From d21d7df407189778d2300a09ddc16bc0989663fe Mon Sep 17 00:00:00 2001
From: Yandi <yandirzm@gmail.com>
Date: Sun, 22 Jan 2023 19:15:42 +0100
Subject: [PATCH] [Debug] Trying to understand why does not converge
---
main.py | 2 +-
model.py | 25 +++++++++++++++++++++----
train.py | 12 ++++++------
3 files changed, 28 insertions(+), 11 deletions(-)
diff --git a/main.py b/main.py
index 9b2efb8..3acf979 100644
--- a/main.py
+++ b/main.py
@@ -107,7 +107,7 @@ if __name__ == "__main__":
print(" Validation : Loss : {:.4f}".format(val_loss))
- create_submission.create_submission(network, None)
+ create_submission.create_submission(network, None, device)
"""
logdir = generate_unique_logpath(top_logdir, "linear")
print("Logging to {}".format(logdir))
diff --git a/model.py b/model.py
index a40839c..9a73117 100644
--- a/model.py
+++ b/model.py
@@ -11,11 +11,28 @@ class LinearRegression(nn.Module):
super(LinearRegression, self).__init__()
self.input_size = input_size
self.bias = cfg["LinearRegression"]["Bias"]
- self.regressor = nn.Linear(input_size, 1, self.bias)
- self.activate = nn.ReLU()
+ self.hidden_size = int(cfg["LinearRegression"]["HiddenSize"])
+ self.regressor = nn.Sequential(
+ nn.Linear(input_size,self.hidden_size,self.bias),
+ nn.ReLU(),
+ nn.Linear(self.hidden_size, self.hidden_size, self.bias),
+ nn.ReLU(),
+ nn.Linear(self.hidden_size,self.hidden_size,self.bias),
+ nn.ReLU(),
+ nn.Linear(self.hidden_size,self.hidden_size,self.bias),
+ nn.ReLU(),
+ nn.Linear(self.hidden_size,self.hidden_size,self.bias),
+ nn.ReLU(),
+ nn.Linear(self.hidden_size,self.hidden_size,self.bias),
+ nn.ReLU(),
+ nn.Linear(self.hidden_size,self.hidden_size, self.bias),
+ nn.ReLU(),
+ nn.Linear(self.hidden_size,1, self.bias),
+ nn.ReLU()
+ )
+
def forward(self, x):
- y = self.regressor(x).view((x.shape[0],-1))
- return self.activate(y)
+ return self.regressor(x)
def build_model(cfg, input_size):
return eval(f"{cfg['Model']['Name']}(cfg, input_size)")
diff --git a/train.py b/train.py
index cb594f8..a09e78a 100644
--- a/train.py
+++ b/train.py
@@ -41,14 +41,14 @@ def train(model, loader, f_loss, optimizer, device):
Y = list(model.parameters())[0].grad.cpu().tolist()
- gradients.append(np.mean(Y))
- tar.append(np.mean(outputs.cpu().tolist()))
- out.append(np.mean(targets.cpu().tolist()))
+ #gradients.append(np.mean(Y))
+ #tar.append(np.mean(outputs.cpu().tolist()))
+ #out.append(np.mean(targets.cpu().tolist()))
optimizer.step()
- visualize_gradients(gradients)
- visualize_gradients(tar)
- visualize_gradients(out)
+ #visualize_gradients(gradients)
+ #visualize_gradients(tar)
+ #visualize_gradients(out)
def visualize_gradients(gradients):
print(gradients)
--
GitLab