Instructions to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries
PEFT
How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with PEFT:
```
Task type is invalid.
```

How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

Use Docker

docker model run hf.co/reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

LM Studio
Jan

vLLM

How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

SGLang

How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with Ollama:
```
ollama run hf.co/reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M
```
Unsloth Desktop

How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with Docker Model Runner:
```
docker model run hf.co/reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M
```

Lemonade

How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

Run and chat with the model

lemonade run user.phi-4-mini-reasoning-primaryhealthcare-csci566-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Model Card for Phi-4-mini-reasoning — Primary Healthcare Fine-tune

Fine-tuned version of unsloth/Phi-4-mini-reasoning on a large medical QA dataset for primary healthcare, developed as part of the CSCI 566 (Deep Learning) course project at USC.

Model Details

Model Description

This model is a LoRA fine-tune of Phi-4-mini-reasoning, trained on ~146K medical multiple-choice QA examples to improve performance on clinical reasoning tasks relevant to primary healthcare. Phi-4-mini-reasoning's chain-of-thought reasoning capabilities make it well-suited for step-by-step medical QA. Both LoRA adapter weights and quantized GGUF files are provided.

Model type: Causal Language Model (LoRA fine-tune)
Language(s) (NLP): English
License: MIT
Finetuned from model: unsloth/Phi-4-mini-reasoning

Uses

Direct Use

Answer medical multiple-choice questions and provide evidence-based clinical reasoning for primary healthcare queries.

Downstream Use

Can be further fine-tuned or used as a backbone for medical decision-support tools, clinical education assistants, or healthcare chatbots.

Out-of-Scope Use

This model is not intended for direct clinical diagnosis or as a replacement for qualified medical professionals. It should not be used for emergency medical advice or high-stakes clinical decisions without human oversight.

Bias, Risks, and Limitations

The model may produce plausible-sounding but incorrect medical information.
Performance may degrade on rare conditions, non-English queries, or highly specialized subspecialties.

Recommendations

Always consult a qualified healthcare professional for personal medical advice. Outputs from this model should be treated as informational only and verified against authoritative clinical sources.

How to Get Started with the Model

With LoRA adapter (Transformers + PEFT)

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel

base_model = AutoModelForCausalLM.from_pretrained("unsloth/Phi-4-mini-reasoning")
tokenizer = AutoTokenizer.from_pretrained("reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566", subfolder="adapter")
model = PeftModel.from_pretrained(base_model, "reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566", subfolder="adapter")

With GGUF (llama.cpp / Ollama)

huggingface-cli download reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566 gguf/Phi-4-mini-reasoning.Q4_K_M.gguf

Training Details

Training Data

~146,354 medical multiple-choice QA examples in ShareGPT chat format, covering a broad range of primary care topics including internal medicine, pharmacology, and clinical reasoning. A validation split of ~5,548 examples was held out during training.

The system prompt used during training:

"You are a knowledgeable medical AI assistant. Provide accurate, evidence-based medical information. Always recommend consulting a qualified healthcare professional for personal medical advice."

Training Procedure

Training Hyperparameters

Training regime: bf16 mixed precision
Fine-tuning method: LoRA via Unsloth + TRL SFTTrainer
LoRA rank (r): 32
LoRA alpha: 64
LoRA dropout: 0.0
RSLoRA: Yes
Target modules: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj
Max sequence length: 1024
Per-device batch size: 4
Gradient accumulation steps: 8 (effective batch size: 32)
Learning rate: 2e-4
Warmup steps: 400
Weight decay: 0.01
Epochs: 1
Optimizer: AdamW

Evaluation

Testing Data, Factors & Metrics

Testing Data

Held-out validation split (~5,548 examples) from the same medical QA distribution as training data, plus standard medical benchmarks (MedQA, MedMCQA).

Metrics

Accuracy on multiple-choice medical QA benchmarks

Results

Benchmark	Base Model	Fine-tuned
CareQA	43.23%	56.77%
HEADQA	45.92%	60.69%
MedQA	41.63%	50.51%
MMLU (Medical)	50.10%	63.63%
MMLU-Pro (Medical)	31.42%	34.47%
PubMedQA	53.50%	69.70%

Summary

Phi-4-mini-reasoning shows the most consistent and significant improvements from fine-tuning across all models in this comparison. Every benchmark improved substantially, with HEADQA gaining +14.77%, MMLU +13.53%, CareQA +13.54%, and PubMedQA +16.20%. This suggests the base model's reasoning capabilities were well-suited to adapting to the medical domain, and the fine-tuning successfully transferred domain-specific knowledge without degrading general reasoning ability.

Technical Specifications

Model Architecture and Objective

Decoder-only transformer (Phi-4-mini-reasoning) with LoRA adapters applied to attention and MLP projection layers. Trained with next-token prediction (SFT) on chat-formatted medical QA data.

Compute Infrastructure

Hardware

NVIDIA RTX 5090 GPU: 24GB (CUDA 12.8)

Software

Python 3.13
Unsloth
TRL (SFTTrainer)
PEFT
Transformers

Model Card Authors

Reyansh Gupta — USC CSCI 566

Model Card Contact

reyanshg@usc.edu

Downloads last month: 29

GGUF

Model size

4B params

Architecture

phi3

Hardware compatibility

4-bit

Model tree for reyanshgupta/phi-4-mini-reasoning-primaryhealthcare-csci566

Base model

microsoft/Phi-4-mini-reasoning

Finetuned

unsloth/Phi-4-mini-reasoning

Adapter

(1)

this model

Evaluation results

Accuracy on CareQA
self-reported

56.770
Accuracy on HEADQA (EN)
self-reported

60.690
Accuracy on MedQA USMLE
self-reported

50.510
Accuracy on MMLU (Medical)
self-reported

63.630
Accuracy on MMLU-Pro (Medical)
self-reported

34.470
Accuracy on PubMedQA
self-reported

69.700