File size: 2,729 Bytes
4099389 90c8efc 4099389 22fc030 4099389 22fc030 4099389 22fc030 4099389 22fc030 4099389 22fc030 4099389 22fc030 4099389 22fc030 4099389 22fc030 14c88b7 4099389 4cd16ae 4099389 4cd16ae 4099389 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 |
---
license: cc-by-nc-sa-4.0
---
# Triplex: a SOTA LLM for knowledge graph construction.
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.
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.
## Benchmark
## Usage:
- **Blog:** [https://www.sciphi.ai/blog/triplex](https://www.sciphi.ai/blog/triplex)
- **Demo:** [kg.sciphi.ai](kg.sciphi.ai)
- **Cookbook:** [https://r2r-docs.sciphi.ai/cookbooks/knowledge-graph](https://r2r-docs.sciphi.ai/cookbooks/knowledge-graph)
- **Python:**
```python
import json
from transformers import AutoModelForCausalLM, AutoTokenizer
def triplextract(model, tokenizer, text, entity_types, predicates):
input_format = """
**Entity Types:**
{entity_types}
**Predicates:**
{predicates}
**Text:**
{text}
"""
message = input_format.format(
entity_types = json.dumps({"entity_types": entity_types}),
predicates = json.dumps({"predicates": predicates}),
text = text)
messages = [{'role': 'user', 'content': message}]
input_ids = tokenizer.apply_chat_template(messages, add_generation_prompt = True, return_tensors="pt").to("cuda")
output = tokenizer.decode(model.generate(input_ids=input_ids, max_length=2048)[0], skip_special_tokens=True)
return output
model = AutoModelForCausalLM.from_pretrained("sciphi/triplex", trust_remote_code=True).to('cuda').eval()
tokenizer = AutoTokenizer.from_pretrained("sciphi/triplex", trust_remote_code=True)
entity_types = [ "LOCATION", "POSITION", "DATE", "CITY", "COUNTRY", "NUMBER" ]
predicates = [ "POPULATION", "AREA" ]
text = """
San Francisco,[24] officially the City and County of San Francisco, is a commercial, financial, and cultural center in Northern California.
With a population of 808,437 residents as of 2022, San Francisco is the fourth most populous city in the U.S. state of California behind Los Angeles, San Diego, and San Jose.
"""
prediction = triplextract(model, tokenizer, text, entity_types, predicates)
print(prediction)
``` |