let's add the tsv file

FACTS.tsv

@@ -0,0 +1,35 @@
+model	size	Separate Grounding Score	Separate Quality Score	Combined Score
+deepseek-ai/DeepSeek-R1-Distill-Qwen-14B	14	0.817797	0.542373	0.457627
+VIDraft/Gemma-3-R1984-27B	27	0.93617	0.459574	0.434043
+meta-llama/Llama-3.3-70B-Instruct	70	0.842553	0.510638	0.425532
+Qwen/Qwen3-30B-A3B	30	0.812766	0.540426	0.425532
+Qwen/Qwen3-4B	4	0.770213	0.540426	0.425532
+google/medgemma-27b-text-it	  27	0.940678	0.440678	0.419492
+Qwen/Qwen3-32B	32	0.740426	0.553191	0.417021
+deepseek-ai/DeepSeek-R1-0528-Qwen3-8B	8	0.834746	0.478814	0.411017
+deepseek-ai/DeepSeek-R1-Distill-Llama-8B	8	0.766949	0.516949	0.40678
+Qwen/Qwen3-8B	8	0.748936	0.523404	0.4
+Qwen/Qwen3-14B	14	0.778723	0.502128	0.382979
+google/gemma-3-27b-it	27	0.936	0.391	0.378
+google/medgemma-4b-it	4	0.834746	0.423729	0.377119
+Qwen/Qwen2.5-VL-32B-Instruct	32	0.621277	0.570213	0.357447
+meta-llama/Llama-3.1-70B-Instruct	70	0.855932	0.389831	0.334746
+google/gemma-3-12b-it	12	0.944	0.343	0.313
+google/gemma-3-4b-it	4	0.9	0.33	0.3
+Qwen/Qwen3-1.7B	1.7	0.702128	0.451064	0.297872
+deepseek-ai/DeepSeek-R1-Distill-Qwen-7B	7	0.59322	0.449153	0.275424
+Qwen/Qwen3-0.6B	0.6	0.682203	0.330508	0.266949
+Qwen/Qwen2.5-7B-Instruct	7	0.731915	0.310638	0.255319
+Qwen/Qwen2.5-14B-Instruct-1M	14	0.70339	0.300847	0.254237
+nvidia/Llama-Nemotron-Nano-8B	8	0.576271	0.402542	0.241525
+OpenScholar/Llama-3.1-OpenScholar-8B	8	0.690678	0.283898	0.241525
+Qwen/Qwen2.5-7B-Instruct-1M	7	0.737288	0.271186	0.207627
+nvidia/Llama-Nemotron-Nano-4B-v1.1	4	0.548936	0.340426	0.2
+google/gemma-3-1b-it	1	0.65	0.28	0.19
+mistralai/Ministral-8B-Instruct-2410	8	0.94	0.184	0.175
+meta-llama/Llama-3.1-8B-Instruct	8	0.665254	0.194915	0.169492
+mistralai/Mistral-Small-3.1-24B-Instruct-2503	24	0.953191	0.165957	0.157447
+mistralai/Mistral-Small-24B-Instruct-2501	24	0.95339	0.135593	0.131356
+open-thoughts/OpenThinker-7B	7	0.478814	0.152542	0.110169
+PleIAs/Pleias-RAG-350M	0.35	0.236264	0.021978	0.010989
+PleIAs/Pleias-RAG-1B	1	0.190476	0.037037	0

app.py

@@ -1,23 +1,16 @@
 import gradio as gr
 import pandas as pd
 import numpy as np
-import requests
 from io import StringIO
 import os
 
-dropbox_url = os.getenv("FACTS_RES_URL")
-
 try:
-    # Download the TSV file
-    response = requests.get(dropbox_url)
-    response.raise_for_status()  # Raise an exception for bad status codes
-
-    # Read the TSV data
-    df = pd.read_csv(StringIO(response.text), sep='\t')
-    print(f"Successfully loaded {len(df)} models from Dropbox")
+    # Read the local TSV file
+    df = pd.read_csv("FACTS.tsv", sep='\t')
+    print(f"Successfully loaded {len(df)} models from local file")
 except Exception as e:
-    print(f"Error loading data from Dropbox: {e}")
-    # Show sample data when download fails
+    print(f"Error loading data from local file: {e}")
+    # Show sample data when file reading fails
     df = pd.DataFrame({
         'model': [
             'deepseek-ai/DeepSeek-R1-Distill-Qwen-14B',
@@ -30,7 +23,7 @@ except Exception as e:
         'Separate Quality Score': [0.542373, 0.510638, 0.540426, 0.391],
         'Combined Score': [0.457627, 0.425532, 0.425532, 0.378]
     })
-    print("Showing sample data (download failed)")
+    print("Showing sample data (file read failed)")
 
 # Clean up the data
 df = df.dropna()  # Remove any rows with missing values