Skip to content

Combinatorial Probabilistic Diabetes Diagnosis

Author: Jiawei Zhang
(Released: July 9, 2024; latest Revision: July 9, 2024.)
 Jupyter Logo Yaml Logo Python Logo

In this example, we will build a 1-layer RPN model with combinatorial_normal_expansion, identity_reconciliation and linear_remainder for diagnosing the diabetes disease based on the Pima Indians Diabetes dataset.

We use mps as the device for the model config file provided below.

Python Code and Model Configurations

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
from tinybig.config import rpn_config
from tinybig.util import set_random_seed
from tinybig.metric import accuracy

print('loading configs...')
config_file_name = 'diabetes_configs'
config_obj = rpn_config()
config = config_obj.load_yaml(cache_dir='./configs', config_file=config_file_name + '.yaml')
print(config)

print('setting up environments...')
device = config['configurations'].pop('device')
random_seed = config['configurations'].pop('random_seed')
set_random_seed(random_seed)
print('device: ', device, '; random_seed: ', random_seed)

print('instantiating objects from config...')
object_dict = config_obj.instantiate_object_from_config(config['configurations'])
data_obj, model_obj, learner_obj, metric_obj, result_obj = [object_dict[name] for name in ['data', 'model', 'learner', 'metric', 'result']]
print('parameter num: ', sum([parameter.numel() for parameter in model_obj.parameters()]))

print('loading dataset...')
data_loader = data_obj.load(train_percentage=0.9, normalize_X=True)

print('training model...')
training_records = learner_obj.train(model=model_obj, data_loader=data_loader, device=device, metric=metric_obj, disable_tqdm=True, display_step=100)
model_obj.save_ckpt(cache_dir='./ckpt', checkpoint_file=f'{config_file_name}_checkpoint')

print('testing model...')
test_result = learner_obj.test(model=model_obj, test_loader=data_loader['test_loader'], device=device,
                               metric=metric_obj)

print('evaluating result...')
print(metric_obj.__class__.__name__, metric_obj.evaluate(y_true=test_result['y_true'], y_pred=test_result['y_pred'], y_score=test_result['y_score'], ))

y_rounded_label = [[round(y[0])] for y in test_result['y_score']]
test_result['y_pred'] = y_rounded_label

acc_metric = accuracy('accuracy_metric')
print('evaluating rounded prediction labels...')
print(acc_metric.__class__.__name__, acc_metric.evaluate(y_true=test_result['y_true'], y_pred=test_result['y_pred'], y_score=test_result['y_score'], ))

print('saving result...')
result_obj.save(test_result, cache_dir='./result', output_file='{}_result'.format(config_file_name))
result_obj.save(training_records, cache_dir='./result', output_file='{}_record'.format(config_file_name))

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
configurations:
  device: &device mps
  random_seed: 1234

  data_configs:
    data_class: tinybig.data.diabetes
    data_parameters:
      name: diabetes
      train_batch_size: 1000
      test_batch_size: 1000

  learner_configs:
    learner_class: tinybig.learner.backward_learner
    learner_parameters:
      name: error_backward_propagation
      n_epochs: 3000
      optimizer_configs:
        optimizer_class: torch.optim.AdamW
        optimizer_parameters:
          lr: 5.0e-4
          weight_decay: 1.0e-05
      loss_configs:
        loss_class: torch.nn.MSELoss
        loss_parameters:
          reduction: mean

  model_configs:
    model_class: tinybig.model.rpn
    device: *device
    model_parameters:
      name: reconciled_polynomial_network
      depth: 1
      depth_alloc: [1]
      layer_configs:
        - layer_class: tinybig.module.rpn_layer
          layer_parameters:
            name: rpn_layer
            m: 8
            n: 1
            width: 1
            fusion_strategy: average
            width_alloc: [1]
            head_configs:
              - head_class: tinybig.module.rpn_head
                head_parameters:
                  l: null
                  channel_num: 1
                  data_transformation_configs:
                    data_transformation_class: tinybig.expansion.combinatorial_normal_expansion
                    data_transformation_parameters:
                      name: combinatorial_normal_expansion
                      d: 2
                  parameter_fabrication_configs:
                    parameter_fabrication_class: tinybig.reconciliation.identity_reconciliation
                    parameter_fabrication_parameters:
                      name: identity_reconciliation
                  remainder_configs:
                    remainder_class: tinybig.remainder.linear_remainder
                    remainder_parameters:
                      name: linear_remainder
                      require_parameters: True
  metric_configs:
    metric_class: tinybig.metric.mse
    metric_parameters:
      name: mse

