arxiv.org/html/{article_id}' 
    

dmarx

simulate the behavior of the following function. respond only with the output, enclosed in brackets.

def pair(node1, node2):
    GeeseGraph = WorldGraph.neighborhood("things related to geese")
    geese_nodes1 = GeeseGraph.nearest_neighbors(node1, n=10)
    geese_nodes2 = GeeseGraph.nearest_neighbors(node2, n=10)
    return geese_nodes1.intersect(geese_nodes2).sample(n=1, bias="funny matches")

>>> GeeseGraph.sample(n=100)
[pond, goose, feather, honk, migration, flock, nest, lake, wintering grounds, birdwatching, grain, park, bread, gosling, water, flyway, wetlands, grass, reeds, sunrise, sunset, quack, wildlife, bird, aviary, down, flapping, waddle, pecking, squawk, preen, aquatic plants, river, stream, marsh, ecosystem, habitat, conservation, ornithology, species, biodiversity, nature reserve, environmental protection, tagging, tracking, molting, territorial, aggression, mating rituals, offspring, survival, adaptation, plumage, waterfowl, aquatic invertebrates, feeding, foraging, resting, roosting, territo

dmarx

import numpy as np
import scipy as sp
from scipy.stats import poisson, expon
import matplotlib.pyplot as plt
import math
import random

random.seed(42)

dmarx

eff_batch_size=2048

#########################
# what you're doing-ish #
#########################

gradient_accumulation_steps = 4

micro_batch_size = eff_batch_size // gradient_accumulation_steps

dmarx

def init_comfy(
    comfyui_path = "/home/dmarx/projects/ComfyUI"
):
    # 1. add ComyUI to path
    import sys
    sys.path.append(comfyui_path)

    # 2. ensure cli parser doesn't cause issues
    from comfy.options import enable_args_parsing
    enable_args_parsing(False)

dmarx

#!pip install omegaconf keyframed

from omegaconf import OmegaConf
from collections import defaultdict

from keyframed import Curve
from keyframed.dsl import curve_from_cn_string
from keyframed.utils import simplify

def update_storyboard_inferred_values(storyboard):

dmarx

from pathlib import Path
import re

images = list(Path('./frames').glob('*.png'))
n_frames = len(images)

def get_frame_index(fpath):
    fpath = str(fpath)
    match = re.findall('[0-9]+', fpath)
    if match:

dmarx

scene1 {transition: t0} | scene2 {transition: t1} | scene3 {transition: t0} | scene4

scenes:
  scene1:
    prompts:
    - foo
    - bar
  scene2:
    prompts:
    - baz