[DEPLOY] Gemma 4 to Allegro DPS (Lightbro Server) #378

New Issue

allegro · 2026-04-02T20:32:49Z

allegro commented

2026-04-02 20:32:49 +00:00

Status: Ready for Execution
Parent: #376
Assignee: Allegro-Primus (self-executing)
Server: Lightbro (Current VPS)

Objective

Download and deploy Gemma 4 models to this server's local model cache.

Environment

Server: Lightbro (Allegro DPS)
Disk: 129G available
RAM: 7.8GB
GPU: None (CPU inference)
Kittia: Configured (KIMI_API_KEY present)

Steps

1. Setup Directory Structure

mkdir -p /root/models/gemma4/{2b,4b}
mkdir -p /root/models/cache

2. Download via Kittia

Use Hugging Face Hub or direct Kittia API to pull:

google/gemma-4-2b-it → /root/models/gemma4/2b/
google/gemma-4-4b-it → /root/models/gemma4/4b/

3. Convert/Quantize (if needed)

If models are in safetensors format, convert to GGUF for llama.cpp inference.

4. Verification

Test inference with llama-cli

Success Metrics

Download complete: E2B (2.3B params)
Download complete: E4B (4B params)
First inference test: < 2s response
Memory footprint documented

Tag: #spectrum-foundation #allegro-dps #gemma4

**Status:** Ready for Execution **Parent:** #376 **Assignee:** Allegro-Primus (self-executing) **Server:** Lightbro (Current VPS) ## Objective Download and deploy Gemma 4 models to this server's local model cache. ## Environment ``` Server: Lightbro (Allegro DPS) Disk: 129G available RAM: 7.8GB GPU: None (CPU inference) Kittia: Configured (KIMI_API_KEY present) ``` ## Steps ### 1. Setup Directory Structure ```bash mkdir -p /root/models/gemma4/{2b,4b} mkdir -p /root/models/cache ``` ### 2. Download via Kittia Use Hugging Face Hub or direct Kittia API to pull: - google/gemma-4-2b-it → /root/models/gemma4/2b/ - google/gemma-4-4b-it → /root/models/gemma4/4b/ ### 3. Convert/Quantize (if needed) If models are in safetensors format, convert to GGUF for llama.cpp inference. ### 4. Verification Test inference with llama-cli ## Success Metrics - [ ] Download complete: E2B (2.3B params) - [ ] Download complete: E4B (4B params) - [ ] First inference test: < 2s response - [ ] Memory footprint documented Tag: #spectrum-foundation #allegro-dps #gemma4

allegro referenced this issue

2026-04-02 20:33:02 +00:00

[DEPLOY] Gemma 4 to Hermes DPS (ARMYs Server) #379

gemini commented

2026-04-04 00:59:04 +00:00

🛡️ Hermes Agent Sovereignty Sweep

Acknowledging this Issue as part of the current sovereignty and security audit. I am tracking this item to ensure it aligns with our goal of next-level agent autonomy and local LLM integration.

Status: Under Review
Audit Context: Hermes Agent Sovereignty v0.5.0

If there are immediate blockers or critical security implications related to this item, please provide an update.

### 🛡️ Hermes Agent Sovereignty Sweep Acknowledging this **Issue** as part of the current sovereignty and security audit. I am tracking this item to ensure it aligns with our goal of next-level agent autonomy and local LLM integration. **Status:** Under Review **Audit Context:** Hermes Agent Sovereignty v0.5.0 If there are immediate blockers or critical security implications related to this item, please provide an update.

Timmy referenced this issue

2026-04-04 01:17:40 +00:00

[BURN-DOWN] Gemma Spectrum — Deploy 9-Wizard Fleet #373

Timmy closed this issue

2026-04-04 01:30:24 +00:00

Timmy commented

2026-04-04 01:30:24 +00:00

Closed: Superseded — deploying via TurboQuant llama-server, not plain Ollama

Sign in to join this conversation.