shreyaspimpalgaonkar
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -3,28 +3,28 @@ license: cc-by-nc-sa-4.0
|
|
3 |
|
4 |
---
|
5 |
|
6 |
-
# Triplex
|
7 |
|
8 |
-
|
9 |
|
10 |
-
Triplex is a
|
11 |
|
12 |
-
|
13 |
|
14 |
-
|
15 |
|
16 |
-
|
17 |
|
18 |
-
|
19 |
|
20 |
-
### Model Sources
|
21 |
|
22 |
- **Blog:** [https://www.sciphi.ai/blog/triplex](https://www.sciphi.ai/blog/triplex)
|
23 |
- **Demo:** [kg.sciphi.ai](kg.sciphi.ai)
|
24 |
-
- **
|
25 |
|
26 |
```python
|
27 |
|
|
|
28 |
import json
|
29 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
30 |
|
|
|
3 |
|
4 |
---
|
5 |
|
6 |
+
# Triplex: a SOTA LLM for knowledge graph construction.
|
7 |
|
8 |
+
Knowledge graphs, like Microsoft's Graph RAG, enhance RAG methods but are expensive to build. Triplex offers a 98% cost reduction for knowledge graph creation, outperforming GPT-4 at 1/60th the cost and enabling local graph building with SciPhi's R2R.
|
9 |
|
10 |
+
Triplex is a finetuned version of Phi3-3.8B for creating knowledge graphs from unstructured data developed by [SciPhi.AI](https://www.sciphi.ai). It works by extracting triplets - simple statements consisting of a subject, predicate, and object - from text or other data sources.
|
11 |
|
12 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/668d8d7a2413acbd544530d1/kcUC5FDEoziMSEcjVHQ3-.png)
|
13 |
|
14 |
+
## Benchmark
|
15 |
|
16 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/668d8d7a2413acbd544530d1/xsZ2UPZE5mnTFvgAsQwtl.png)
|
17 |
|
18 |
+
## Usage:
|
19 |
|
|
|
20 |
|
21 |
- **Blog:** [https://www.sciphi.ai/blog/triplex](https://www.sciphi.ai/blog/triplex)
|
22 |
- **Demo:** [kg.sciphi.ai](kg.sciphi.ai)
|
23 |
+
- **Cookbook:** [https://r2r-docs.sciphi.ai/cookbooks/knowledge-graph](https://r2r-docs.sciphi.ai/cookbooks/knowledge-graph)
|
24 |
|
25 |
```python
|
26 |
|
27 |
+
|
28 |
import json
|
29 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
30 |
|