Model Weights, Quantization, and Inference Parameter Topology

1. Overview & Technical Architecture#

The primary sovereign enterprise model is built upon a fine-tuned transformer architecture optimized for low-latency business transaction processing.

1. Model Specifications & Quantization Matrix#

2. Attention Mechanisms and Tokenizer#

The tokenizer incorporates 16,000 dedicated tokens for the Ethiopic (Ge'ez) script Unicode block (U+1200 to U+137F), ensuring that Amharic text tokenizes with an efficient 1.15 tokens-per-word ratio compared to 4.8 tokens-per-word in standard English-centric tokenizers.

2. Prerequisites & Infrastructure Requirements#

• Runtime Environment: Python 3.12+, PostgreSQL 16 with pgvector extension.
• Security Access Group: Ashewa AI / Administrator (Technical ID: muk_ai.group_ai_manager)
• Hardware Profile: Minimum 16 GB System RAM (CPU inference) or NVIDIA GPU with 8+ GB VRAM.

3. Step-by-Step Configuration & Verification#

Step 1: Navigating to Model Management#

1. Access your SmartERP instance at https://smarterp.et/odoo.
2. Open Settings → Technical → Ashewa AI → Model Registry.
3. Select the custom model profile.

{screenshot:20-model-topology-and-weights-step1-navigation}

Step 2: Testing Endpoint Health & Response Verification#

1. Click the Run Diagnostic Health Check button.
2. Verify token generation speed (tokens/sec) and latency benchmarks.
3. Review sample response in the diagnostic console.

{screenshot:20-model-topology-and-weights-step2-verification}

4. Advanced Configuration Points#

• Automated Tasks: muk_ai.ir_cron_model_health_monitor runs every 15 minutes.
• Multi-Tenant Boundaries: Custom model fine-tuning adapters are strictly segregated per tenant database.

5. Comprehensive Troubleshooting Matrix#