You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 7 Next »




Opis

Tensorflow je python knjižnica namijenjena razvoju aplikacija temeljenih na dubokom učenju koja se oslanja na ubrzanje grafičkim procesorima. Jedna od glavnih značajki ove knjižnice je postojanje API-a za brži razvoj modela strojnog učenja Keras, koja u sebi sadrži module i funkcije za svaki dio pipelinea u tipičnoj ML aplikaciji (preprocessing podataka, definicija modela, načina optimizacije i validacije)

Verzije

verzijamodulred
2.10.1scientific/tensorflow/2.10.1-ngcgpu

Dokumentacija

Primjeri

Ispod se nalaze primjeri aplikacija umjetnog benchmarka koji testira performanse na modelu Resnet50.

Primjeri su redom:

  • singlegpu.* - skripte za pokretanje na jednom grafičkom procesoru
  • multigpu-singlenode.* - skripte za pokretanje na više grafičkih procesora na jednom čvoru
  • multigpu-multinode.* - skripte za pokretanje na više grafičkih procesora na jednom čvoru

Jedan grafički procesor

singlegpu.sh 
 #!/bin/bash

#PBS -q gpu
#PBS -l select=1:ncpus=8:ngpus=1:mem=10GB
#PBS -o output/
#PBS -e output/

# pozovi modul
module load scientific/tensorflow/2.10.1-ngc

# pomakni se u direktorij gdje se nalazi skripta
cd ${PBS_O_WORKDIR:-""}

# potjeraj skriptu
run-singlenode.sh singlegpu.py
singlegpu.py 
 #!/usr/bin/env python3

# source:
# - https://github.com/leondgarse/Keras_insightface/discussions/17

import sys
import time
import argparse
import numpy as np
import tensorflow as tf

def main():

    # vars
    batch_size = 256
    samples = 256 * 20
    epochs = 10

    # do not allocate all GPU memory
    gpus = tf.config.experimental.list_physical_devices('GPU')
    for gpu in gpus:
        tf.config.experimental.set_memory_growth(gpu, True)

    # use fp16 for faster inference
    tf.keras.mixed_precision.set_global_policy('mixed_float16')

    # strategy
    gpus = tf.config.experimental.list_physical_devices('GPU')
    devices = [ gpu.name[-5:] for gpu in gpus ]
    strategy = tf.distribute.OneDeviceStrategy(device=devices[0])

    # dataset
    data = np.random.uniform(size=[samples, 224, 224, 3])
    target = np.random.uniform(size=[samples, 1], low=0, high=999).astype("int64")
    dataset = tf.data.Dataset.from_tensor_slices((data, target))
    dataset = dataset.batch(batch_size*strategy.num_replicas_in_sync)

    # define model
    with strategy.scope():
        model = tf.keras.applications.ResNet50(weights=None)
        loss = tf.keras.losses.SparseCategoricalCrossentropy()
        optimizer = tf.optimizers.SGD(0.01)
        model.compile(optimizer=optimizer, loss=loss)

    # fit
    callbacks = []
    model.fit(dataset,
              callbacks=callbacks,
              epochs=epochs,
              verbose=2)

if __name__ == "__main__":
    main()

Više grafičkih procesora na jednom čvoru

multigpu-singlenode.sh 
 #!/bin/bash

#PBS -q gpu
#PBS -l select=1:ncpus=16:ngpus=2:mem=10GB
#PBS -o output/
#PBS -e output/

# pozovi modul
module load scientific/tensorflow/2.10.1-ngc

# pomakni se u direktorij gdje se nalazi skripta
cd ${PBS_O_WORKDIR:-""}

# potjeraj skriptu
run-singlenode.sh multigpu-singlenode.py
multigpu-singlenode.py 
 #!/usr/bin/env python3

# source:
# - https://github.com/leondgarse/Keras_insightface/discussions/17

import sys
import time
import argparse
import numpy as np
import tensorflow as tf

def main():

    # vars
    batch_size = 256
    samples = 256 * 20
    epochs = 10

    # do not allocate all GPU memory
    gpus = tf.config.experimental.list_physical_devices('GPU')
    for gpu in gpus:
        tf.config.experimental.set_memory_growth(gpu, True)

