AttributeError: 'KMeans' object has no attribute 'labels_' with pytorch

first of all I thank , I tried to train model with pytorch but I got the following error: AttributeError: ‘KMeans’ object has no attribute ‘labels_’.I am trying to model a extract features point cloud using deep learning in pytorch. I get the following error . Could anyone help on this? ************** *************** Thanks!

# Training loop
def training_loop(gpu, training_dataloader, model, loss_fn, optimizer):
    losses = []
    correct = 0
    batch_results = dict()
    conf_mat = np.zeros((10,10))
    for batch_n, batch in enumerate(training_dataloader): #batch[batch, pos, ptr, y]
        batch_size = int(batch.batch.size()[0] / sample_points)        
        
        if dimensionality == 3:
            # Input dim [:,3] for your geometry x,y,z
            X = batch.pos.cuda(non_blocking=True).view(batch_size, sample_points, -1) + torch.normal(
                torch.zeros(batch_size, sample_points, dimensionality), torch.full((batch_size, sample_points,
                                                                                   dimensionality), fill_value=0.1)).cuda(gpu)
        else:
            # Input dim [:,6] for your geometry x,y,z and normals nx,ny,nz
            X = torch.cat((batch.pos.cuda(non_blocking=True), batch.normal.cuda(non_blocking=True)), 1).view(batch_size, sample_points, -1) + torch.normal(
                torch.zeros(batch_size, sample_points, dimensionality), torch.full((batch_size, sample_points,
                                                                                   dimensionality), fill_value=0.1)).cuda(gpu)
        
        y = batch.y.cuda(non_blocking=True).flatten() #size (batch_size) --> torch.Size([8])
        
        # Compute predictions
        pred = model(None, X) #size (batch_size,classes) --> torch.Size([8, 10])
        
        if overall_classes_loss:
            # weighted CE Loss over all classes
            loss = loss_fn(pred, y)
        else:
            # weighted batchwise Loss
            sample_count = np.array([[x, batch.y.tolist().count(x)] for x in batch.y])[:,1]
            batch_weights = 1. / sample_count
            batch_weights = torch.from_numpy(batch_weights)
            batch_weights = batch_weights.double()
            loss = element_weighted_loss(pred, batch.y, batch_weights, gpu)
        
        correct += (pred.argmax(1) == y).type(torch.float).sum().item()
        print(f"Loss: {loss}")

        tensor_list_y =  [torch.ones_like(y) for _ in range(dist.get_world_size())]
        tensor_list_pred = [torch.ones_like(y) for _ in range(dist.get_world_size())]
        torch.distributed.all_gather(tensor_list_y, y, group=None, async_op=False)
        torch.distributed.all_gather(tensor_list_pred, pred.argmax(1), group=None, async_op=False)
        tensor_list_y = torch.cat(tensor_list_y)
        tensor_list_pred = torch.cat(tensor_list_pred)
        
        # Confusion Matrix
        conf_mat += confusion_matrix(tensor_list_y.cpu().detach().numpy(), tensor_list_pred.cpu().detach().numpy(), labels=np.arange(0,10))
        
        # Backpropagation
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        losses.append(loss.item())
        
        # Save batch predictions
        batch_results[batch_n] = {'true':tensor_list_y, 'pred':tensor_list_pred}
        
        if verbosity == True:
            print(f"\n\nTRAIN on GPU:{gpu}: True Label {y} - Prediction {pred.argmax(1)} - Loss {loss}")
            truevalue = '\t\t'.join(classes[items] for items in y.tolist())
            predvalues = '\t\t'.join(classes[items] for items in pred.argmax(1).tolist())
            print(f"INFO on GPU:{gpu}: TRAIN - True Value\t {truevalue}")
            print(f"INFO on GPU:{gpu}: TRAIN - Predictions\t {predvalues}")
        
        if batch_n % 25 == 0:
            torch.distributed.reduce(loss, 0)
            
    return torch.tensor(losses, device=f"cuda:{gpu}"), torch.tensor(correct, device=f"cuda:{gpu}"), batch_results, conf_mat

