Update app.py

app.py

@@ -3,15 +3,20 @@ from textPreprocessing import text2prompt
 from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
 import torch
 
+"""
+Необходимо раскомментить при досутпе к GPU
+
 bnb_config = BitsAndBytesConfig(
     load_in_4bit=True,
     bnb_4bit_use_double_quant=True,
     bnb_4bit_quant_type="fp4",
     bnb_4bit_compute_dtype=torch.bfloat16
-)
+)"""
 
-model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-Instruct-v0.1",
-                                             quantization_config=bnb_config)
+model = AutoModelForCausalLM.from_pretrained(
+    "mistralai/Mistral-7B-Instruct-v0.1",
+    # quantization_config=bnb_config # Необходимо раскомментить при досутпе к GPU
+)
 tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.1")
 
 def predict(input_text, t, m):
@@ -25,7 +30,7 @@ def predict(input_text, t, m):
             - Instruct-based модель
     """
     prompt = text2prompt(input_text)
-    inputs = tokenizer(prompt, return_tensors="pt")
+    inputs = tokenizer(prompt, return_tensors="np")
     generate_ids = model.generate(inputs.input_ids, max_new_tokens=128)
     answer = tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
     return answer.replace(prompt, "")