Update app.py

app.py

@@ -8,15 +8,15 @@ access_token = os.getenv("read_access")
 from transformers import AutoModelForCausalLM, AutoTokenizer
 device = "cpu" # the device to load the model onto
 
-tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2-0.5B-Instruct")
+#tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2-0.5B-Instruct")
 
-model1 = AutoModelForCausalLM.from_pretrained(
-    "Qwen/Qwen2-1.5B-Instruct",
-    device_map="auto"
-)
+#model1 = AutoModelForCausalLM.from_pretrained(
+#    "Qwen/Qwen2-1.5B-Instruct",
+#    device_map="auto"
+#)
 
-tokenizer2 = AutoTokenizer.from_pretrained("google/gemma-2-2b-it", token=access_token)
-model2 = AutoModelForCausalLM.from_pretrained(
+tokenizer = AutoTokenizer.from_pretrained("google/gemma-2-2b-it", token=access_token)
+model = AutoModelForCausalLM.from_pretrained(
     "google/gemma-2-2b-it",
     device_map="auto",
     token=access_token
@@ -106,9 +106,9 @@ async def read_droot():
         tokenize=False,
         add_generation_prompt=True
     )
-    model_inputs = tokenizer2([text], return_tensors="pt").to(device)
+    model_inputs = tokenizer([text], return_tensors="pt").to(device)
         
-    generated_ids = model2.generate(
+    generated_ids = model.generate(
         model_inputs.input_ids,
         max_new_tokens=64
     )
@@ -116,11 +116,10 @@ async def read_droot():
         output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
     ]
         
-    response = tokenizer2.batch_decode(generated_ids, skip_special_tokens=True)[0]
-    respons = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
     print(response)
     end_time = time.time()
     time_taken = end_time - starttime 
     print(time_taken)
-    return {"Hello": respons}
+    return {"Hello": "resps"}
     #return {response: time}