# Test loop
def test_loop(gpu, test_dataloader, model, loss_fn):
    test_losses = []
    correct = 0
    batch_results = dict()
    conf_mat = np.zeros((10,10))
    with torch.no_grad():
        for batch_n, batch in enumerate(test_dataloader):
            batch_size = int(batch.batch.size()[0] / sample_points)
            
            if dimensionality == 3:
                # Input dim [:,3] for your geometry x,y,z
                X = batch.pos.cuda(non_blocking=True).view(batch_size, sample_points, -1)
            else:
                # Input dim [:,6] for your geometry x,y,z and normals nx,ny,nz
                X = torch.cat((batch.pos.cuda(non_blocking=True), batch.normal.cuda(non_blocking=True)), 1).view(batch_size, sample_points, -1)
                
            y = batch.y.cuda(non_blocking=True).flatten()
            pred = model(None, X) #size (batch,classes) per batch_n
            
            if overall_classes_loss:
                # weighted CE Loss over all classes
                loss = loss_fn(pred, y)
            else:
                # weighted batchwise Loss
                sample_count = np.array([[x, batch.y.tolist().count(x)] for x in batch.y])[:,1]
                batch_weights = 1. / sample_count
                batch_weights = torch.from_numpy(batch_weights)
                batch_weights = batch_weights.double()
                loss = element_weighted_loss(pred, batch.y, batch_weights, gpu)
            
            test_losses.append(loss.item())
            correct += (pred.argmax(1) == y).type(torch.float).sum().item()
            print(f"Loss: {loss}")
        
            tensor_list_y =  [torch.ones_like(y) for _ in range(dist.get_world_size())]
            tensor_list_pred = [torch.ones_like(y) for _ in range(dist.get_world_size())]
            torch.distributed.all_gather(tensor_list_y, y, group=None, async_op=False)
            torch.distributed.all_gather(tensor_list_pred, pred.argmax(1), group=None, async_op=False)
            tensor_list_y = torch.cat(tensor_list_y)
            tensor_list_pred = torch.cat(tensor_list_pred)
            
            # Confusion Matrix
            conf_mat += confusion_matrix(tensor_list_y.cpu().detach().numpy(), tensor_list_pred.cpu().detach().numpy(), labels=np.arange(0,10))
            
            # Save batch predictions
            batch_results[batch_n] = {'true':tensor_list_y, 'pred':tensor_list_pred}
            
            if verbosity == True:
                print(f"\n\nTEST on GPU:{gpu}: True Label {y} - Prediction {pred.argmax(1)} - Loss {loss}")
                truevalue = '\t\t'.join(classes[items] for items in y.tolist())
                predvalues = '\t\t'.join(classes[items] for items in pred.argmax(1).tolist())
                print(f"INFO on GPU:{gpu}: TEST - True Value\t {truevalue}")
                print(f"INFO on GPU:{gpu}: TEST - Predictions\t {predvalues}")

    test_loss = statistics.mean(test_losses)
    return torch.tensor(correct, device=f"cuda:{gpu}"), torch.tensor(test_loss, device=f"cuda:{gpu}"), batch_results, conf_mat
def train_optimisation(gpu, gpus, training_dataloader, test_dataloader, model, loss_fn, optimizer, scheduler, dir_path, initial_epoch):
    epoch_losses = []
    training_accuracies = []
    test_losses = []
    test_accuracies = []
    learning_rates = []
    counter = 0 #early stopping counter
    batchwise_results = dict()
    
    # Learning Rate Scheduler
    scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, mode='min', patience=20)

    for i in range(initial_epoch, initial_epoch + epochs):
        if gpu == 0:
            if initial_epoch > 0:
                print(f"\n\nEpoch {i}\n-------------------------------")
            else:
                print(f"\n\nEpoch {i + 1}\n-------------------------------")

