def main():
train_alignment, train_reviews = getdata.readFiles("train.npz")
train_alignment = getdata.decode(train_alignment)
# train_alignment = getdata.describeLabels(train_alignment)
train_reviews = getdata.decode(train_reviews)
test_align, test_reviews = getdata.readFiles("test.npz")
test_align = getdata.decode(test_align)
# test_align = getdata.describeLabels(test_align)
test_reviews = getdata.decode(test_reviews)
model = analyze.train(train_reviews, test_reviews, train_alignment)
# analyze.predict(test_reviews,model)
analyze.predict(test_reviews, test_align, model)
def set_d(file, Kernel, df, df1):
cons = 1.19
eps = 0.008
gam = 0.707
p=1
predict(df, Kernel, cons, eps, p, gam, "ARCHITECTURE2")
predict(df1, Kernel, cons, eps, p, gam, "ARCHITECTURE1")
print("====================================================")
async def receiver(websocket, path):
"""
Socket that handles segmentation and predict queries
INPUT: websocket - the tcp connection
path - path to the image
OUTPUT: none (except this function sends result back to nodejs client)
"""
jsonString = await websocket.recv()
jsonObject = json.loads(jsonString)
print(jsonObject["type"])
if (jsonObject["type"] == "segmentation"):
filename = jsonObject["filename"]
path = "../communication/rawimage/" + filename
norm_points = segmentation(path)
obj = {
"type": "segmentation",
"boxes": norm_points,
"iou": 0.65,
}
string = json.dumps(obj)
await websocket.send(string)
if (jsonObject["type"] == "predict"):
filename = jsonObject["filename"]
path = "../communication/rawimage/" + filename
global model
norm_points, labels, percentages = predict(model, path)
print("done")
obj = {
"type": "predict",
"boxes": norm_points,
"labels": labels,
}
string = json.dumps(obj)
print("sent")
await websocket.send(string)
def set_e(file, Kernel, df, df1):
cons = 1.00
eps = 0.011
gam = 0.841
p=1
predict(df, Kernel, cons, eps, p, gam, "ARCHITECTURE2")
predict(df1, Kernel, cons, eps, p, gam, "ARCHITECTURE1")
print("====================================================")
def set_a(file, Kernel, df, df1):
cons = 26.91
eps = 0.0004
gam = 0.105
p=1
predict(df, Kernel, cons, eps, p, gam, "ARCHITECTURE2")
predict(df1, Kernel, cons, eps, p, gam, "ARCHITECTURE1")
print("====================================================")
def set_c(file, Kernel, df, df1):
cons = 1.41
eps = 0.006
gam = 0.297
p=1
#validate(df, Kernel, cons, eps, p, gam)
predict(df, Kernel, cons, eps, p, gam, "ARCHITECTURE2")
predict(df1, Kernel, cons, eps, p, gam, "ARCHITECTURE1")
print("====================================================")
def set_f(file, Kernel, df, df1):
cons = 1.00
eps = 0.011
gam = 0.189
p=1
#validate(df, Kernel, cons, eps, p, gam)
predict(df, Kernel, cons, eps, p, gam, "ARCHITECTURE2")
predict(df1, Kernel, cons, eps, p, gam, "ARCHITECTURE1")
print("====================================================")
inputs, targets = dataset()
model.train()
optimizer.zero_grad()
x = torch.tensor(inputs).to(device)
y = torch.tensor(targets).to(device)
logits, (state_h, state_c) = model(x, (state_h, state_c))
loss = criterion(logits.transpose(1, 2), y)
loss_value = loss.item()
loss.backward()
state_h = state_h.detach()
state_c = state_c.detach()
torch.nn.utils.clip_grad_norm_(model.parameters(), flags.gradients_norm)
optimizer.step()
if iterator % 100 == 0:
print('Epoch: {}/{}'.format(epoch, flags.epochs),
'Iteration: {}'.format(iterator), 'Loss: {}'.format(loss_value))
if iterator % (epoch * flags.max_batch) == 0:
predict(device, model, dataset.vocabulary, top_k=5)
torch.save(model.state_dict(),
'LSTM-word-level/states_testing/epoch-{}.pth'.format(epoch))
