bling-phi-3.5-gguf

bling-phi-3.5-gguf is part of the BLING ("Best Little Instruct No-GPU") model series, RAG-instruct trained on top of a Microsoft Phi-3.5 base model, and 4_K_M quantized with GGUF for fast local inference.

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: 100 correct out of 100
--Not Found Classification: 85.0%
--Boolean: 95.0%
--Math/Logic: 90.0%
--Complex Questions (1-5): 4 (Above Average - multiple-choice, causal)
--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).

Please note that this is the model version used in the test results to replicate the most common inference environment (rather than the original Pytorch version).

Note: compare results with bling-phi-3-gguf and bling-phi-2.

Model Description

• Developed by: llmware
• Model type: bling
• Language(s) (NLP): English
• License: Apache 2.0
• Finetuned from model: Microsoft Phi-3.5

Uses

The intended use of BLING models is two-fold:

1. Provide high-quality RAG-Instruct models designed for fact-based, no "hallucination" question-answering in connection with an enterprise RAG workflow.

2. BLING models are fine-tuned on top of leading base foundation models, generally in the 1-3B+ range, and purposefully rolled-out across multiple base models to provide choices and "drop-in" replacements for RAG specific use cases.

Direct Use

BLING is designed for enterprise automation use cases, especially in knowledge-intensive industries, such as financial services, legal and regulatory industries with complex information sources.

BLING models have been trained for common RAG scenarios, specifically: question-answering, key-value extraction, and basic summarization as the core instruction types without the need for a lot of complex instruction verbiage - provide a text passage context, ask questions, and get clear fact-based responses.

Bias, Risks, and Limitations

Any model can provide inaccurate or incomplete information, and should be used in conjunction with appropriate safeguards and fact-checking mechanisms.

How to Get Started with the Model

To pull the model via API:

from huggingface_hub import snapshot_download           
snapshot_download("llmware/bling-phi-3.5-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  

# to load the model and make a basic inference
model = ModelCatalog().load_model("llmware/bling-phi-3.5-gguf", temperature=0.0, sample=False)
response = model.inference(query, add_context=text_sample)  

Details on the prompt wrapper and other configurations are on the config.json file in the files repository.

How to Get Started with the Model

The BLING model was fine-tuned with closed-context samples, which assume generally that the prompt consists of two sub-parts:

1. Text Passage Context, and
2. Specific question or instruction based on the text passage

To get the best results, package "my_prompt" as follows:

my_prompt = {{text_passage}} + "\n" + {{question/instruction}}

Model Card Contact

Darren Oberst & llmware team