        # TRAIN
        losses, training_accuracy, train_batch_result, train_conf_mat = training_loop(gpu, training_dataloader, model, loss_fn, optimizer)
        average_loss = torch.mean(losses)
        torch.distributed.reduce(average_loss, 0, torch.distributed.ReduceOp.SUM)
        torch.distributed.reduce(training_accuracy, 0, torch.distributed.ReduceOp.SUM)
        
        # TEST
        test_accuracy, test_loss, test_batch_result, test_conf_mat = test_loop(gpu, test_dataloader, model, loss_fn)
        torch.distributed.reduce(test_accuracy, 0, torch.distributed.ReduceOp.SUM)
        torch.distributed.reduce(test_loss, 0, torch.distributed.ReduceOp.SUM)
                
        # save results
        batchwise_results[i] = {'train':train_batch_result, 'test':test_batch_result}
        if gpu == 0:  # the following operations are performed only by the process running in the first gpu
            average_loss = average_loss / torch.tensor(gpus, dtype=torch.float)  # average loss among all gpus
            test_accuracy = test_accuracy / torch.tensor(len(test_dataloader.dataset),
                                                         dtype=torch.float) * torch.tensor(100.0)
            training_accuracy = training_accuracy / torch.tensor(len(training_dataloader.dataset),
                                                                 dtype=torch.float) * torch.tensor(100.0)
            test_loss = test_loss / torch.tensor(gpus, dtype=torch.float)
            epoch_losses.append(average_loss.item())
            training_accuracies.append(training_accuracy.item())
            test_losses.append(test_loss.item())
            test_accuracies.append(test_accuracy.item())
            learning_rates.append((optimizer.param_groups[0])["lr"])
            print(f"\nBatch size: {batch_size * int(gpus)}")
            print(f"average Training Loss: {average_loss.item():.6f}")
            print(f"average Test Loss: {test_loss.item():.6f}")
            print(f"\naverage Training Acc: {training_accuracy.item():.6f}")
            print(f"average Test Acc: {test_accuracy.item():.6f}")
            printLearningRate(optimizer)
            scheduler.step(test_loss)
            
                        
            # saving model checkpoint
            save_checkpoint(model, optimizer, scheduler, i, epoch_losses, training_accuracies, test_losses, test_accuracies, learning_rates,
                            os.path.join(dir_path, f"epoch{i}.pth"), {key: value for key, value in batchwise_results[i].items() if key == 'train'}, {key: value for key, value in batchwise_results[i].items() if key == 'test'}, train_conf_mat, test_conf_mat)
            #TODO: implement ONNX Export
            # early stopping scheduler
            if early_stopping(test_losses) == True:
                counter += 1
                print(f"Early Stopping counter: {counter} of {patience}")
            else:
                counter += 0                        
            if counter < patience:
                pass
            else:
                print("\n\nEarly Stopping activated")
                print(f"Training stopped at Epoch{i + 1}")
                dist.destroy_process_group()
                exit()

def train(gpu, gpus, world_size):
    torch.manual_seed(0)
    torch.cuda.set_device(gpu)
    try:
        dist.init_process_group(backend='nccl', world_size=world_size, rank=gpu) #for distributed GPU training
    except RuntimeError:
        print("\n\nINFO:RuntimeError is raised >> Used gloo backend instead of nccl!\n")
        dist.init_process_group(backend='gloo', world_size=world_size, rank=gpu) #as a fallback option
    
    dir_path = None
    if gpu == 0:
        dir_path = "stackgraphConvPool3DPnet"
        createdir(dir_path)
        training_number = next_training_number(dir_path)
        dir_path = os.path.join(dir_path, f"train{training_number}")
        createdir(dir_path)
        #save hyper-parameters in txt protocol file
        save_hyperparameters(dir_path, 'hyperparameters.txt')
        print("\nINFO: Protocol File saved successfully . . .")
        
