Delentia/delentia-slm-jitna-v0.4
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: Delentia/delentia-slm-jitna-v0.4
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-08-06 15:55:17 +08:00
commit 14fcab74b5
32 changed files with 5495 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

63
docs/AIR_GAPPED_ENTERPRISE_SDK_GUIDE.md Normal file
View File

@@ -0,0 +1,63 @@
# 🛡️ Delentia OS — Air-Gapped Enterprise SDK & Deployment Guide (v0.4.1)
**Document Classification:** Enterprise Architecture & Deployment Standard
**Compliance Scope:** 100% Offline Air-Gapped Networks (PDPA / GDPR Certified)
**Target Infrastructure:** Enterprise On-Premise Servers & Consumer Edge Terminals
---
## 💻 System Hardware Requirements Specification
To ensure high throughput and zero VRAM memory swapping during operational deployment, hardware hosts must fulfill the following certified profiles:
### 1. Consumer Edge Terminal Profile (Minimum Operational Spec)
Designed for local workstation deployment and offline employee terminals:
* **Operating System:** Windows 11 / Windows 10 (64-bit), Ubuntu 22.04 LTS, or macOS Sonoma
* **Central Processing Unit (CPU):** 8-Core x86_64 or ARM64 Processor (AVX2 support recommended)
* **System RAM Memory:** **8.0 GB Minimum** (Delentia GGUF Engine consumes **~4.9 GB RAM**)
* **Disk Storage Space:** 10.0 GB free SSD storage (NVMe recommended)
* **Graphics Acceleration (Optional):** Integrated Graphics or NVIDIA Consumer GPU (2GB VRAM cap)
### 2. Enterprise Server Profile (Production High-Throughput Spec)
Designed for multi-user enterprise gateways, database firewalls, and automated RCTDB synchronization:
* **Operating System:** RHEL 9 / Ubuntu Server 24.04 LTS / Windows Server 25
* **Central Processing Unit (CPU):** 16-Core / 32-Thread Enterprise Server Processor
* **System RAM Memory:** 32.0 GB High-Speed ECC RAM
* **Graphics Accelerator (GPU):** NVIDIA Data Center GPU (NVIDIA L4 24GB, A10 24GB, or T4 16GB)
* **Storage Infrastructure:** Enterprise NVMe RAID 1 (Zero-Latency I/O)
---
## 🔒 Air-Gapped Offline Installation Protocol
Delentia OS is engineered to run completely decoupled from external cloud networks. The deployment workflow follows three deterministic steps:
```
+--------------------------+ +--------------------------+ +--------------------------+
| 1. Download Offline Package| ----> | 2. Verify Cryptographic | ----> | 3. Launch Native C++ |
| (Delentia-OS.zip) | | Hashes (SHA-256) | | GGUF/Ollama Engine |
+--------------------------+ +--------------------------+ +--------------------------+
```
### Step 1: Transfer Offline Package
Transfer the certified `Delentia-OS.zip` (18.49 MB) and model binaries (`delentia-jitna-v0.4-Q4_K_M.gguf` - 4.92 GB) via secure USB or internal enterprise artifact repository (Artifactory / Nexus).
### Step 2: Cryptographic Verification
Verify payload integrity using SHA-256 hash validation to ensure zero tampering during air-gapped transport:
```bash
CertUtil -hashfile models/gguf/delentia-jitna-v0.4-Q4_K_M.gguf SHA256
```
### Step 3: Launch Native GGUF Runtime
Execute the native GGUF engine or register via Ollama enterprise CLI:
```bash
ollama create delentia-os -f models/Modelfile
ollama run delentia-os
```
---
## ⚖️ Regulatory Compliance & Privacy Attestation
* **Zero Cloud Egress:** 100% of telemetry, user prompts, and structured JSON tool calls remain strictly bound to local memory.
* **PDPA / GDPR Enforcement:** The Guardian Pillar automatically sanitizes sensitive PII (Personally Identifiable Information) prior to memory persistence in RCTDB.