Update app.py

app.py

@@ -45,24 +45,34 @@ hf_token = os.getenv("hf_token")
 login(token=hf_token)
 
 # Define the model pipeline with additional generation parameters
-model_pipeline = pipeline(
-   # model="meta-llama/Llama-3.2-1B",
-    model="dunzhang/stella_en_1.5B_v5",
-    #pad_token_id=50256,
-    #use_auth_token=hf_token,
-    #max_length=1000,  # You can increase this if needed
-    max_new_tokens=900  # Limit how many tokens are generated
-)
+#model_pipeline = pipeline(
+#   # model="meta-llama/Llama-3.2-1B",
+#    model="dunzhang/stella_en_1.5B_v5",
+#    #pad_token_id=50256,
+#    #use_auth_token=hf_token,
+#    #max_length=1000,  # You can increase this if needed
+#    max_new_tokens=900  # Limit how many tokens are generated
+#)
 
 # Use the pipeline in HuggingFacePipeline
-llm = HuggingFacePipeline(pipeline=model_pipeline)
+#llm = HuggingFacePipeline(pipeline=model_pipeline)
+###################
+
+
+llm = HuggingFacePipeline.from_model_id(
+    model_id="bigscience/bloom-1b7",
+    task="text-generation",
+    model_kwargs={"temperature": 0, "max_length": 64},
+)
+
 
 ##### Alternative
 from transformers import pipeline
 import torch
 from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
 
-#READER_MODEL_NAME = "HuggingFaceH4/zephyr-7b-beta"
+#READER_MODEL_NAME = "HuggingFaceH4/zephyr-7b-beta" #    model="dunzhang/stella_en_1.5B_v5",
+
 
 #bnb_config = BitsAndBytesConfig(
 #    load_in_4bit=True,