        
    model = Classifier(shrinkingLayers, mlpClassifier)

    torch.cuda.set_device(gpu)
    model.cuda(gpu)
    
    #setting up optimizer
    if optimizer_str == "SGD":
        optimizer = torch.optim.SGD(model.parameters(), learning_rate, momentum=momentum, weight_decay=weight_decay)
    elif optimizer_str == "RMSprop":
        optimizer = torch.optim.RMSprop(model.parameters(), learning_rate, weight_decay=weight_decay)
    else:
        optimizer = torch.optim.Adam(model.parameters(), learning_rate, weight_decay=weight_decay)
    
    # single-program multiple-data training paradigm (Distributed Data-Parallel Training)
    model = DDP(model, device_ids=[gpu])
    
    if dimensionality == 3:
        training_data = ModelNet("ModelNet10_train_data", transform=lambda x: NormalizeScale()(SamplePoints(num=sample_points)(x)))
    else:
        training_data = ModelNet("ModelNet10_train_data", transform=lambda x: NormalizeScale()(NormalizeRotation()(SamplePoints(num=sample_points, remove_faces=True, include_normals=True)(x))))
    
    training_sampler = DistributedWeightedSampler(training_data, num_replicas=world_size) #weight unbalanced classes by 1/cls_count
    training_dataloader = DataLoader(dataset=training_data, batch_size=batch_size, shuffle=data_shuffle, num_workers=0,
                                          pin_memory=True, sampler=training_sampler)
    
    if dimensionality == 3:
        test_data = ModelNet("ModelNet10_test_data", train=False, transform=lambda x: NormalizeScale()(SamplePoints(num=sample_points)(x)))
    else:
        test_data = ModelNet("ModelNet10_test_data", train=False, transform=lambda x: NormalizeScale()(NormalizeRotation()(SamplePoints(num=sample_points, remove_faces=True, include_normals=True)(x))))
    
    test_sampler = DistributedWeightedSampler(test_data, num_replicas=world_size) #weight unbalanced classes by 1/cls_count
    test_dataloader = DataLoader(dataset=test_data, batch_size=batch_size, shuffle=data_shuffle, num_workers=0,
                                      pin_memory=True, sampler=test_sampler)
    
    
    # weighted CE Loss over all Classes C
    class_sample_count = np.array([len(np.where(training_data.data.y == t)[0]) for t in np.unique(training_data.data.y)])
    weight = 1. / class_sample_count
    weight = torch.from_numpy(weight)
    weight = weight.float()
    loss_fn = nn.CrossEntropyLoss(weight=weight).cuda(gpu)
    
    # continue training from certain checkpoint
    continue_from_scratch = True if args.resume is None else False
    if continue_from_scratch: 
        if gpu == 0:
            print("\nINFO: Train from scratch has started . . .")
        train_optimisation(gpu, gpus, training_dataloader, test_dataloader, model, loss_fn, optimizer, None, dir_path, 0)
    else:
        checkpoint_path = "stackgraphConvPool3DPnet/" + args.resume
        if gpu == 0:
            print(f"\nINFO: Train has started from certain checkpoint {checkpoint_path.split('/')[2].split('.')[0]} in {checkpoint_path.split('/')[1]} . . .")
        model.load_state_dict(torch.load(checkpoint_path)['model_state_dict'], strict=False)
        optimizer.load_state_dict(torch.load(checkpoint_path)['optimizer_state_dict'])
        final_epoch = (torch.load("stackgraphConvPool3DPnet/" + args.resume)['epoch'])+1
        train_optimisation(gpu, gpus, training_dataloader, test_dataloader, model, loss_fn, optimizer, None, dir_path, final_epoch)

code tools:



class KMeansInitMostDistantFromMean:
    def __call__(self, *args, **kwargs):
        X, k = args
        mean = np.mean(X, axis=0)
        arg_sorted = np.argsort(np.apply_along_axis(lambda y: euclidean(mean, y), 1, X))
        output = X[np.flip(arg_sorted)[:k]]
        return output

