aurotripathy’s gists

aurotripathy / run-onnx-model.py

Last active March 13, 2026 18:55

	mport onnxruntime as ort
	import numpy as np
	import torch


	print(f"onnx runtime version:{ort.__version__}")

	# Check if PyTorch can see an available GPU \

	if torch.cuda.is_available():

aurotripathy / onnx-2-safetensors.py

Last active March 12, 2026 19:50

	import argparse
	import os

	import onnx
	import onnx_safetensors


	INPUT_DIR = "onnx-files" # onnx files have to go here
	OUTPUT_DIR = "safetensors-files" # out put files will generated here

aurotripathy / resnet-on-onnxruntime.py

Last active March 13, 2026 01:17

resnet on onnx runtime

	import onnxruntime # to inference ONNX models, we use the ONNX Runtime
	import onnx
	from onnx import numpy_helper
	import urllib.request
	import json
	import time
	import torch
	import numpy as np

aurotripathy / sample-pt-load-convert-to-onnx-infer.py

Created March 10, 2026 19:06

	import onnxruntime
	import numpy as np

	from ultralytics import YOLO
	# Load a pretrained YOLOv8 model
	model = YOLO('yolov8n.pt')
	# Export to ONNX format
	model.export(format='onnx')

aurotripathy / onnx-model-metadata.py

Created March 10, 2026 04:09

	import onnx
	import onnx_tool

	# Load the ONNX model from a file
	# model_path = "resnet34_1_3_416_640.onnx"
	# model_path = "pointpillar_custom.onnx"
	model_path = "detr_1_3_512_512.onnx"
	print(f"Model: {model_path}")

	# Use onnx.load to get the model proto object

aurotripathy / quadlet_script_run.sh

Created January 29, 2026 14:12

perf script

	#!/bin/bash
	if [ "$#" -ne 2 ]; then
	echo "Usage: $0 <results_suffix> <port_num>"
	exit 1
	fi


	results_suffix=$1
	port_num=$2

aurotripathy / flowsettings.py

Last active June 4, 2025 21:15

Kotaemon flowsettings.py file for configuring the project

	import os
	from importlib.metadata import version
	from inspect import currentframe, getframeinfo
	from pathlib import Path

	from decouple import config
	from ktem.utils.lang import SUPPORTED_LANGUAGE_MAP
	from theflow.settings.default import * # noqa

	cur_frame = currentframe()

aurotripathy / get-context-window.py

Last active June 2, 2025 17:16

 # from https://community.openai.com/t/easy-way-to-get-a-context-window-for-a-model/552099/4
 # the hard way...
 # Makes a gigantic meaningless OpenAI chat-completion promp call into the vLLM server
 # Parses the (error) return and determines the actual context window supported
 # Usage: python context-window-discovery.py --model gpt-4o-mini --base-url http://localhost:8080/v1
 from openai import OpenAI
 import re
 import argparse

aurotripathy / RNGD-sample-RAG-App.py

Last active May 7, 2025 15:57

A canonical RAG sample application. Uses (furiosa-llm server + OpenAI embeddings + Chroma Vector DB + LangChain framework)

	# A canonical RAG sample application
	# Uses (furiosa-llm server + OpenAI embeddings + Chroma Vector DB + LangChain framework)
	# 100% cursor-generated code
	# Needs a text file in the documents directory,
	# I used the doc, https://gist.github.com/wey-gu/75d49362d011a0f0354d39e396404ba2

	from typing import List, Dict, Optional
	from langchain.text_splitter import RecursiveCharacterTextSplitter
	from langchain.embeddings import OpenAIEmbeddings
	from langchain.vectorstores import Chroma

aurotripathy / calc-perplexity-from-logprobs.py

Last active February 11, 2025 22:42

	from openai import OpenAI
	from dotenv import load_dotenv
	import numpy as np
	load_dotenv()

	def chat_with_gpt(messages):
	client = OpenAI()
	try:
	completion = client.chat.completions.create(
	model="gpt-3.5-turbo",

Auro Tripathy aurotripathy