|
--- |
|
license: apache-2.0 |
|
inference: false |
|
--- |
|
|
|
# DRAGON-YI-9B-GGUF |
|
|
|
<!-- Provide a quick summary of what the model is/does. --> |
|
|
|
**dragon-yi-9b-gguf** is a fact-based question-answering model, optimized for complex business documents, finetuned on top of 01-ai/yi-v1.5-9b base and quantizedwith 4_K_M GGUF quantization, providing an inference implementation for use on CPUs. |
|
|
|
|
|
## Benchmark Tests |
|
|
|
Evaluated against the benchmark test: RAG-Instruct-Benchmark-Tester |
|
1 Test Run (temperature=0.0, sample=False) with 1 point for correct answer, 0.5 point for partial correct or blank / NF, 0.0 points for incorrect, and -1 points for hallucinations. |
|
|
|
--Accuracy Score: **98.0** correct out of 100 |
|
--Not Found Classification: 90.0% |
|
--Boolean: 97.5% |
|
--Math/Logic: 95% |
|
--Complex Questions (1-5): 5 (Very Strong) |
|
--Summarization Quality (1-5): 4 (Above Average) |
|
--Hallucinations: No hallucinations observed in test runs. |
|
|
|
For test run results (and good indicator of target use cases), please see the files ("core_rag_test" and "answer_sheet" in this repo). |
|
|
|
|
|
To pull the model via API: |
|
|
|
from huggingface_hub import snapshot_download |
|
snapshot_download("llmware/dragon-yi-9b-gguf", local_dir="/path/on/your/machine/", local_dir_use_symlinks=False) |
|
|
|
|
|
Load in your favorite GGUF inference engine, or try with llmware as follows: |
|
|
|
from llmware.models import ModelCatalog |
|
model = ModelCatalog().load_model("dragon-yi-9b-gguf") |
|
response = model.inference(query, add_context=text_sample) |
|
|
|
Note: please review [**config.json**](https://huggingface.co/llmware/dragon-yi-9b-gguf/blob/main/config.json) in the repository for prompt wrapping information, details on the model, and full test set. |
|
|
|
|
|
### Model Description |
|
|
|
<!-- Provide a longer summary of what this model is. --> |
|
|
|
- **Developed by:** llmware |
|
- **Model type:** GGUF |
|
- **Language(s) (NLP):** English |
|
- **License:** Apache 2.0 |
|
|
|
|
|
## Model Card Contact |
|
|
|
Darren Oberst & llmware team |
