Update app.py

app.py

@@ -2,7 +2,10 @@ import streamlit as st
 import sacrebleu
 from bert_score import score as bert_score
 import jieba
-import traceback
+import torch
+
+# Ensure CUDA is not used to avoid potential issues on Hugging Face Spaces
+torch.cuda.is_available = lambda : False
 
 def calculate_bleu(translations, references):
     return sacrebleu.corpus_bleu(translations, [references]).score
@@ -14,7 +17,7 @@ def calculate_chrf(translations, references):
     return sacrebleu.corpus_chrf(translations, [references]).score
 
 def calculate_bertscore(translations, references, lang):
-    P, R, F1 = bert_score(translations, references, lang=lang)
+    P, R, F1 = bert_score(translations, references, lang=lang, device='cpu')
     return F1.mean().item()
 
 def tokenize_chinese(text):
@@ -25,18 +28,15 @@ st.write("Input the translated text and the reference translation to compute BLE
 
 languages = {
     "English": "en", "Chinese": "zh", "French": "fr", "German": "de", "Spanish": "es",
-    "Russian": "ru", "Japanese": "ja", "Korean": "ko", "Arabic": "ar", "Italian": "it",
-    "Dutch": "nl", "Portuguese": "pt", "Turkish": "tr", "Polish": "pl", "Czech": "cs",
-    "Swedish": "sv", "Danish": "da", "Finnish": "fi", "Greek": "el", "Hungarian": "hu",
-    "Indonesian": "id", "Norwegian": "no", "Romanian": "ro", "Thai": "th", "Vietnamese": "vi",
-    "Hebrew": "he", "Hindi": "hi", "Bengali": "bn", "Tamil": "ta", "Urdu": "ur", "Other": "other"
+    "Japanese": "ja", "Korean": "ko", "Arabic": "ar", "Italian": "it",
+    "Dutch": "nl", "Portuguese": "pt", "Russian": "ru"
 }
 
 source_lang = st.selectbox("Select Source Language", list(languages.keys()))
 target_lang = st.selectbox("Select Target Language", list(languages.keys()))
 
-source_lang_code = st.text_input("Enter Source Language Code (ISO 639-1):", value=languages[source_lang]) if source_lang == "Other" else languages[source_lang]
-target_lang_code = st.text_input("Enter Target Language Code (ISO 639-1):", value=languages[target_lang]) if target_lang == "Other" else languages[target_lang]
+source_lang_code = languages[source_lang]
+target_lang_code = languages[target_lang]
 
 translation_input = st.text_area("Translated Text", height=200)
 reference_input = st.text_area("Reference Translation", height=200)
@@ -56,25 +56,14 @@ if st.button("Evaluate"):
             else:
                 references = [reference_input.strip()]
 
-            st.write("Debug: Inputs processed")
-            st.write(f"Translation: {translations}")
-            st.write(f"Reference: {references}")
-
-            st.write("Debug: Calculating scores...")
+            st.write("Calculating scores...")
 
             bleu_score = calculate_bleu(translations, references)
-            st.write(f"Debug: BLEU score calculated: {bleu_score}")
-
             ter_score = calculate_ter(translations, references)
-            st.write(f"Debug: TER score calculated: {ter_score}")
-
             chrf_score = calculate_chrf(translations, references)
-            st.write(f"Debug: CHRF score calculated: {chrf_score}")
-
             bertscore = calculate_bertscore(translations, references, target_lang_code)
-            st.write(f"Debug: BERTScore calculated: {bertscore}")
 
-            st.write("Final Scores:")
+            st.write("Evaluation Scores:")
             st.write(f"**BLEU Score:** {bleu_score:.2f}")
             st.write(f"**TER Score:** {ter_score:.2f}")
             st.write(f"**CHRF Score:** {chrf_score:.2f}")
@@ -82,9 +71,5 @@ if st.button("Evaluate"):
 
         except Exception as e:
             st.error(f"An error occurred: {str(e)}")
-            st.write("Debug: Full traceback")
-            st.code(traceback.format_exc())
     else:
-        st.error("Please provide both translated text and reference translation.")
-
-st.write("Debug: End of script reached")
+        st.error("Please provide both translated text and reference translation.")