    # use fp16 for faster inference
    tf.keras.mixed_precision.set_global_policy('mixed_float16')

    # strategy
    gpus = tf.config.experimental.list_physical_devices('GPU')
    devices = [ gpu.name[-5:] for gpu in gpus ]
    strategy = tf.distribute.MirroredStrategy(devices=devices)

    # dataset
    data = np.random.uniform(size=[samples, 224, 224, 3])
    target = np.random.uniform(size=[samples, 1], low=0, high=999).astype("int64")
    dataset = tf.data.Dataset.from_tensor_slices((data, target))
    dataset = dataset.batch(batch_size*strategy.num_replicas_in_sync)

    # define model
    with strategy.scope():
        model = tf.keras.applications.ResNet50(weights=None)
        loss = tf.keras.losses.SparseCategoricalCrossentropy()
        optimizer = tf.optimizers.SGD(0.01)
        model.compile(optimizer=optimizer, loss=loss)

    # fit
    callbacks = []
    model.fit(dataset,
              callbacks=callbacks,
              epochs=epochs,
              verbose=2)

if __name__ == "__main__":
    main()

Više grafičkih procesora na više čvorova

Pri definiranju traženih resursa, potrebno je osigurati jednak broj grafičkih procesora po čvoru.

multigpu-multinode.sh 
#!/bin/bash

#PBS -q gpu
#PBS -l select=2:ncpus=8:ngpus=2:mem=10GB
#PBS -l place=free
#PBS -o output/
#PBS -e output/

# pozovi modul
module load scientific/tensorflow/2.10.1-ngc

# pomakni se u direktorij gdje se nalazi skripta
cd ${PBS_O_WORKDIR:-""}

# potjeraj skriptu
run-multinode.sh multigpu-multinode.py
multigpu-multinode.py 
 #!/usr/bin/env python3

# source:
# - https://github.com/leondgarse/Keras_insightface/discussions/17

import os
import sys
import time
import socket
import argparse
import numpy as np
import tensorflow as tf

def main():

    # vars
    batch_size = 256
    samples = 256*20
    epochs = 10

    # do not allocate all GPU memory
    gpus = tf.config.experimental.list_physical_devices('GPU')
    for gpu in gpus:
        tf.config.experimental.set_memory_growth(gpu, True)

    # use fp16 for faster inference
    tf.keras.mixed_precision.set_global_policy('mixed_float16')

    # strategy
    communication_options = tf.distribute.experimental.CommunicationOptions(
        implementation=tf.distribute.experimental.CommunicationImplementation.NCCL)
    strategy = tf.distribute.MultiWorkerMirroredStrategy(
        communication_options=communication_options)

    # dataset
    data = np.random.uniform(size=[samples, 224, 224, 3])
    target = np.random.uniform(size=[samples, 1], low=0, high=999).astype("int64")
    dataset = tf.data.Dataset.from_tensor_slices((data, target))
    dataset = dataset.batch(batch_size*strategy.num_replicas_in_sync)

    # define model
    with strategy.scope():
        model = tf.keras.applications.ResNet50(weights=None)
        loss = tf.keras.losses.SparseCategoricalCrossentropy()
        optimizer = tf.optimizers.SGD(0.01)
        model.compile(optimizer=optimizer, loss=loss)

    # fit
    callbacks = []
    verbose = 2 if os.environ['PMI_RANK'] == '0' else 0
    model.fit(dataset,
              callbacks=callbacks,
              epochs=epochs,
              verbose=verbose)

if __name__ == "__main__":
    main()


Napomene

Apptainer i run-singlenode.sh

Ova knjižnica je dostavljena u obliku kontejnera, zbog opterećenja koje pip/conda virtualna okruženja stvaraju na Lustre dijeljenim datotečnim sustavima.

Za ispravno izvršavanje python aplikacija, potrebno ih je koristiti wrappere run-singlenode.sh ili run-multinode.sh u skriptama sustava PBS:

...
run-singlenode.sh moja_python_skripta.py
...

Korištenje više grafičkih procesora

Tensorflow ne raspodjeljuje automatski aplikaciju na više grafičkih procesora.

Pri razvoju aplikacije, bitno je koristiti odgovarajuće funkcionalnosti poput primjera iznad:

MirroredStrategy primjer
...
strategy = tf.distribute.MirroredStrategy()
with strategy.scope():
	model = ... definirati model ...
	model.compile()
  • No labels