# ValueError: Expected input batch\_size (192) to match target batch\_size (64) **URL:** \ **Category:** vision\ **Created:** [June 24, 2020, 2:41pm UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739 "2020-06-24T14:41:28Z")\ **Posts on this page:** 10\ **Page:** 1
**Author:** ![TheDoctor](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/t/e274bd/32.png) [@TheDoctor](https://discuss.pytorch.org/u/TheDoctor)\ **Post date:** [June 24, 2020, 2:41pm UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/1 "2020-06-24T14:41:28Z")
Hi, i’m working with an RNN for Image Classification, my problem which i assume is that my loaded Images are 3 Channels RGB, if i only had one channel it would work i guess. Since it has 3 channels the input\_batchsize is 3 times higher than the target. Any ideas how i can solve this? I also checked [this post](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-324-to-match-target-batch-size-4/24498) which helped me already this far. But im stuck with the 3 channels now. The Input Shape is torch.Size([64, 3, 224, 224]) **X** after permute is torch.Size([224, 192, 224]) My parameters: # parameters - BATCH\_SIZE = 64 - N\_STEPS = 28 - N\_INPUTS = 224 - N\_CHANNELS = 3 - N\_NEURONS = 150 - N\_OUTPUTS = 21 - N\_EPOCHS = 5 - N\_PIXELS = 224 ```auto class ImageRNN(nn.Module): def __init__ (self, batch_size, n_steps, n_inputs, n_neurons, n_outputs): super(ImageRNN, self). __init__ () self.n_neurons = n_neurons self.batch_size = batch_size self.n_steps = n_steps self.n_inputs = n_inputs self.n_outputs = n_outputs self.basic_rnn = nn.RNN(self.n_inputs, self.n_neurons) self.FC = nn.Linear(self.n_neurons, self.n_outputs) def init_hidden(self,): # (num_layers, batch_size, n_neurons) return (torch.zeros(1, self.batch_size, self.n_neurons)) def forward(self, X): # transforms X to dimensions: n_steps X batch_size X n_inputs print(X.shape) X = X.permute(1, 0, 2) print(X.shape) print(X.size) self.batch_size = X.size(1) self.hidden = self.init_hidden() lstm_out, self.hidden = self.basic_rnn(X, self.hidden) out = self.FC(self.hidden) return out.view(-1, self.n_outputs) # batch_size X n_output ``` ```auto for i, data in enumerate(trainloader): # zero the parameter gradients optimizer.zero_grad() # reset hidden states model.hidden = model.init_hidden() # get the inputs inputs, labels = data inputs = inputs.view(-1, N_PIXELS,N_PIXELS) # forward + backward + optimize outputs = model(inputs) loss = criterion(outputs, labels) loss.backward() optimizer.step() ``` ---
**Author:** ![chetan\_patil](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/chetan_patil/32/25122_2.png) [@chetan\_patil](https://discuss.pytorch.org/u/chetan_patil)\ **Post date:** [June 25, 2020, 4:21am UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/2 "2020-06-25T04:21:43Z")
How do you want the size to be after `.permute` ? ---
**Author:** ![TheDoctor](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/t/e274bd/32.png) [@TheDoctor](https://discuss.pytorch.org/u/TheDoctor)\ **Post date:** [June 25, 2020, 12:01pm UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/3 "2020-06-25T12:01:47Z")
I want the size to be 64. ---
**Author:** ![chetan\_patil](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/chetan_patil/32/25122_2.png) [@chetan\_patil](https://discuss.pytorch.org/u/chetan_patil)\ **Post date:** [June 25, 2020, 2:58pm UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/4 "2020-06-25T14:58:45Z")
Before `permute` it was `[64,3,224,224]`. After permute, `[64, ?, ?]` ? based on your code, the RNN must accept size `(64, 224)` , how can you squash `(64,3,224,224)` into `(64,224)` ? ---
**Author:** ![TheDoctor](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/t/e274bd/32.png) [@TheDoctor](https://discuss.pytorch.org/u/TheDoctor)\ **Post date:** [June 25, 2020, 4:03pm UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/5 "2020-06-25T16:03:26Z")
Is there anyway i can use the 3 channels without loosing information? ---
**Author:** ![chetan\_patil](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/chetan_patil/32/25122_2.png) [@chetan\_patil](https://discuss.pytorch.org/u/chetan_patil)\ **Post date:** [June 26, 2020, 11:46am UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/6 "2020-06-26T11:46:37Z")
I’m not sure ny exact way, but you can flatten out `X` so that its shape now becomes `[64,3*224*224]`, and change your RNNs input size to be `3*224*224`. ---
**Author:** ![TheDoctor](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/t/e274bd/32.png) [@TheDoctor](https://discuss.pytorch.org/u/TheDoctor)\ **Post date:** [June 26, 2020, 2:05pm UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/7 "2020-06-26T14:05:14Z")
Thanks for the Input, i was thinking of reshaping my [64, 3, 224, 224] to a [64, 1, 672, 224]. So i just append the channels. I wouldnt loose information. What do you think about that solution, does it make sense? ---
**Author:** ![chetan\_patil](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/chetan_patil/32/25122_2.png) [@chetan\_patil](https://discuss.pytorch.org/u/chetan_patil)\ **Post date:** [June 26, 2020, 2:28pm UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/8 "2020-06-26T14:28:24Z")
Yes, if `X` is reshaped to something like `[64,1,672,672]`, you would again need to arrive at a shape of `[64,n_inputs]` ie `[64, 224]` so that it can be fed into the RNN. 1. Reshape it from `[64, 3, 224, 224]` to `[64, 150528]`, apply a feed-forward layer to get the desired size for the `basic_rnn` layer. 2. Apply 2 or 3 Convolutional layers and bring them to the desired shape. The 2nd approach will learn some spatial features as well, due to the Conv layers. ---
**Author:** ![TheDoctor](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/t/e274bd/32.png) [@TheDoctor](https://discuss.pytorch.org/u/TheDoctor)\ **Post date:** [June 26, 2020, 2:53pm UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/9 "2020-06-26T14:53:15Z")
Okay, what i did now is before putting my Images Batch into the Network i change inputs = inputs.view(-1, N\_N\_PIXELS,N\_PIXELS) to inputs = inputs.view(-1, N\_ **CHANNELS\*N** \_PIXELS,N\_PIXELS) which just appends all channels. Thanks again for your help @chetan_patil. ---
**Author:** ![chetan\_patil](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/chetan_patil/32/25122_2.png) [@chetan\_patil](https://discuss.pytorch.org/u/chetan_patil)\ **Post date:** [June 26, 2020, 3:00pm UTC](https://discuss.pytorch.org/t/valueerror-expected-input-batch-size-192-to-match-target-batch-size-64/86739/10 "2020-06-26T15:00:04Z")
You’re welcome @TheDoctor.