Jatin Mehra commited on
Commit
e9800a3
·
1 Parent(s): 2f7e733

Add detailed RAG system metrics section to README(HF_SPACES)

Browse files
Files changed (1) hide show
  1. README_hf.md +33 -0
README_hf.md CHANGED
@@ -16,6 +16,7 @@ An advanced PDF document analysis tool that combines RAG (Retrieval Augmented Ge
16
 
17
  - [Overview](#overview)
18
  - [Features](#features)
 
19
  - [Architecture](#architecture)
20
  - [Technical Stack](#technical-stack)
21
  - [Installation](#installation)
@@ -44,6 +45,38 @@ The application employs an agentic approach that can augment the document's info
44
  - **Hugging Face Integration**: Automatic deployment to Hugging Face Spaces
45
  - **Android Application**: Native mobile client
46
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
47
  ## Architecture
48
 
49
  The application follows a modular architecture with these main components:
 
16
 
17
  - [Overview](#overview)
18
  - [Features](#features)
19
+ - [RAG System Metrics](#rag-system-metrics)
20
  - [Architecture](#architecture)
21
  - [Technical Stack](#technical-stack)
22
  - [Installation](#installation)
 
45
  - **Hugging Face Integration**: Automatic deployment to Hugging Face Spaces
46
  - **Android Application**: Native mobile client
47
 
48
+ ## RAG System Metrics
49
+
50
+ 1. **Key Metrics Overview**:
51
+
52
+ | Metric | Value |
53
+ | ----------------------------------- | ------- |
54
+ | **Semantic Similarity (Mean)** | `0.852` |
55
+ | **ROUGE-L F1 Score (Mean)** | `0.395` |
56
+ | **Semantic Similarity (Max)** | `1.000` |
57
+ | **ROUGE-L F1 Score (Max)** | `1.000` |
58
+ | **Semantic Similarity (Min)** | `0.592` |
59
+ | **ROUGE-L F1 Score (Min)** | `0.099` |
60
+ | **Standard Deviation (Similarity)** | `0.089` |
61
+ | **Standard Deviation (ROUGE-L F1)** | `0.217` |
62
+
63
+
64
+ 2. **Quantile Distribution**:
65
+
66
+ | Percentile | Semantic Similarity | ROUGE-L F1 Score |
67
+ | ---------- | ------------------- | ---------------- |
68
+ | **25%** | `0.7946` | `0.2516` |
69
+ | **50%** | `0.8732` | `0.3256` |
70
+ | **75%** | `0.9181` | `0.4951` |
71
+
72
+
73
+ 3. **Evaluation Status**:
74
+
75
+ | Status | Count | Percentage |
76
+ | ------ | ----- | ---------- |
77
+ | PASS | `64` | `85.3%` |
78
+ | FAIL | `11` | `14.7%` |
79
+
80
  ## Architecture
81
 
82
  The application follows a modular architecture with these main components: