Update app.py

app.py

@@ -32,16 +32,17 @@ import mysql.connector
 from mysql.connector import errorcode, pooling
 from dotenv import load_dotenv
 from huggingface_hub import login
+from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
 
 model_name = "openlm-research/open_llama_3b_v2"
 tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False, legacy=False)
-model = OpenLlamaForCausalLM.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name)  # Use AutoModelForCausalLM
 
 # Determine the maximum supported length for the model
-max_supported_length = 8000  # You might need to adjust this
+max_supported_length = 2048  # You might need to adjust this
 
 openllama_pipeline = pipeline(
-    "GenerationMixin",
+    "text-generation",  # Use "text-generation" 
     model=model,
     tokenizer=tokenizer,
     truncation=True,
@@ -49,19 +50,6 @@ openllama_pipeline = pipeline(
     # ... other parameters
 )
 
-model_name = "openlm-research/open_llama_3b_v2"  # Or another OpenLlama variant
-tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False, legacy=False)
-model = OpenLlamaForCausalLM.from_pretrained(model_name)
-
-openllama_pipeline = pipeline(
-    "GenerationMixin",
-    model=model,
-    tokenizer=tokenizer,
-    truncation=True,
-    max_length=max_tokens,  # Assuming max_tokens is your max_length variable
-    # ... other parameters
-)
-
 nlp = AutoTokenizer.from_pretrained("bert-base-uncased")
 
 HUGGINGFACE_TOKEN = os.getenv("HUGGINGFACE_TOKEN")