Update README.md

README.md

@@ -21,29 +21,32 @@ Neurazum is an exciting technological step that has the potential to revolutioni
 ## Featured Models and Systems
 
 ### **bai Family — EEG-Intelligent Models**
-A scalable family of brain–AI architectures designed for real-time EEG interpretation and multimodal neural signal analysis.
+A modular family of brain–AI architectures built for real-time EEG interpretation, cognitive state analysis, and neural signal intelligence.  
+Each model is task-specific, lightweight, and optimized for adaptive Brain–Computer Interface (BCI) use.
 
-- **bai-2, bai-4, bai-6, bai-8, bai-32, bai-64, bai-128, bai-256**  
-  Each model name reflects the **number of EEG electrodes** it processes.  
-  Optimized for everyday EEG use, supporting flexible Brain–Computer Interface (BCI) configurations and adaptive signal handling.
+- **bai-{Task}-{Channels} (e.g. bai-Epilepsy-6, bai-Emotion-8, bai-Mind2Text-4)**  
+  The naming system reflects both the **EEG channel count** and the **intended task**.  
+  Each model is trained for a single, well-defined function such as **seizure detection**, **emotion classification**, or **EEG-to-text reasoning**.  
+  This modular design allows models to be used independently or combined into larger multimodal systems.
 
-- **bai-x.y Versions (e.g., bai-6.1, bai-6.2)**  
-  Version identifiers denote functional evolution across iterations.  
-  Successive updates enhance the models from **single-task** performance (e.g., seizure prediction) to **multi-task capabilities**, enabling advanced signal pattern recognition and hybrid cognitive modeling.
+- **bai-{Task}-{Channels}-v{x} (e.g. bai-Epilepsy-6-v1, bai-Emotion-8-v2)**  
+  Version numbers indicate progressive refinements in accuracy, data diversity, and real-time stability.  
+  Each major version (v1, v2, v3, …) represents a full upgrade with new datasets, improved architectures, or enhanced task performance.  
+  Minor updates are integrated into the next major version rather than numbered separately.
 
-- **bai-xP Series (e.g., bai-6P, bai-64P, bai-256P)**  
-  The **Professional line**, built for clinical and research-grade use.  
-  These variants share the same naming logic but are trained on **medical datasets** for high-precision diagnostic performance and regulatory compliance.
+- **bai-{Task}-{Channels}P-v{x} (e.g. bai-Epilepsy-6P-v1, bai-Emotion-8P-v2)**  
+  The **Professional series**, trained on **medical-grade EEG datasets** and validated for clinical or research deployment.  
+  These models adhere to strict signal fidelity, diagnostic reliability, and interpretability standards for regulated environments.  
 
 ### **Vbai Family — Imaging and Brain Models**
 A collection of open-access models for neurological imaging, analysis, and 3D brain modeling.
 
-- **Vbai-x.x (e.g., Vbai-2.5, Vbai-3.0)**  
+- **Vbai-{x.y} (e.g. Vbai-2.5, Vbai-2.6, Vbai-3.0)**  
   Fully **open-source** models designed to handle multi-domain biomedical tasks.  
-  The major version coefficient (e.g., 2.0 → 3.0) changes with the introduction of new core functionalities.  
+  The major version coefficient (e.g. 2.0 → 3.0) changes with the introduction of new core functionalities.  
   Example: **Vbai-2.5** supports simultaneous **tumor and dementia detection**, enabling complex dual diagnostic workflows.
 
-- **Vbai-3D Series (e.g., Vbai-3D-v1, Vbai-3D-v2)**  
+- **Vbai-3D-v{x} (e.g. Vbai-3D-v1, Vbai-3D-v2)**  
   Specialized **3D-capable** open models for volumetric brain data analysis.  
   Each version (v1, v2, etc.) represents incremental improvements in precision, processing speed, and multimodal fusion.  
   These models can directly process 3D brain MRI inputs to detect structural and degenerative anomalies in real time.
@@ -52,12 +55,12 @@ A collection of open-access models for neurological imaging, analysis, and 3D br
 A family of lightweight transformer-based models designed to bridge natural language and neural data interpretation.  
 Tbai models function similarly to small-scale tokenizers like **T5**, optimized for concise reasoning and domain-specific understanding.
 
-- **Tbai-HF-v.x (e.g., Tbai-HF-v1, Tbai-HF-v2)**  
+- **Tbai-HF-v{x} (e.g. Tbai-HF-v1, Tbai-HF-v2)**  
   Models trained for **neuroimaging-based interpretation**.  
   Capable of generating analytical or descriptive text directly from **brain imaging data** such as MRI or CT scans.  
   Each version enhances domain adaptation, linguistic fluency, and cross-modal comprehension between visual brain patterns and textual insight.
 
-- **Tbai-NAI-v.x (e.g., Tbai-NAI-v1, Tbai-NAI-v2)**  
+- **Tbai-NAI-v{x} (e.g. Tbai-NAI-v1, Tbai-NAI-v2)**  
   Models specialized for **EEG-based cognitive commentary**.  
   Designed to interpret brainwave signals and translate them into meaningful linguistic explanations.  
   Subsequent versions expand accuracy in recognizing temporal EEG events, emotional states, and neurological anomalies.
@@ -65,7 +68,7 @@ Tbai models function similarly to small-scale tokenizers like **T5**, optimized
 ### **Lbai Family — Large-Scale Cognitive Reasoning Models**
 A suite of large language models (LLMs) trained on **medical, neuroscientific, and cognitive** corpora to perform advanced reasoning and diagnostic dialogue.
 
-- **Lbai-x-params_billion (e.g., Lbai-1-7B, Lbai-2-7B, Lbai-1-30B)**  
+- **Lbai-{x}-{params} (e.g. Lbai-1-7B, Lbai-2-7B, Lbai-1-30B)**  
   Each model’s identifier specifies its parameter scale in billions.  
   These models emulate expert-level reasoning, trained to **think and respond like medical professionals**, with a strong focus on neuroscience, neurophysiology, and clinical linguistics.  
   Lbai models aim to achieve near-complete domain comprehension across **brain science, medicine, and cognitive AI research**.