afspies

# EDIT 10/04/2022 - This version was provided by @jayelm who fixed some bugs and made the function much more robust

import os
import subprocess
import time

def assign_free_gpus(threshold_vram_usage=1500, max_gpus=2, wait=False, sleep_time=10):
    """
    Assigns free gpus to the current process via the CUDA_AVAILABLE_DEVICES env variable
    This function should be called after all imports,

kklemon

import torch
import torch.distributed as dist
import torch.multiprocessing as mp

from torch.utils.data import IterableDataset, DataLoader


class DistributedIterableDataset(IterableDataset):
    """
    Example implementation of an IterableDataset that handles both multiprocessing (num_workers > 0) 

zphang

import torch
import torch.nn as nn
import torch.nn.functional as F

from dataclasses import dataclass

import transformers.modeling_bert as modeling_bert
import nlpr.shared.torch_utils as torch_utils
import nlpr.shared.model_resolution as model_resolution

SuperShinyEyes

def f1_loss(y_true:torch.Tensor, y_pred:torch.Tensor, is_training=False) -> torch.Tensor:
    '''Calculate F1 score. Can work with gpu tensors
    
    The original implmentation is written by Michal Haltuf on Kaggle.
    
    Returns
    -------
    torch.Tensor
        `ndim` == 1. 0 <= val <= 1
    

lopuhin

import os
os.environ['OMP_NUM_THREADS'] = '1'
import argparse
import resource
import sys
import time

import numpy as np
import torch
from torchvision.models import resnet34

vi

#!/bin/bash

IN=$1
OUT=$2

true ${SD_PARAMS:="-55dB:d=0.3"};
true ${MIN_FRAGMENT_DURATION:="20"};
export MIN_FRAGMENT_DURATION

if [ -z "$OUT" ]; then

alsrgv

from __future__ import print_function

import argparse
import torch.backends.cudnn as cudnn
import torch.nn.functional as F
import torch.optim as optim
import torch.utils.data.distributed
from torchvision import models
import horovod.torch as hvd
import timeit

thomwolf

model.zero_grad()                                   # Reset gradients tensors
for i, (inputs, labels) in enumerate(training_set):
    predictions = model(inputs)                     # Forward pass
    loss = loss_function(predictions, labels)       # Compute loss function
    loss = loss / accumulation_steps                # Normalize our loss (if averaged)
    loss.backward()                                 # Backward pass
    if (i+1) % accumulation_steps == 0:             # Wait for several backward steps
        optimizer.step()                            # Now we can do an optimizer step
        model.zero_grad()                           # Reset gradients tensors
        if (i+1) % evaluation_steps == 0:           # Evaluate the model when we...

okomarov

# memory footprint support libraries/code
!ln -sf /opt/bin/nvidia-smi /usr/bin/nvidia-smi
!pip install gputil
!pip install psutil
!pip install humanize

import psutil
import humanize
import os
import GPUtil as GPU

danielwatson6

"""Automatically build a vanilla or Cudnn RNN."""

import numpy as np
import tensorflow as tf


def gru(inputs,
        num_layers,
        num_units,
        direction='unidirectional',