Franc v0.9

.idea/.gitignore

@@ -0,0 +1,8 @@
+# Default ignored files
+/shelf/
+/workspace.xml
+# Editor-based HTTP Client requests
+/httpRequests/
+# Datasource local storage ignored files
+/dataSources/
+/dataSources.local.xml

.idea/franc-v0.9.iml

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="PYTHON_MODULE" version="4">
+  <component name="NewModuleRootManager">
+    <content url="file://$MODULE_DIR$" />
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+</module>

.idea/inspectionProfiles/profiles_settings.xml

@@ -0,0 +1,6 @@
+<component name="InspectionProjectProfileManager">
+  <settings>
+    <option name="USE_PROJECT_PROFILE" value="false" />
+    <version value="1.0" />
+  </settings>
+</component>

.idea/misc.xml

@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.11" project-jdk-type="Python SDK" />
+</project>

.idea/modules.xml

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectModuleManager">
+    <modules>
+      <module fileurl="file://$PROJECT_DIR$/.idea/franc-v0.9.iml" filepath="$PROJECT_DIR$/.idea/franc-v0.9.iml" />
+    </modules>
+  </component>
+</project>

.idea/vcs.xml

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="VcsDirectoryMappings">
+    <mapping directory="" vcs="Git" />
+  </component>
+</project>

main.py

@@ -0,0 +1,115 @@
+import os
+
+import gradio as gr
+import pinecone
+from langchain import PromptTemplate
+from langchain.chains import RetrievalQA
+from langchain.embeddings.huggingface import HuggingFaceEmbeddings
+from langchain.llms import HuggingFaceEndpoint
+from langchain.vectorstores import Pinecone
+from torch import cuda
+
+LLAMA_2_7B_CHAT_HF_FRANC_V0_9 = os.environ.get("LLAMA_2_7B_CHAT_HF_FRANC_V0_9")
+HUGGING_FACE_HUB_TOKEN = os.environ.get("HUGGING_FACE_HUB_TOKEN")
+PINECONE_API_KEY = os.environ.get('PINECONE_API_KEY')
+PINECONE_ENVIRONMENT = os.environ.get('PINECONE_ENVIRONMENT')
+
+# Set up Pinecone vector store
+
+pinecone.init(
+    api_key=PINECONE_API_KEY,
+    environment=PINECONE_ENVIRONMENT
+)
+
+index_name = 'stadion-6237'
+index = pinecone.Index(index_name)
+embedding_model_id = 'sentence-transformers/paraphrase-mpnet-base-v2'
+device = f'cuda:{cuda.current_device()}' if cuda.is_available() else 'cpu'
+embedding_model = HuggingFaceEmbeddings(
+    model_name=embedding_model_id,
+    model_kwargs={'device': device},
+    encode_kwargs={'device': device, 'batch_size': 32}
+)
+text_key = 'text'
+
+vector_store = Pinecone(
+    index, embedding_model.embed_query, text_key
+)
+
+B_INST, E_INST = "[INST] ", " [/INST]"
+B_SYS, E_SYS = "<<SYS>>\n", "\n<</SYS>>\n\n"
+
+
+def get_prompt_template(instruction, system_prompt):
+    system_prompt = B_SYS + system_prompt + E_SYS
+    prompt_template = B_INST + system_prompt + instruction + E_INST
+    return prompt_template
+
+
+template = get_prompt_template(
+    """Use the following pieces of context to answer the question at the end.
+
+{context}
+
+Question: {question}
+Helpful Answer:""",
+    """Your name is Franc.
+You are a running coach and exercise physiologist.
+You communicate in the style of Hal Higdon.
+Your answers are always 512-character long or less.
+If you don't know the answer to a question, please don't share false information."""
+)
+
+endpoint_url = (
+    LLAMA_2_7B_CHAT_HF_FRANC_V0_9
+)
+
+llm = HuggingFaceEndpoint(
+    endpoint_url=endpoint_url,
+    huggingfacehub_api_token=HUGGING_FACE_HUB_TOKEN,
+    task="text-generation",
+    model_kwargs={
+        "max_new_tokens": 512,
+        "temperature": 0.3,
+        "repetition_penalty": 1.2,
+        "return_full_text": True,
+    }
+)
+
+rag_chain = RetrievalQA.from_chain_type(
+    llm=llm,
+    chain_type='stuff',
+    retriever=vector_store.as_retriever(),
+    chain_type_kwargs={
+        "prompt": PromptTemplate(
+            template=template,
+            input_variables=["context", "question"],
+        ),
+    },
+)
+
+
+def generate(message):
+    return rag_chain(message)
+
+
+gr.ChatInterface(
+    generate,
+    title="Franc v0.1",
+    description="Meet Franc, a running coach who has become the go-to expert for runners, whether they are just "
+                "lacing up their shoes for the first time or seasoned marathoners. With his deep understanding of "
+                "exercise physiology, Franc answers all sorts of questions, guiding individuals on their unique "
+                "running journeys. But don't be fooled – even with his vast knowledge, Franc has been known to make "
+                "mistakes, and his advice should always be taken with a grain of salt. What sets him apart from other "
+                "coaches, however, is his approach, dedication, and a surprising revelation: Franc is actually a "
+                "fine-tuned LLM (Language Model), blending technology with human-like empathy to help his runners "
+                "achieve their goals.",
+    examples=[
+        "What's VO2max?"
+    ],
+    cache_examples=True,
+    theme=gr.themes.Soft(),
+    submit_btn="Ask Franc",
+    retry_btn="Do better, Franc!",
+    autofocus=True,
+).queue().launch()

requirements.txt

@@ -0,0 +1,8 @@
+gradio
+huggingface-hub
+langchain
+pinecone-client
+sentence-transformers
+text-generation
+torch
+transformers