import flor
log = flor.log
@flor.track
def fib(idx):
fib = {}
fib[log.param(0)] = log.metric(0)
fib[log.param(1)] = log.metric(1)
fib[log.param(2)] = log.metric(2)
for i in range(3, idx + 1):
fib[log.param(i)] = log.metric(fib[i - 1] + fib[i - 2])
with flor.Context('fib'):
fib(5)
optimizer.zero_grad()
loss.backward()
optimizer.step()
if (i + 1) % 100 == 0:
log.metric(epoch)
log.metric(i)
log.metric(loss.item())
# Test the model
# In test phase, we don't need to compute gradients (for memory efficiency)
with torch.no_grad():
correct = 0
total = 0
for images, labels in test_loader:
images = images.reshape(-1, 28 * 28).to(device)
labels = labels.to(device)
outputs = model(images)
_, predicted = torch.max(outputs.data, 1)
total += labels.size(0)
correct += (predicted == labels).sum().item()
acc = 100 * (correct / total)
log.metric(acc)
print('Accuracy of the network on the 10000 test images: {} %'.format(acc))
with flor.Context('pytorch_demo_nn'):
main()
import flor
from sklearn import datasets
from sklearn import svm
from sklearn.model_selection import train_test_split
log = flor.log
@flor.track
def fit_and_score_model(gamma, C, test_size, random_state):
iris = datasets.load_iris()
X_tr, X_te, y_tr, y_te = train_test_split(iris.data, iris.target,
test_size=log.param(test_size),
random_state=log.param(random_state))
clf = svm.SVC(gamma=log.param(gamma), C=log.param(C))
clf.fit(X_tr, y_tr)
score = log.metric(clf.score(X_te, y_te))
with flor.Context('iris'):
fit_and_score_model(gamma=0.001, C=100.0, test_size=0.15, random_state=430)
@flor.track
def main(x, y, z):
# Load the Data
movie_reviews = pd.read_json(log.read('data.json'))
movie_reviews['rating'] = movie_reviews['rating'].map(lambda x: 0
if x < z else 1)
# Do train/test split-
X_tr, X_te, y_tr, y_te = train_test_split(movie_reviews['text'],
movie_reviews['rating'],
test_size=log.param(x),
random_state=log.param(y))
# Vectorize the English sentences
vectorizer = TfidfVectorizer()
vectorizer.fit(X_tr)
X_tr = vectorizer.transform(X_tr)
X_te = vectorizer.transform(X_te)
# Fit the stateful
for i in [1, 5]:
clf = train_model(i, X_tr, y_tr)
test_model(clf, X_te, y_te)
the_answer_to_everything = log.param(42)
with flor.Context('basic'):
main(0.2, 92, 5)