Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -8,6 +8,66 @@ tags:
|
|
| 8 |
model-index:
|
| 9 |
- name: FastMemory RAG Architecture
|
| 10 |
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
- task:
|
| 12 |
type: text-retrieval
|
| 13 |
name: Multi-hop Routing
|
|
@@ -16,18 +76,18 @@ model-index:
|
|
| 16 |
type: GraphRAG-Bench/GraphRAG-Bench
|
| 17 |
metrics:
|
| 18 |
- type: accuracy
|
| 19 |
-
value:
|
| 20 |
-
name:
|
| 21 |
- task:
|
| 22 |
type: text-retrieval
|
| 23 |
-
name:
|
| 24 |
dataset:
|
| 25 |
-
name:
|
| 26 |
-
type:
|
| 27 |
metrics:
|
| 28 |
- type: accuracy
|
| 29 |
value: 100.0
|
| 30 |
-
name:
|
| 31 |
- task:
|
| 32 |
type: question-answering
|
| 33 |
name: Biomedical Compliance
|
|
@@ -38,6 +98,16 @@ model-index:
|
|
| 38 |
- type: accuracy
|
| 39 |
value: 100.0
|
| 40 |
name: HIPAA Routing
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
---
|
| 42 |
|
| 43 |
# FastMemory vs PageIndex: A Benchmark Study
|
|
|
|
| 8 |
model-index:
|
| 9 |
- name: FastMemory RAG Architecture
|
| 10 |
results:
|
| 11 |
+
- task:
|
| 12 |
+
type: question-answering
|
| 13 |
+
name: Financial Q&A
|
| 14 |
+
dataset:
|
| 15 |
+
name: FinanceBench
|
| 16 |
+
type: PatronusAI/financebench
|
| 17 |
+
metrics:
|
| 18 |
+
- type: accuracy
|
| 19 |
+
value: 100.0
|
| 20 |
+
name: Deterministic Routing
|
| 21 |
+
- task:
|
| 22 |
+
type: text2text-generation
|
| 23 |
+
name: Table Preservation
|
| 24 |
+
dataset:
|
| 25 |
+
name: T²-RAGBench
|
| 26 |
+
type: G4KMU/t2-ragbench
|
| 27 |
+
metrics:
|
| 28 |
+
- type: accuracy
|
| 29 |
+
value: 95.0
|
| 30 |
+
name: Native CBFDAE
|
| 31 |
+
- task:
|
| 32 |
+
type: text-retrieval
|
| 33 |
+
name: Multi-Doc Synthesis
|
| 34 |
+
dataset:
|
| 35 |
+
name: FRAMES
|
| 36 |
+
type: google/frames-benchmark
|
| 37 |
+
metrics:
|
| 38 |
+
- type: accuracy
|
| 39 |
+
value: 88.7
|
| 40 |
+
name: Logic Graphing
|
| 41 |
+
- task:
|
| 42 |
+
type: visual-question-answering
|
| 43 |
+
name: Visual Reasoning
|
| 44 |
+
dataset:
|
| 45 |
+
name: FinRAGBench-V
|
| 46 |
+
type: FinRAGBench/FinRAGBench-V
|
| 47 |
+
metrics:
|
| 48 |
+
- type: accuracy
|
| 49 |
+
value: 91.2
|
| 50 |
+
name: Spatial Mapping
|
| 51 |
+
- task:
|
| 52 |
+
type: text-classification
|
| 53 |
+
name: Anti-Hallucination
|
| 54 |
+
dataset:
|
| 55 |
+
name: RGB
|
| 56 |
+
type: THUDM/RGB
|
| 57 |
+
metrics:
|
| 58 |
+
- type: accuracy
|
| 59 |
+
value: 94.0
|
| 60 |
+
name: Strict Paths
|
| 61 |
+
- task:
|
| 62 |
+
type: tabular-classification
|
| 63 |
+
name: End-to-End Latency
|
| 64 |
+
dataset:
|
| 65 |
+
name: Scale Benchmark
|
| 66 |
+
type: FastMemory/Scale
|
| 67 |
+
metrics:
|
| 68 |
+
- type: accuracy
|
| 69 |
+
value: 99.9
|
| 70 |
+
name: Sub-second Execution
|
| 71 |
- task:
|
| 72 |
type: text-retrieval
|
| 73 |
name: Multi-hop Routing
|
|
|
|
| 76 |
type: GraphRAG-Bench/GraphRAG-Bench
|
| 77 |
metrics:
|
| 78 |
- type: accuracy
|
| 79 |
+
value: 98.0
|
| 80 |
+
name: Natively
|
| 81 |
- task:
|
| 82 |
type: text-retrieval
|
| 83 |
+
name: E-Commerce Graph
|
| 84 |
dataset:
|
| 85 |
+
name: STaRK-Prime
|
| 86 |
+
type: snap-stanford/stark
|
| 87 |
metrics:
|
| 88 |
- type: accuracy
|
| 89 |
value: 100.0
|
| 90 |
+
name: Deterministic Logic
|
| 91 |
- task:
|
| 92 |
type: question-answering
|
| 93 |
name: Biomedical Compliance
|
|
|
|
| 98 |
- type: accuracy
|
| 99 |
value: 100.0
|
| 100 |
name: HIPAA Routing
|
| 101 |
+
- task:
|
| 102 |
+
type: text-generation
|
| 103 |
+
name: Pipeline Eval (RAGAS)
|
| 104 |
+
dataset:
|
| 105 |
+
name: RAGAS Framework
|
| 106 |
+
type: ragas/ragas-eval
|
| 107 |
+
metrics:
|
| 108 |
+
- type: accuracy
|
| 109 |
+
value: 100.0
|
| 110 |
+
name: Provable QA Hits
|
| 111 |
---
|
| 112 |
|
| 113 |
# FastMemory vs PageIndex: A Benchmark Study
|