🌘w🌖

app.py

@@ -1,7 +1,6 @@
 import gradio as gr
-from datasets import load_dataset, Dataset
+from datasets import load_dataset
 
-# import faiss
 import os
 import spaces
 from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
@@ -31,6 +30,7 @@ data = load_dataset("not-lain/wikipedia-small-3000-embedded", subset="train")
 # index dataset
 data.add_faiss_index("embedding", device=1)
 
+
 @spaces.GPU
 def search(query: str, k: int = TOP_K):
     embedded_query = model.encode(query)
@@ -68,7 +68,9 @@ def talk(message, history):
             cleaned_past = item[1].split("\nRESOURCES:\n")[0]
             chat.append({"role": "assistant", "content": cleaned_past})
     chat.append({"role": "user", "content": message})
-    messages = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True)
+    messages = tokenizer.apply_chat_template(
+        chat, tokenize=False, add_generation_prompt=True
+    )
     # Tokenize the messages string
     model_inputs = tokenizer([messages], return_tensors="pt").to(device)
     streamer = TextIteratorStreamer(
@@ -100,12 +102,14 @@ TITLE = "RAG"
 
 DESCRIPTION = """
 ## Resources used to build this project
-* https://huggingface.co/mixedbread-ai/mxbai-colbert-large-v1
-* me 😎
-## Models
-the models used in this space are : 
-* google/gemma-7b-it
-* mixedbread-ai/mxbai-colbert-v1
+* embedding model : https://huggingface.co/mixedbread-ai/mxbai-colbert-large-v1
+* dataset : https://huggingface.co/datasets/not-lain/wikipedia-small-3000-embedded (used mxbai-colbert-large-v1 to create the embedding column )
+* faiss docs : https://huggingface.co/docs/datasets/v2.18.0/en/package_reference/main_classes#datasets.Dataset.add_faiss_index 
+* chatbot : google/gemma-7b-it
+
+If you want to support my work please click on the heart react button ❤️🤗
+
+<sub><sup><sub><sup>psst, I am still open for work if please reach me out at https://not-lain.github.io/</sup></sub></sup></sub>
 """
 
 demo = gr.ChatInterface(