- Pushed logs from NEC test runs

- Naming corrections for consistency, comments, removed superfluous files

.gitignore

@@ -118,4 +118,4 @@ train.json
 logs
 
 # evaluation output
-results
+#results

data/PileNERtype/pile_interface.py

@@ -96,7 +96,7 @@ def get_annotated_sentences():
                                 for ent in entities:
                                     annotations.append((ent, entity_type))
                         except json.JSONDecodeError:
-                            # If parsing fails, we can skip or log the error.
+                            # If parsing fails, skip the error.
                             pass
                     i += 2  # Skip the GPT answer
                 else:

src/common_interface.py

@@ -2,7 +2,7 @@
 Provides functions for probing / NER that can be called with model name and the required data.
 Makes evaluating models easier.
 """
-from src.models.llms_interface import available_models as llms
+from src.models.LLM_interface import available_models as llms
 from src.models.GLiNER import find_entities as find_entities_gliner
 from src.models.GLiNER import classify_entity as classify_entity_gliner
 from src.models.T5_NLI import classify_entity as classify_entity_t5_nli

src/experiments/NEC_evaluation/evaluation.py

@@ -27,7 +27,7 @@ def run_NEC_tests(model_name, dataset, results_dir, test_instances=10):
     labels = data_manager.get_labels(dataset)
     data = data_manager.get_annotated_sentences(dataset, test_instances)
 
-    if (model_name == "Word2Vec" or model_name == "T5-MLM-entity"):
+    if model_name == "Word2Vec" or model_name == "T5-MLM-entity":
         label_dict = data_manager.get_label_dict(dataset, test_instances)
         print(label_dict)
         set_label_dict(model_name, label_dict)
@@ -58,7 +58,7 @@ def run_NEC_tests(model_name, dataset, results_dir, test_instances=10):
 
 
 def run_NEC_tests_all():
-    models = ["GLiNER", "Llama-3.1-8B", "T5-NLI", "T5-MLM-label", "T5-MLM-entity", "Word2Vec", "DeepSeek-R1-Distill-Qwen-32B"]
+    models = ["DeepSeek-R1-Distill-Qwen-32B"]  # "GLiNER", "Llama-3.1-8B", "T5-NLI", "T5-MLM-label", "T5-MLM-entity", "Word2Vec"
     datasets = ["CoNLL", "FIGER-coarse", "FIGER-fine"]  # "Pile-NER-type"]
     for model in models:
         for dataset in datasets:
@@ -97,5 +97,5 @@ def read_NEC_metrics(directory):
         print(f"Model: {model}, Dataset: {dataset}, Accuracy: {avg_accuracy:.2f}%")
 
 
-run_NEC_tests_all()
+#run_NEC_tests_all()
 read_NEC_metrics("results")