  result_configs:
    result_class: tinybig.output.rpn_output
    result_parameters:
      name: prediction_output
rpn with identity reconciliation for mnist classification output 
training model...
Epoch: 0, Test Loss: 0.5340700745582581, Test Score: 0.5340701268297842, Time Cost: 0.006468772888183594
Epoch: 100, Test Loss: 0.29197242856025696, Test Score: 0.2919724107636688, Time Cost: 0.0055277347564697266
Epoch: 200, Test Loss: 0.2766035795211792, Test Score: 0.27660356945729825, Time Cost: 0.005342245101928711
Epoch: 300, Test Loss: 0.26146456599235535, Test Score: 0.2614645354786587, Time Cost: 0.005430936813354492
Epoch: 400, Test Loss: 0.24704204499721527, Test Score: 0.24704204207501665, Time Cost: 0.005308866500854492
Epoch: 500, Test Loss: 0.23358845710754395, Test Score: 0.233588461016266, Time Cost: 0.005259990692138672
Epoch: 600, Test Loss: 0.22124238312244415, Test Score: 0.22124238750083064, Time Cost: 0.005246162414550781
Epoch: 700, Test Loss: 0.21007299423217773, Test Score: 0.21007298251503007, Time Cost: 0.0055370330810546875
Epoch: 800, Test Loss: 0.2001015841960907, Test Score: 0.2001015802964822, Time Cost: 0.005261659622192383
Epoch: 900, Test Loss: 0.1913149207830429, Test Score: 0.19131491785419621, Time Cost: 0.0054361820220947266
Epoch: 1000, Test Loss: 0.1836737096309662, Test Score: 0.1836737053254299, Time Cost: 0.005302906036376953
Epoch: 1100, Test Loss: 0.17711907625198364, Test Score: 0.17711907725023052, Time Cost: 0.005380153656005859
Epoch: 1200, Test Loss: 0.17157748341560364, Test Score: 0.1715774714107086, Time Cost: 0.0056056976318359375
Epoch: 1300, Test Loss: 0.16696523129940033, Test Score: 0.16696524386354172, Time Cost: 0.0053501129150390625
Epoch: 1400, Test Loss: 0.16319161653518677, Test Score: 0.16319160136611355, Time Cost: 0.0054242610931396484
Epoch: 1500, Test Loss: 0.16016244888305664, Test Score: 0.16016245362139456, Time Cost: 0.0053369998931884766
Epoch: 1600, Test Loss: 0.15778301656246185, Test Score: 0.15778301431480563, Time Cost: 0.005410194396972656
Epoch: 1700, Test Loss: 0.15595999360084534, Test Score: 0.1559599993896924, Time Cost: 0.005301952362060547
Epoch: 1800, Test Loss: 0.15460419654846191, Test Score: 0.15460419713485304, Time Cost: 0.00522303581237793
Epoch: 1900, Test Loss: 0.15363219380378723, Test Score: 0.153632188729002, Time Cost: 0.005263090133666992
Epoch: 2000, Test Loss: 0.15296749770641327, Test Score: 0.15296750178082663, Time Cost: 0.005364894866943359
Epoch: 2100, Test Loss: 0.15254199504852295, Test Score: 0.15254199940270272, Time Cost: 0.0052661895751953125
Epoch: 2200, Test Loss: 0.15229643881320953, Test Score: 0.15229643199782647, Time Cost: 0.005544900894165039
Epoch: 2300, Test Loss: 0.15218046307563782, Test Score: 0.1521804629655372, Time Cost: 0.005230903625488281
Epoch: 2400, Test Loss: 0.15215271711349487, Test Score: 0.15215270496868652, Time Cost: 0.005433797836303711
Epoch: 2500, Test Loss: 0.15218019485473633, Test Score: 0.15218018954948764, Time Cost: 0.005357980728149414
Epoch: 2600, Test Loss: 0.15223759412765503, Test Score: 0.1522375949802569, Time Cost: 0.005576133728027344
Epoch: 2700, Test Loss: 0.15230637788772583, Test Score: 0.15230638071102204, Time Cost: 0.005346059799194336
Epoch: 2800, Test Loss: 0.15237337350845337, Test Score: 0.15237337481388422, Time Cost: 0.0055010318756103516
Epoch: 2900, Test Loss: 0.15243025124073029, Test Score: 0.15243023843696565, Time Cost: 0.005589008331298828
model checkpoint saving to ./ckpt/diabetes_configs_checkpoint...

evaluating result...
mse 0.15247175085902862

evaluating rounded prediction labels...
accuracy 0.8051948051948052