class KMeansInit:
    def __call__(self, *args, **kwargs):
        X, k = args
        current_centroids = np.expand_dims(np.mean(X, axis=0), 0)
        for i in range(k - 1):
            X, current_centroids = self.next_centroid(X, current_centroids)

        return current_centroids

    def next_centroid(self, X, curr_centroids):
        highest_dist = 0.0
        next_centroid = None
        next_centroid_index = None
        for i, x in enumerate(X):
            max_dist = np.amax(np.apply_along_axis(lambda y: euclidean(x, y), 1, curr_centroids))
            if max_dist > highest_dist:
                next_centroid = x
                highest_dist = max_dist
                next_centroid_index = i

        return np.delete(X, next_centroid_index, 0), np.append(curr_centroids, np.expand_dims(next_centroid, 0), 0)


class Conv(gnn.MessagePassing):
   
    def __init__(self, sigma: nn.Module, F: nn.Module, W: nn.Module, M: nn.Module, C: int, P: int):
       
        super().__init__(aggr="mean")
        self.sigma = sigma
        self.F = F
        self.W = W
        self.M = M
        self.C = C
        self.P = P
        self.B = torch.randn(C+P, requires_grad=True)

    def forward(self, feature_matrix, edge_index):
        return self.propagate(edge_index, feature_matrix=feature_matrix)

    def message(self, feature_matrix_i, feature_matrix_j):
        message = self.F(feature_matrix_j - feature_matrix_i)
        message = message.view(-1, self.C + self.P, self.C)
        feature_matrix_i_ = feature_matrix_i.unsqueeze(2)
        output = torch.bmm(message, feature_matrix_i_).squeeze()
        return output

    def update(self, aggr_out, feature_matrix):
        Weight = self.M(aggr_out)
        aggr_out = aggr_out * Weight
        transform = self.W(feature_matrix)
        transform = transform.view(-1, self.C + self.P, self.C)
        feature_matrix = feature_matrix.unsqueeze(2)
        transformation = torch.bmm(transform, feature_matrix).squeeze()
        aggr_out = aggr_out + transformation
        output = aggr_out + self.B
        output = self.sigma(output)
        return output


class Aggregation(nn.Module):
    
    def __init__(self, mlp1: nn.Module, mlp2: nn.Module):
       
        super().__init__()
        self.mlp1 = mlp1
        self.mlp2 = mlp2
        self.softmax = nn.Softmax(0)

    def forward(self, feature_matrix_batch: torch.Tensor, conv_feature_matrix_batch: torch.Tensor):
        N, I, D = feature_matrix_batch.size()
        N_, I_, D_ = conv_feature_matrix_batch.size()
        augmentation = D_ - D
        if augmentation > 0:
            feature_matrix_batch = F.pad(feature_matrix_batch, (0, augmentation))
        S1 = torch.mean(feature_matrix_batch, 1)
        S2 = torch.mean(conv_feature_matrix_batch, 1)
        Z1 = self.mlp1(S1)
        Z2 = self.mlp2(S2)
        M = self.softmax(torch.stack((Z1, Z2), 0))
        M1 = M[0]
        M2 = M[1]
        M1 = M1.unsqueeze(1).expand(-1, I, -1)
        M2 = M2.unsqueeze(1).expand(-1, I, -1)
        output = (M1 * feature_matrix_batch) + (M2 * conv_feature_matrix_batch)
        return output


class MaxPool(nn.Module):
    
    def __init__(self, k: int):
        
        super().__init__()
        self.k = k

    def forward(self, feature_matrix_batch: torch.Tensor, cluster_index: torch.Tensor):
       
