mtreviso

mtreviso

"""
>>> python remove_unused_bibs.py -h

usage: remove_unused_bibs.py [-h] [-b BIB_FILES] [-p PAPER_FILES] [-o OUT_FILE]

Remove unused bibtex entries

options:
  -h, --help            show this help message and exit
  -b BIB_FILES, --bib_files BIB_FILES

mtreviso

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""psqueue.py
This script displays squeue in a pretty table.
"""
import os
import subprocess
import sys

try:

mtreviso

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""psinfo.py
This script displays sinfo in a pretty table.
"""
import platform
import subprocess
import sys

try:

mtreviso

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""quotacheck.py
This script displays user quotas for a given mount point.
"""

import os
import platform
import subprocess
import sys

mtreviso

import numpy as np


def compute_kendall_taus(h, m):
    """
    Compute multiple variants of Kendall's Tau correlations between two rank arrays.

    This function calculates several variants of Kendall's Tau correlations (tau_a, tau_b, tau_c, 
    tau_10, tau_13, tau_14, tau_23) between two input rank arrays. It employs vectorized operations 
    for computation efficiency and can handle tied ranks. This function is an implementation based on 

mtreviso

"""Adapted from https://github.com/bastings/interpretable_predictions"""
import math

import distrax
import jax
import jax.numpy as jnp


EPS = 1e-6

mtreviso

mtreviso

import torch

def create_bow(words, vocab_size, pad_id=None):
    """
    Create a bag of words matrix using torch.sparse.FloatTensor.
    
    Args:
        words (torch.LongTensor): tensor containing ids for words in
            your vocabulary. Shape of (batch_size, seq_len)
        vocab_size (int): size of the words vocabulary (including special

mtreviso

@staticmethod
def select_word_pieces(features, bounds, method='first'):
    """
    Args:
        features (torch.Tensor): output of BERT. Shape of (bs, ts, h_dim)
        bounds (torch.LongTensor): the indexes where the word pieces start.
            Shape of (bs, ts)
            e.g. Welcome to the jungle -> Wel_ _come _to _the _jungle
                 bounds[0] = [0, 2, 3, 4] 
            indexes for padding positions are expected to be equal to -1