FastaiMonitor

Simple two Monitors for fast.ai, using TensorBoard or Neptune.ml

Colab Example: https://colab.research.google.com/drive/1HEVbnOyKSLSIkMhpYLWO83Wut3hJrP42

How to use with TensorBoard

List of saved parameters:

• Scalars with all metrics and training loss for each batch
• Images for top losses images
• Graph for the model
• Distribution and Histograms for each parameters in layers
• Hyperparameter Text for learning rate (lr), momentum (mom), weight decay (wd), and beta (beta)
• Text for Model, Summary, Model save path, and Total Epochs

Google Colab Install

1. Clone project

git clone https://github.com/alexiej/FastaiMonitor.git

2. Install all requirements, and run ngrok.io to use tensorboard

!pip install tensorboardX
!wget https://bin.equinox.io/c/4VmDzA7iaHb/ngrok-stable-linux-amd64.zip
!unzip ngrok-stable-linux-amd64.zip

LOG_DIR = '/content/log'
get_ipython().system_raw(
    'tensorboard --logdir {} --host 0.0.0.0 --port 6006 &'
    .format(LOG_DIR)
)

get_ipython().system_raw('/content/ngrok http 6006 &')
! curl -s http://localhost:4040/api/tunnels | python3 -c \
    "import sys, json; print(json.load(sys.stdin)['tunnels'][0]['public_url'])"
#End Google Colabc

Local install

1. install tensorboardX, and clone repository

pip install tensorboardX
git clone https://github.com/alexiej/FastaiMonitor.git

2. run tensorboard

tensorboard --logdir ./log --host 0.0.0.0 --port 6006 &

Run Code

1. Add monitor for your learning fast.ai model

from fastai import *
from fastai.vision import *
from FastaiMonitor.TensorBoardMonitor import TensorBoardMonitor

path = untar_data(URLs.MNIST_TINY);
bs = 64
data = ImageDataBunch.from_folder(path,
                                   ds_tfms=get_transforms(), 
                                   size=224, 
                                   bs=bs
                                  ).normalize(imagenet_stats)
learn = cnn_learner(data, models.resnet34, metrics=error_rate)
TensorBoardMonitor(learn)

learn.fit_one_cycle(2, max_lr=1e-2)

4. Go to the tensorboard http://localhost:6006/ (or to link for ngrok.io )

How to use with Neptune.ml

List of exported parameters:

• all metrics
• last_loss batch
• model.txt - info about the model
• opt.txt - OptiWrapper
• Properties for learning rate (lr), momentum (mom), weight decay (wd), and beta (beta)
• Use of CPU/GPU, RAM

1. Create your account on the Neptune.ml (https://neptune.ml/), and get your API Token, and project name

2. Clone git project

git clone https://github.com/alexiej/FastaiMonitor.git

2. Install neptune-client

pip install neptune-client

3. Use NeptuneMonitor with your fast.ai code

API_TOKEN = '---'
PROJECT_NAME = '---'

from fastai import *
from fastai.vision import *
from FastaiMonitor.NeptuneMonitor import NeptuneMonitor

path = untar_data(URLs.MNIST_TINY);
bs = 64
data = ImageDataBunch.from_folder(path,
                                   ds_tfms=get_transforms(), 
                                   size=224, 
                                   bs=bs
                                  ).normalize(imagenet_stats)
learn = cnn_learner(data, models.resnet34, metrics=error_rate)

nm = NeptuneMonitor(learn,
                    api_token=API_TOKEN,
                    project=PROJECT_NAME)
nm.start(name='first-experiment')
learn.fit_one_cycle(2, max_lr=1e-2)

nm.stop()

4. Go to the list of the experiments and watch your experiments.