check

app.py

@@ -1,68 +1,79 @@
 from gradio_client import Client
 import gradio as gr
-import requests
-from langchain.chains import RetrievalQA
-import pinecone
-from langchain.vectorstores import Pinecone
-import os
-import openai
-import time
-from langchain.embeddings.huggingface import HuggingFaceEmbeddings
-import transformers
-from langchain.chains import RetrievalQA
+# import requests
+# from langchain.chains import RetrievalQA
+# import pinecone
+# from langchain.vectorstores import Pinecone
+# import os
+# import openai
+# import time
+# from langchain.embeddings.huggingface import HuggingFaceEmbeddings
+# import transformers
+# from langchain.chains import RetrievalQA
 
-API_URL = "https://api-inference.huggingface.co/models/HuggingFaceH4/zephyr-7b-beta"
-headers = {"Authorization": f"Bearer {os.environ.get('API_KEY')}"}
+# API_URL = "https://api-inference.huggingface.co/models/HuggingFaceH4/zephyr-7b-beta"
+# headers = {"Authorization": f"Bearer {os.environ.get('API_KEY')}"}
 retrieval = Client("https://ishaan-mital-ncert-helper-vector-db.hf.space/--replicas/7f5fz9pvt/")
+llm = Client("https://library-samples-zephyr-7b.hf.space/--replicas/b7p4f/")
+# embed_model_id = 'sentence-transformers/all-MiniLM-L6-v2'
 
-embed_model_id = 'sentence-transformers/all-MiniLM-L6-v2'
+# embed_model = HuggingFaceEmbeddings(
+#     model_name=embed_model_id,
+# )
 
-embed_model = HuggingFaceEmbeddings(
-    model_name=embed_model_id,
-)
 
+# pinecone.init(
+#     api_key=os.environ.get('PINECONE_API_KEY'),
+#     environment=os.environ.get('PINECONE_ENVIRONMENT')
+# )
 
-pinecone.init(
-    api_key=os.environ.get('PINECONE_API_KEY'),
-    environment=os.environ.get('PINECONE_ENVIRONMENT')
-)
+# index_name='llama-rag'
+# index = pinecone.Index(index_name)
+# index.describe_index_stats()
+# text_field = 'text'
 
-index_name='llama-rag'
-index = pinecone.Index(index_name)
-index.describe_index_stats()
-text_field = 'text'
+# vectorstore = Pinecone(
+#     index, embed_model.embed_query, text_field
+# )
 
-vectorstore = Pinecone(
-    index, embed_model.embed_query, text_field
-)
 
-
-headers = {"Authorization": "Bearer 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"}
-generate_text = transformers.pipeline(
-    model="HuggingFaceH4/zephyr-7b-beta",
-    return_full_text=True,  # langchain expects the full text
-    task='text-generation',
-    # we pass model parameters here too
-    temperature=0.7,  # 'randomness' of outputs, 0.0 is the min and 1.0 the max
-    max_new_tokens=512,  # mex number of tokens to generate in the output
-    repetition_penalty=1.1,  # without this output begins repeating
-    do_sample=True
-)
+# headers = {"Authorization": "Bearer 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"}
+# generate_text = transformers.pipeline(
+#     model="HuggingFaceH4/zephyr-7b-beta",
+#     return_full_text=True,  # langchain expects the full text
+#     task='text-generation',
+#     # we pass model parameters here too
+#     temperature=0.7,  # 'randomness' of outputs, 0.0 is the min and 1.0 the max
+#     max_new_tokens=512,  # mex number of tokens to generate in the output
+#     repetition_penalty=1.1,  # without this output begins repeating
+#     do_sample=True
+# )
 	
-from langchain.llms import HuggingFacePipeline
+# from langchain.llms import HuggingFacePipeline
 
-llm = HuggingFacePipeline(pipeline=generate_text)
+# llm = HuggingFacePipeline(pipeline=generate_text)
 
-rag_pipeline = RetrievalQA.from_chain_type(
-    llm=llm, chain_type='stuff',
-    retriever=vectorstore.as_retriever()
-)
+# rag_pipeline = RetrievalQA.from_chain_type(
+#     llm=llm, chain_type='stuff',
+#     retriever=vectorstore.as_retriever()
+# )
 
 def main(question):
     # return rag_pipeline(question)
     global chatbot
-    chatbot = requests.post(API_URL, headers=headers, json={"inputs": question}).json()
-    return rag_pipeline(question)
+    context = retrieval.predict(question)
+    answer = llm.predict(
+        f'Question: {question} and context: {context}', 
+            "NCERT Helper!!",	# str  in 'System prompt' Textbox component
+            2048,	# float (numeric value between 1 and 2048) in 'Max new tokens' Slider component
+            0.1,	# float (numeric value between 0.1 and 4.0) in 'Temperature' Slider component
+            0.05,	# float (numeric value between 0.05 and 1.0) in 'Top-p (nucleus sampling)' Slider component
+            3,	# float (numeric value between 1 and 1000) in 'Top-k' Slider component
+            1,	# float (numeric value between 1.0 and 2.0) in 'Repetition penalty' Slider component
+            api_name="/chat"
+    )
+    chatbot = answer
+    return answer
 
 demo = gr.Interface(main, inputs = "text", outputs = "text")
 

requirements.txt

@@ -2,10 +2,10 @@ hugchat
 gradio
 gradio_client
 gtts
-openai==0.28
-pydantic==1.10.9
-langchain
-pinecone-client==2.2.2
-faiss-cpu
-sentence_transformers
-transformers
+# openai==0.28
+# pydantic==1.10.9
+# langchain
+# pinecone-client==2.2.2
+# faiss-cpu
+# sentence_transformers
+# transformers