test metrics for english

app.py

@@ -1,3 +1,6 @@
+import json
+from pathlib import Path
+
 import pandas as pd
 import streamlit as st
 
@@ -10,6 +13,12 @@ model_label = {en: "Select model", ru: "Выберете модель"}
 title_label = {en: "Title", ru: "Название статьи"}
 authors_label = {en: "Author(s)", ru: "Автор(ы)"}
 abstract_label = {en: "Abstract", ru: "Аннотация"}
+metrics_label = {en: "Test metrics", ru: "Метрики на тренировочном датасете"}
+
+with open(
+    Path(__file__).parent / "category_classification" / "test_results.json", "r"
+) as metric_f:
+    metrics = json.load(metric_f)
 
 
 def text_area_height(line_height: int):
@@ -21,6 +30,7 @@ def load_class_model(name):
     model = class_models.get_model(name)
     return model
 
+
 lang = st.pills(label=langs_str, options=langs)
 if lang is None:
     lang = en
@@ -38,3 +48,8 @@ if title:
     results = model(input)
     results = process_results(results, lang)
     st.dataframe(results, hide_index=True)
+
+lang_metrics = pd.DataFrame(metrics[lang])
+if not lang_metrics.empty:
+    with st.expander(metrics_label[lang]):
+        st.dataframe(lang_metrics)

category_classification/test_results.json

@@ -0,0 +1,20 @@
+{
+    "en": {
+        "Accuracy": {
+            "HibiscusMaximus.scibert_paper_classification": 0.8396666666666667,
+            "oracat.bert-paper-classifier-arxiv": 0.21833333333333332
+        },
+        "Macro F1": {
+            "HibiscusMaximus.scibert_paper_classification": 0.5253199699435153,
+            "oracat.bert-paper-classifier-arxiv": 0.06976813826285358
+        },
+        "Cross-entropy loss": {
+            "HibiscusMaximus.scibert_paper_classification": 356.9575782795263,
+            "oracat.bert-paper-classifier-arxiv": 726.6457145760214
+        }
+    },
+    "ru": {
+        "F1": {},
+        "Macro F1": {}
+    }
+}