        N, I, D = feature_matrix_batch.size()
        feature_matrix_batch = feature_matrix_batch.view(-1, D)
        output = scatter_max(feature_matrix_batch, cluster_index, dim=0)[0]
        output = output.view(N, self.k, -1)
        
        return output


class GraphConvPool3DPnet(nn.Module):
    
    def __init__(self, shrinkingLayers: [ShrinkingUnit], mlp: nn.Module):
        super().__init__()
        self.neuralNet = nn.Sequential(*shrinkingLayers, mlp)

    def forward(self, x: torch.Tensor, pos: torch.Tensor):
        feature_matrix_batch = torch.cat((pos, x), 2) if x is not None else pos
        return self.neuralNet(feature_matrix_batch)

class ShrinkingUnitStack(nn.Module):
   
    def __init__(self, input_stack: int, stack_fork: int, mlp: nn.Module, learning_rate: int, k: int, kmeansInit, n_init, sigma: nn.Module, F: nn.Module, W: nn.Module,
                 M: nn.Module, C, P, mlp1: nn.Module, mlp2: nn.Module):
       
        super().__init__()
        self.stack_fork = stack_fork
        stack_size = input_stack * stack_fork
        self.selfCorrStack = SelfCorrelationStack(stack_size, mlp, learning_rate)
        self.kmeansConvStack = KMeansConvStack(stack_size, k, kmeansInit, n_init, sigma, F, W, M, C, P)
        self.localAdaptFeaAggreStack = AggregationStack(stack_size, mlp1, mlp2)
        self.graphMaxPoolStack = MaxPoolStack(stack_size, k)

    def forward(self, feature_matrix_batch):
        
        feature_matrix_batch = torch.repeat_interleave(feature_matrix_batch, self.stack_fork, dim=0)
        
        feature_matrix_batch = self.selfCorrStack(feature_matrix_batch)
        
        feature_matrix_batch_, conv_feature_matrix_batch, cluster_index = self.kmeansConvStack(feature_matrix_batch)
        feature_matrix_batch = self.localAdaptFeaAggreStack(feature_matrix_batch, conv_feature_matrix_batch)
        output = self.graphMaxPoolStack(feature_matrix_batch, cluster_index)
       
        return output


class SelfCorrelationStack(nn.Module):
   
    def __init__(self, stack_size: int, mlp: nn.Module, learning_rate: int = 1.0):
       
        super().__init__()
        self.selfCorrelationStack = nn.ModuleList([SelfCorrelation(copy.deepcopy(mlp), learning_rate) for i in range(stack_size)])
        self.apply(init_weights)

    def forward(self, feature_matrix_batch: torch.Tensor):
        # feature_matrix_batch size = (S,N,I,D) where S=stack_size, N=batch number, I=members, D=member dimensionality
        output = selfCorrThreader(self.selfCorrelationStack, feature_matrix_batch)
        # output size = (S,N,I,D) where where S=stack_size, N=batch number, I=members, D=member dimensionality
        return output


class KMeansConvStack(nn.Module):
    
    def __init__(self, stack_size: int, k: int, kmeansInit, n_init: int, sigma: nn.Module, F: nn.Module, W: nn.Module,
                 M: nn.Module, C: int, P: int):
        
        super().__init__()
        self.kmeansConvStack = nn.ModuleList([
            KMeansConv(k, kmeansInit, n_init, copy.deepcopy(sigma), copy.deepcopy(F), copy.deepcopy(W),
                       copy.deepcopy(M), C, P) for i in range(stack_size)])
        self.apply(init_weights)

    def forward(self, feature_matrix_batch: torch.Tensor):
        # feature_matrix_batch size = (S,N,I,D) where S=stack size, N=batch number, I=members, D=member dimensionality
        feature_matrix_batch, conv_feature_matrix_batch, cluster_index = kmeansConvThreader(self.kmeansConvStack,
                                                                                            feature_matrix_batch)
  
        return feature_matrix_batch, conv_feature_matrix_batch, cluster_index


class AggregationStack(nn.Module):
    
    def __init__(self, stack_size: int, mlp1: nn.Module, mlp2: nn.Module):
       
