Nur Lan 0xnurl

Installing OpenFst Native Python Extension on MacOS

Starting from version 1.5, OpenFst has offered a native Python module, making the use of external wrappers like PyFst unnecessary. This has been greatly helpful since PyFst doesn't support Python 3.

1. Install OpenFst

Download OpenFst 1.5.4 or above from http://openfst.cs.nyu.edu/twiki/bin/view/FST/FstDownload
Unzip, untar: $ tar xzvf openfst-1.5.4.tar.gz
$ ./configure --enable-python --enable-far

A simple MPI logging and compute example

The simple logging example can be run via:

mpiexec -n 4 python mpi_logger.py

For the compute and logging example:

mpiexec -n 4 python mock_execution.py

Some remarks on Large Language Models

Yoav Goldberg, January 2023

Audience: I assume you heard of chatGPT, maybe played with it a little, and was imressed by it (or tried very hard not to be). And that you also heard that it is "a large language model". And maybe that it "solved natural language understanding". Here is a short personal perspective of my thoughts of this (and similar) models, and where we stand with respect to language understanding.

Intro

Around 2014-2017, right within the rise of neural-network based methods for NLP, I was giving a semi-academic-semi-popsci lecture, revolving around the story that achieving perfect language modeling is equivalent to being as intelligent as a human. Somewhere around the same time I was also asked in an academic panel "what would you do if you were given infinite compute and no need to worry about labour costs" to which I cockily responded "I would train a really huge language model, just to show that it doesn't solve everything!". We


	mkdir -p /work/

	# Clone the source from Github
	cd /work/ && git clone — recurse-submodules https://github.com/tensorflow/serving

	# Pin the version of Tensorflow Serving and its submodule
	TENSOR_SERVING_COMMIT_HASH=85db9d3
	TENSORFLOW_COMMIT_HASH=dbe5e17
	cd /work/serving && git checkout $TENSOR_SERVING_COMMIT_HASH

	# Clone the Tensorflow Serving source
	git clone https://github.com/tensorflow/serving
	cd serving && git checkout <commit_hash>

	# Build the docker image (time to go get yourself a coffee, maybe a meal as well, this will take a while.)
	docker build -t some_user_namespace/tensorflow-serving:latest -f ./serving/tensorflow_serving/tools/docker/Dockerfile.devel .

	# Run up the Docker container in terminal
	docker run -ti some_user_namespace/tensorflow-serving:latest

	#coding: utf-8
	#demo of beam search for seq2seq model
	import numpy as np
	import random

	vocab = {
	0: 'a',
	1: 'b',
	2: 'c',
	3: 'd',

	#!/usr/bin/env python3

	import os
	import paramiko

	ssh_key_filename = os.getenv('HOME') + '/.ssh/id_rsa'

	jumpbox_public_addr = '168.128.52.199'
	jumpbox_private_addr = '10.0.5.10'
	target_addr = '10.0.5.20'