# Low accuracy when loading the model and testing

**URL:** <https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991>\
**Category:** Uncategorized\
**Created:** [May 11, 2019, 2:43pm UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991 "2019-05-11T14:43:59Z")\
**Posts on this page:** 11\
**Page:** 1

<div class="post-metadata">

**Author:** ![srikanthram](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/s/9fc348/32.png) [@srikanthram](https://discuss.pytorch.org/u/srikanthram)\
**Post date:** [May 11, 2019, 2:43pm UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/1 "2019-05-11T14:43:59Z")

</div>

Hey guys,

I trained my CIFAR-10 full precision network using vgg architecture. I got a 92.27 percent accuracy on the validation set. However, when I saved and loaded the model and then tested using the loop, I am getting only 35 percent accuracy. I am pretty sure the file is not corrupted.

```auto
model=torch.load('/content/cifar_fullprecison_vgg8.pth')
import torch
class Ternary_batch_rel(torch.nn.Module):
    
  def __init__ (self,batchnorm_size):
    super(Ternary_batch_rel,self). __init__ ()
    self.l1=torch.nn.Sequential(
    torch.nn.ReLU(),
    torch.nn.BatchNorm2d(batchnorm_size)
    )
  
  def forward(self,x):
    out=self.l1(x)
    return out
    
z1=Ternary_batch_rel(128).to(device)
z2=Ternary_batch_rel(256).to(device)
z3=Ternary_batch_rel(512).to(device)
class Ternary_max_pool(torch.nn.Module):
  def __init__ (self):
    
    super(Ternary_max_pool,self). __init__ ()
    self.l1=torch.nn.Sequential(
    torch.nn.MaxPool2d(kernel_size=2,stride=2))
      
  def forward(self,x):
    out=self.l1(x)
    return out

zm=Ternary_max_pool().to(device)

```

Testing loop

```auto
correct=0
total=0
for images,labels in (train_loader):
  images=images.to(device)
  labels=labels.to(device)
  y1=F.conv2d(images,model['layer1.0.weight'],padding=1)
  y2=z1(y1)
  y2=F.conv2d(y2,model['layer1.3.weight'],padding=1)
  y3=z1(y2)
  y3=zm(y3)
  y4=F.conv2d(y3,model['layer2.0.weight'],padding=1)
  y4=z2(y4)
  y5=F.conv2d(y4,model['layer2.3.weight'],padding=1)
  y5=z2(y5)
  y6=zm(y5)
  y7=F.conv2d(y6,model['layer3.0.weight'],padding=1)
  y8=z3(y7)
  y9=F.conv2d(y8,model['layer3.3.weight'],padding=1)
  y10=z3(y9)
  y11=zm(y10)
  y11=y11.view(y11.size(0),-1)
  y12=F.linear(y11,model['layer4.0.weight'])
  y13=F.relu(y12)
  y14=F.dropout(y13)
  y15=F.linear(y14,model['layer4.3.weight'])
  _,predicted=torch.max(y15,1)
  total+=labels.size(0)
  correct+=(predicted==labels).sum().item()
print('Test accuracy of the model on the 10000 test images:{}%'.format((correct/total)*100))

```

---

<div class="post-metadata">

**Author:** ![InnovArul](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/innovarul/32/5282_2.png) [@InnovArul](https://discuss.pytorch.org/u/InnovArul)\
**Post date:** [May 11, 2019, 3:07pm UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/2 "2019-05-11T15:07:13Z")

</div>

I’m not sure why you are using the model components explicitly than using the model definition file’s `forward` method. Any good reason for that?

What about bias terms in conv, linear layers? Are you using them too?

---

<div class="post-metadata">

**Author:** ![srikanthram](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/s/9fc348/32.png) [@srikanthram](https://discuss.pytorch.org/u/srikanthram)\
**Post date:** [May 11, 2019, 4:12pm UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/3 "2019-05-11T16:12:34Z")

</div>

I am training without the bias even the trained model that I am loading currently.

Also, I am implementing a paper that does not train weights rather than parameters of a weight drawn from some distribution and hence testing would not be straight forward( since weights need to be sampled from a distribution and then passed to infer the data)

---

<div class="post-metadata">

**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\
**Post date:** [May 11, 2019, 5:13pm UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/4 "2019-05-11T17:13:08Z")

</div>

Did you call `model.eval()` after loading the model?  
This is usually needed, if your model contains layers such as dropout and batchnorm.

