Update README.md

README.md

@@ -13,75 +13,67 @@ tags:
 - code
 - conversational
 ---
-# Phi-4-Mini-Reasoning (GGUF Q4_KM)
+# Phi-4-Mini-Reasoning (GGUF Q4_KM) - Sandlogic Lexicons
 
-## Overview
+## Model Summary
 
-**Phi-4-Mini-Reasoning** is a compact, high-performance language model optimized for advanced mathematical reasoning tasks. Built upon the Phi-4-Mini architecture, this 3.8B parameter model excels in multi-step, logic-intensive problem-solving, particularly in environments with limited computational resources. :contentReference[oaicite:2]{index=2}:contentReference[oaicite:3]{index=3}
+**Phi-4-Mini-Reasoning** is a lightweight open-source model from the Phi-4 family, designed with a strong focus on high-quality, reasoning-dense synthetic data. It has been further fine-tuned for advanced mathematical reasoning tasks and supports a 128K token context length. This model is especially optimized for logic-intensive scenarios while maintaining a compact size, making it ideal for memory and compute-constrained environments.
 
-## Model Highlights
+- **Model Family**: Phi-4
+- **Parameter Count**: 3.8B
+- **Architecture**: Dense decoder-only Transformer
+- **Context Length**: 128K tokens
+- **Quantization**: GGUF Q4_KM
+- **Supported Language**: English
+- **Release Date**: April 2025
+- **Cutoff Date**: February 2025
 
-- **Architecture**: :contentReference[oaicite:5]{index=5}
-- **Vocabulary**: :contentReference[oaicite:8]{index=8}
-- **Attention Mechanism**: :contentReference[oaicite:11]{index=11}
-- **Context Length**: :contentReference[oaicite:14]{index=14}
-- **Training Data**: :contentReference[oaicite:17]{index=17}
-- **Training Duration**: :contentReference[oaicite:20]{index=20}
-- **Training Date**: :contentReference[oaicite:23]{index=23}
-- **Data Cutoff**: :contentReference[oaicite:26]{index=26}
-- **Release Date**: :contentReference[oaicite:29]{index=29}
-- **Supported Language**: :contentReference[oaicite:32]{index=32}:contentReference[oaicite:34]{index=34}
+## Intended Uses
 
-## Intended Use Cases
+### Primary Use Cases
 
-### Primary Applications
+Phi-4-Mini-Reasoning is designed to excel at:
 
-- :contentReference[oaicite:36]{index=36}
-- :contentReference[oaicite:39]{index=39}
-- :contentReference[oaicite:42]{index=42}
-- :contentReference[oaicite:45]{index=45}
-- :contentReference[oaicite:48]{index=48}:contentReference[oaicite:50]{index=50}
+- Multi-step mathematical reasoning
+- Formal proof generation
+- Symbolic computation
+- Solving advanced word problems
+- Tasks requiring structured logic and analytical thinking
 
-### Deployment Scenarios
+Its high context length and reasoning capabilities make it suitable for latency-bound applications and deployments on resource-constrained hardware.
 
-- :contentReference[oaicite:52]{index=52}
-- :contentReference[oaicite:55]{index=55}
-- :contentReference[oaicite:58]{index=58}:contentReference[oaicite:60]{index=60}
+### Use Case Considerations
 
-## Limitations and Considerations
+- This model is **optimized specifically for mathematical reasoning tasks**.
+- It is **not evaluated for general-purpose downstream tasks** such as conversational AI or creative writing.
+- Developers should:
+  - Assess use case suitability.
+  - Account for limitations in multi-language support.
+  - Evaluate performance, safety, and fairness—especially in high-risk or regulated environments.
+  - Ensure compliance with all applicable laws and regulations (e.g., privacy and trade compliance).
 
-- **Domain Specificity**: :contentReference[oaicite:62]{index=62}
-- **Language Support**: :contentReference[oaicite:65]{index=65}
-- **Ethical Use**: :contentReference[oaicite:68]{index=68}
-- **Risk Mitigation**: :contentReference[oaicite:71]{index=71}:contentReference[oaicite:73]{index=73}
+## Training Details
 
-## Training Methodology
+- **Model Architecture**: Same as Phi-4-Mini with 3.8B parameters
+- **Notable Enhancements**: 
+  - 200K vocabulary
+  - Grouped-query attention
+  - Shared input/output embeddings
+- **Training Dataset Size**: 150B tokens
+- **Training Duration**: 2 days
+- **Hardware Used**: 128 × H100-80G GPUs
+- **Training Date**: February 2024
+- **Output**: Generated text
+- **Input Format**: Text (chat-style prompts recommended)
 
-The training process for Phi-4-Mini-Reasoning involved a multi-stage approach:
+## Integration in Lexicons
 
-1. **Mid-Training**: :contentReference[oaicite:75]{index=75}
-2. **Supervised Fine-Tuning**: :contentReference[oaicite:78]{index=78}
-3. **Rollout DPO**: :contentReference[oaicite:81]{index=81}
-4. **Reinforcement Learning**: :contentReference[oaicite:84]{index=84} :contentReference[oaicite:86]{index=86}:contentReference[oaicite:87]{index=87}
+This quantized GGUF Q4_KM version of Phi-4-Mini-Reasoning is included in our [Sandlogic Lexicons](https://huggingface.co/SandLogicTechnologies) model zoo, making it readily available for efficient inference in edge deployments and research use cases focused on math reasoning.
 
-## Performance Benchmarks
-
-:contentReference[oaicite:89]{index=89} :contentReference[oaicite:91]{index=91}:contentReference[oaicite:92]{index=92}
-
-## Format and Integration
-
-- **Model Format**: :contentReference[oaicite:94]{index=94}
-- **Integration**: :contentReference[oaicite:97]{index=97}
-- **Lexicon Addition**: :contentReference[oaicite:100]{index=100}:contentReference[oaicite:102]{index=102}
-
-## License and Usage
-
-:contentReference[oaicite:104]{index=104}:contentReference[oaicite:106]{index=106}
+---
 
-## References
+*For optimal results, we recommend using Phi-4-Mini-Reasoning in tasks that require deep mathematical analysis and structured problem solving.*
 
-- :contentReference[oaicite:108]{index=108}
-- :contentReference[oaicite:111]{index=111}:contentReference[oaicite:113]{index=113}