Skip to content

Combinatorial Probabilistic Banknote Authentication

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 banknote disease based on the Banknote Authentication 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 = 'banknote_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.banknote
    data_parameters:
      name: banknote
      train_batch_size: 2000
      test_batch_size: 1000

  learner_configs:
    learner_class: tinybig.learner.backward_learner
    learner_parameters:
      name: error_backward_propagation
      n_epochs: 1500
      optimizer_configs:
        optimizer_class: torch.optim.AdamW
        optimizer_parameters:
          lr: 1.0e-02
          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: 4
            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: 1.9411336183547974, Test Score: 1.9411336696155832, Time Cost: 0.007844209671020508
Epoch: 100, Test Loss: 0.36659175157546997, Test Score: 0.366591744963078, Time Cost: 0.009141921997070312
Epoch: 200, Test Loss: 0.23675186932086945, Test Score: 0.23675185419647374, Time Cost: 0.008933782577514648
Epoch: 300, Test Loss: 0.17674371600151062, Test Score: 0.1767437075688446, Time Cost: 0.007529020309448242
Epoch: 400, Test Loss: 0.13803423941135406, Test Score: 0.1380342434864005, Time Cost: 0.0074062347412109375
Epoch: 500, Test Loss: 0.11002819240093231, Test Score: 0.11002819068762702, Time Cost: 0.007805824279785156
Epoch: 600, Test Loss: 0.08940108120441437, Test Score: 0.08940108904258437, Time Cost: 0.0074310302734375
Epoch: 700, Test Loss: 0.07411758601665497, Test Score: 0.07411758835844745, Time Cost: 0.0074651241302490234
Epoch: 800, Test Loss: 0.06268753856420517, Test Score: 0.06268753147146049, Time Cost: 0.007480144500732422
Epoch: 900, Test Loss: 0.054036639630794525, Test Score: 0.054036637264635624, Time Cost: 0.007607936859130859
Epoch: 1000, Test Loss: 0.04742664471268654, Test Score: 0.04742664752789012, Time Cost: 0.008695125579833984
Epoch: 1100, Test Loss: 0.04236285015940666, Test Score: 0.042362847759040485, Time Cost: 0.007869958877563477
Epoch: 1200, Test Loss: 0.03850637003779411, Test Score: 0.03850636577531528, Time Cost: 0.00732111930847168
Epoch: 1300, Test Loss: 0.03560876473784447, Test Score: 0.03560876589215595, Time Cost: 0.007487058639526367
Epoch: 1400, Test Loss: 0.03347256779670715, Test Score: 0.03347256692348597, Time Cost: 0.00847315788269043
model checkpoint saving to ./ckpt/banknote_configs_checkpoint...

evaluating result...
mse 0.03194525463030612

evaluating rounded prediction labels...
accuracy 0.9710144927536232