Fabrizio Milo Mistobaan

OpenAI golf parameter

This is a vibe coded script by codex to check the average document length after tokenization on the pre-tokenized dataset of parameter golf.

Full train-split summary from the current cached dataset

documents: 6,292,940
tokens_accounted_for: 8,000,000,000

uv tool run opik harbor trials start -p ./examples/tasks/k8s-hello-world/ \
        --agent "terminus-2" \
        --agent-kwarg "api_base=$OPENAI_API_BASE" \
  --model "hosted_vllm/$MODEL" \
  --environment-type "docker"

OPIK: Opik tracking enabled for Harbor
Starting trial: k8s-hello-world__wv9bPao
Task: k8s-hello-world

	# # Serverless OpenThinker Agent v1 with vLLM and Modal

	# This example shows how to serve Hugging Face's
	# [open-thoughts/OpenThinker-Agent-v1](https://huggingface.co/open-thoughts/OpenThinker-Agent-v1)
	# with vLLM on Modal.
	# OpenThinker-Agent-v1 is an agentic model post-trained from Qwen/Qwen3-8B via supervised fine-tuning and reinforcement learning,
	# and it excels on agent benchmarks such as Terminal-Bench 2.0 and SWE-Bench.

	# We include the same Modal best practices as our other inference demos:
	# - GPU snapshots for faster cold starts

	#!/bin/bash

	# Check if the argument is provided
	if [ -z "$1" ]; then
	echo "Usage: $0 <path_storage>"
	exit 1
	fi


	PATH_STORAGE=$1

	<!DOCTYPE html>
	<html xmlns="http://www.w3.org/1999/xhtml" lang xml:lang>

	<head>
	<meta charset="utf-8" />
	<meta name="generator" content="nbh" />
	<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes" />
	<title>Notebook HTML Renderer</title>
	<!-- KaTeX -->
	<script src="https://cdnjs.cloudflare.com/ajax/libs/KaTeX/0.16.9/katex.min.js"

	# install node
	# Download and install nvm:
	curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.3/install.sh \| bash

	# in lieu of restarting the shell
	\. "$HOME/.nvm/nvm.sh"

	# Download and install Node.js:
	nvm install 22

	import math

	def compute_cost(num_nodes, gpus_per_node, cost_per_gpu_hour, days):
	"""
	Computes the cost of running a certain setup (num_nodes, gpus_per_node)
	for a specified number of days at a particular cost_per_gpu_hour.

	:param num_nodes: Number of nodes
	:param gpus_per_node: Number of GPUs in each node
	:param cost_per_gpu_hour: Cost per GPU-hour in USD