collect_inference

collect_inference.py

if config.is_comparison:
      comparison_stats = init_comparison_stats(id2label, config)
  for i, raw_instance in enumerate(test_data):
      print("Testing Artifact: "+str(i+1))
      actual_label_id = str(raw_instance['labels'].item())
      ground_truth_label = id2label[actual_label_id]
      print("\n--------------------------------------------")
      print("Ground Truth Label: " + ground_truth_label)
      if config.is_comparison:
          comparison_stats[ground_truth_label]['count'] += 1
      predictions = {}
      for model_name, model_data in model_containers.items():
          if config.is_comparison:
              raw_instance = get_model_specific_batch(raw_instance, model_name)
          instance = {}
          for input_key, input_value in raw_instance.items():
              if input_key != 'labels':
                  instance[input_key] = input_value.unsqueeze(0)

          with torch.no_grad():
              loaded_model = model_data['loaded_model']
              prediction = loaded_model(**instance)
              result = prediction['logits'].detach().cpu().numpy().argmax(-1)