---

<div class="post-metadata">

**Author:** ![srikanthram](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/s/9fc348/32.png) [@srikanthram](https://discuss.pytorch.org/u/srikanthram)\
**Post date:** [May 11, 2019, 5:24pm UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/5 "2019-05-11T17:24:45Z")

</div>

But the model here is a dictionary since it contains state\_dict . The model here is not an instance of the class.

---

<div class="post-metadata">

**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\
**Post date:** [May 11, 2019, 8:18pm UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/6 "2019-05-11T20:18:17Z")

</div>

It’s not completely clear to me how you are using the model, but even if you call the layers in a functional way, you should set them to `eval` to get the proper validation accuracy.  
You can call `.eval()` on each module separately.

---

<div class="post-metadata">

**Author:** ![srikanthram](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/s/9fc348/32.png) [@srikanthram](https://discuss.pytorch.org/u/srikanthram)\
**Post date:** [May 11, 2019, 11:09pm UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/7 "2019-05-11T23:09:57Z")

</div>

@ptrlblck, lets say if the model were an instance of a neural network class

```auto
model=Conv_net().to(device)

```

Now during the training phase, I would be using the forward function of the class to forward propagate and optimize the parameters. The parameters here are not the weights of the network, rather some parameters of a probability distribution to which the weights belong to.

During the eval() phase, I would have to first sample the trained probability distribution (paramerterized by the weight parameters) and then pass the sampled weights to the neural network  
so during the model.eval() phase, I would not be able to do something like the following:

```auto
model.eval()
for images,labels in valid_loader:
yout=model(images)
...
..

```

However, if i have the samples of the distribution, I want to do the following to see my validation accuracy:

```auto
y1=F.convd(images,w_sampled_1,padding=1)
y2=z1(y1)... 
etc

```

Is there a way to run inference of this kind using model instance rather than a dictionary?

---

<div class="post-metadata">

**Author:** ![srikanthram](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/s/9fc348/32.png) [@srikanthram](https://discuss.pytorch.org/u/srikanthram)\
**Post date:** [May 12, 2019, 12:45am UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/8 "2019-05-12T00:45:59Z")

</div>

@ptrblck you could check out the code written here:

> [@Blown up gradients and loss](https://discuss.pytorch.org/t/blown-up-gradients-and-loss/44886):
>
> I have my entire code written here and the output, I am surprised to see such huge gradients and losses. I am working on this code on CIFAR-10 based on local reparameterization trick by Shayer et.al, I am unable to reproduce the results in the paper. I have been working on this for over a month but no clue where I am going wrong. Could someone help me out Please. class Ternary\_batch\_rel(torch.nn.Module): def \_\_init\_\_(self,batchnorm\_size): super(Ternary\_batch\_rel,self).\_\_init\_\_() self.…

---

<div class="post-metadata">

**Author:** ![muammar](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/muammar/32/76367_2.png) [@muammar](https://discuss.pytorch.org/u/muammar)\
**Post date:** [July 18, 2019, 2:16am UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/9 "2019-07-18T02:16:19Z")

</div>

Did you solve this problem? I have realized that there is accuracy problems when I save the parameters of the models for certain epoch, and try to reproduce its outputs. For instance, suppose I stop the model at epoch 120, and save the model and its outputs. If I load the `state_dict` and predict over the same data points then there is about ~0.1% error. Why would that be? Is there something I should set to save the state\_dict with higher precision?

---

<div class="post-metadata">

**Author:** ![H\_MP](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/h/41988e/32.png) [@H\_MP](https://discuss.pytorch.org/u/H_MP)\
**Post date:** [August 6, 2022, 9:35pm UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/10 "2022-08-06T21:35:28Z")

</div>

I have same problem, I trained my model with 95% accuracy on trainset and 70% on validation, but when I use same data (validation-set) for inference, I’m getting awfull results. did you find an answer?

---

<div class="post-metadata">

**Author:** ![michaelchen78](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/michaelchen78/32/71102_2.png) [@michaelchen78](https://discuss.pytorch.org/u/michaelchen78)\
**Post date:** [June 30, 2024, 6:25am UTC](https://discuss.pytorch.org/t/low-accuracy-when-loading-the-model-and-testing/44991/11 "2024-06-30T06:25:58Z")

</div>

i know this is a bit late but I had a similar problem

for me specifying what device to use fixed it (very strange, no idea why)

I added

device = torch.device(“cpu”)  
model = model.to(device)

before training and corresponding .to(device) calls during training