Update app.py

app.py

@@ -1,7 +1,7 @@
 import gradio as gr
 import spaces
 import torch
-from transformers import T5Tokenizer, T5ForConditionalGeneration, AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
+from transformers import T5Tokenizer, T5ForConditionalGeneration, AutoTokenizer, AutoModelForSeq2SeqLM, AutoModelForCausalLM, pipeline
 import languagecodes
 
 favourite_langs = {"German": "de", "Romanian": "ro", "English": "en", "-----": "-----"}
@@ -14,8 +14,10 @@ models = ["Helsinki-NLP", "t5-base", "t5-small", "t5-large",
           "facebook/nllb-200-distilled-600M",
           "facebook/nllb-200-distilled-1.3B",
           "facebook/mbart-large-50-many-to-many-mmt",
+          "utter-project/EuroLLM-1.7B",
           "Unbabel/TowerInstruct-7B-v0.2",
-          "Unbabel/TowerInstruct-Mistral-7B-v0.2"]
+          "Unbabel/TowerInstruct-Mistral-7B-v0.2"
+          ]
 
 def model_to_cuda(model):
     # Move the model to GPU if available
@@ -26,6 +28,17 @@ def model_to_cuda(model):
         print("CUDA not available! Using CPU.")
     return model
 
+def eurollm(model_id, sl, tl, input_text):
+    model_id = "utter-project/EuroLLM-1.7B"
+    tokenizer = AutoTokenizer.from_pretrained(model_id)
+    model = AutoModelForCausalLM.from_pretrained(model_id)  
+    prompt = f"{sl}: {input_text}. {tl}:"
+    inputs = tokenizer(prompt, return_tensors="pt")
+    outputs = model.generate(**inputs, max_new_tokens=20)
+    output = tokenizer.decode(outputs[0], skip_special_tokens=True) 
+    print(output)
+    return output
+
 @spaces.GPU
 def translate_text(input_text, sselected_language, tselected_language, model_name):
     sl = all_langs[sselected_language]
@@ -43,7 +56,10 @@ def translate_text(input_text, sselected_language, tselected_language, model_nam
                 model = model_to_cuda(AutoModelForSeq2SeqLM.from_pretrained(model_name))
             except EnvironmentError as error:
                 return f"Error finding model: {model_name}! Try other available language combination.", error
-    
+    if 'eurollm' in model_name:
+        translated_text = eurollm(model_name, sl, tl, input_text)
+        return translated_text[0]['translation_text'], message_text
+        
     if 'nllb' in model_name:
         tokenizer = AutoTokenizer.from_pretrained(model_name, src_lang=languagecodes.nllb_language_codes[sselected_language])
         model = AutoModelForSeq2SeqLM.from_pretrained(model_name, device_map="auto")