AbstractIntegratedModule 1.1.7__tar.gz → 1.1.7.5__tar.gz

{abstractintegratedmodule-1.1.7 → abstractintegratedmodule-1.1.7.5}/AbstractIntegratedModule.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: AbstractIntegratedModule
-Version: 1.1.7
+Version: 1.1.7.5
 Summary: Integrated Pipeline with Specialized Non-LLM AI Agent Framework for ARM64 architecture
 Author: Micro-Novelty
 Author-email: Micro-Novelty <hernikpuspita5@gmail.com>

{abstractintegratedmodule-1.1.7 → abstractintegratedmodule-1.1.7.5}/AbstractIntegratedModule.py

@@ -4474,6 +4474,8 @@ class ExplainabilityModule:
     
     def _get_final_output(self, mlp_pred, mlp_conf, trans_pred, trans_conf, attn_weights):
         eps = 1e-5
+        consensus_conf = 0.0 
+
         if isinstance(mlp_conf, np.ndarray):
             mlp_conf = np.clip(np.mean(mlp_conf), 0, 1)
         if isinstance(trans_conf, np.ndarray):
@@ -4481,7 +4483,7 @@ class ExplainabilityModule:
 
         if mlp_pred == trans_pred:
             final_pred = mlp_pred
-            final_conf = max(mlp_conf, trans_conf)
+            consensus_conf = max(mlp_conf, trans_conf)
         else:
             sliced_attention_weight = attn_weights[0]
             if isinstance(sliced_attention_weight, np.ndarray):
@@ -4513,11 +4515,11 @@ class ExplainabilityModule:
             print('[=] Note: Very little Consistency meaning Transformer attention quality is Healthy and focused')
 
         if isinstance(consensus_conf, np.ndarray):
-            consensus_conf = 1.0 / (1.0 + np.exp(-final_conf))
+            consensus_conf = 1.0 / (1.0 + np.exp(-consensus_conf))
             # Apply a sigmoid transformation to ensure the confidence is between 0 and 1
 
         # averaged all confidences to get the final confidence.
-        final_conf = mlp_conf + trans_conf + final_conf / 3
+        final_conf = mlp_conf + trans_conf + consensus_conf / 3
         if np.isnan(final_conf).any() or np.isinf(final_conf).any():
             final_conf = self.pipeline.confidence_threshold
 
@@ -13701,11 +13703,6 @@ class IntegratedPipeline:
                     print(f'[= ! =] High zero-row ratio ({weak_ratio:.0%}), refitting on current batch')
                     self.tfidf.fit(X_raw_generation)
                     X_raw_features = self.tfidf.transform(X_raw_generation).toarray()  
-
-            should_train_transformer, reason = self._should_train_transformer(sequence_inputs, X_raw, min_seq_len=3, min_anisotropy=0.35, min_samples=10, ram_headroom_mb=max_ram_used)   
-            if should_train_transformer:      
-                transformer_features = self.transformer_pooled_features(sequence_inputs)
-                X_raw_features = np.concatenate([X_raw_features, transformer_features], axis=-1)
             
             zero_rows = np.where(X_raw_features.sum(axis=1) == 0)[0]
             if len(zero_rows) > 0:
@@ -13775,12 +13772,16 @@ class IntegratedPipeline:
                     num_classes=n_classes
                 )
                 
+            should_train_transformer, reason = self._should_train_transformer(sequence_inputs, X_raw, min_seq_len=3, min_anisotropy=0.35, min_samples=10, ram_headroom_mb=max_ram_used)                  
             if self.use_transformer and should_train_transformer:
                 print(f'[=] Transformer Training allowed, Reason: {reason}')
                 self.model2.train(sequence_inputs, y_true, epochs=self.transformer_training_epochs, mode=mode, lr=lr, embedded=True, batch_size=batch_size)
-                       
-            X = self.shape_adaptation(hybrid_X, input_dim)      
 
+            if should_train_transformer:      
+                transformer_features = self.transformer_pooled_features(sequence_inputs)
+                hybrid_X = np.concatenate([hybrid_X, transformer_features], axis=-1)  
+
+            X = self.shape_adaptation(hybrid_X, input_dim)      
             self.initialize_model_(X, input_dim, n_classes)
             if not unsuitable:
                 self.model3.train(X, y, epochs=self.mlp_training_epochs, lr=self.mlp_lr, max_samples_for_focused_fit=max_samples_for_focused_fit)