        super().__init__()
        self.localAdaptFeatAggreStack = nn.ModuleList([Aggregation(copy.deepcopy(mlp1), copy.deepcopy(mlp2)) for i
                                                       in range(stack_size)])
        self.apply(init_weights)

    def forward(self, feature_matrix_batch: torch.Tensor, conv_feature_matrix_batch: torch.Tensor):
        
        output = threader(self.localAdaptFeatAggreStack, feature_matrix_batch, conv_feature_matrix_batch)
        
        return output


class MaxPoolStack(nn.Module):
    
    def __init__(self, stack_size: int, k: int):
        
        super().__init__()
        self.graphMaxPoolStack = nn.ModuleList([MaxPool(k) for i in range(stack_size)])
        self.apply(init_weights)

    def forward(self, feature_matrix_batch: torch.Tensor, cluster_index: torch.Tensor):
        
        output = threader(self.graphMaxPoolStack, feature_matrix_batch, cluster_index)
       
        return output


def selfCorrThreader(modules, input_tensor):
    list_append = []
    threads = []
    for i, t in enumerate(input_tensor):
        threads.append(Thread(target=selfCorrAppender, args=(modules[i], t, list_append, i)))
    [t.start() for t in threads]
    [t.join() for t in threads]
    list_append.sort()
    list_append = list(map(lambda x: x[1], list_append))
    return torch.stack(list_append)


def selfCorrAppender(module, tensor, list_append, index):
    list_append.append((index, module(tensor)))


def kmeansConvThreader(modules, input_tensor):
    list1_append = []
    list2_append = []
    list3_append = []
    threads = []
    for i, t in enumerate(input_tensor):
        threads.append(
            Thread(target=kmeansAppender, args=(modules[i], t, list1_append, list2_append, list3_append, i)))
    [t.start() for t in threads]
    [t.join() for t in threads]
    list1_append.sort()
    list2_append.sort()
    list3_append.sort()
    list1_append = list(map(lambda x: x[1], list1_append))
    list2_append = list(map(lambda x: x[1], list2_append))
    list3_append = list(map(lambda x: x[1], list3_append))
    return torch.stack(list1_append), torch.stack(list2_append), torch.stack(list3_append)


def kmeansAppender(module, input, list1_append, list2_append, list3_append, index):
    x, y, z = module(input)
    list1_append.append((index, x))
    list2_append.append((index, y))
    list3_append.append((index, z))


def threader(modules, input_tensor1, input_tensor2):
    list_append = []
    threads = []
    for i, t in enumerate(input_tensor1):
        threads.append(Thread(target=threaderAppender, args=(modules[i], t, input_tensor2[i], list_append, i)))
    [t.start() for t in threads]
    [t.join() for t in threads]
    list_append.sort()
    list_append = list(map(lambda x: x[1], list_append))
    return torch.stack(list_append)


def threaderAppender(module, t1, t2, list_append, index):
    list_append.append((index, module(t1, t2)))


class Classifier(nn.Module):
    
    def __init__(self, shrinkingLayersStack: [ShrinkingUnitStack], mlp: nn.Module):
       
        super().__init__()
        self.neuralNet = nn.Sequential(*shrinkingLayersStack)
        self.mlp = mlp

    def forward(self, x: torch.Tensor, pos: torch.Tensor):
       
        feature_matrix_batch = pos.unsqueeze(0)
        
        output = self.neuralNet(feature_matrix_batch)
        
        output = torch.mean(output, dim=0)
        
        return self.mlp(output)

Error:

enter image description here

enter image description here

enter image description here
enter image description here
enter image description here
enter image description here

thank you for your help

I don’t know how the KMeans object is defined and which attributes would be expected, so you might want to add more information to this topic. It’s also a bit unclear how the screenshots are related to the initial error as it seems different errors are shown.
Posting screenshots is also not a good idea as it would not allow us to copy code and might also be hard to interpret.