Activity Feed

AI & ML interests

None defined yet.

Recent Activity

bartendr604  updated a Space about 8 hours ago
Eternal-Path-Media/README
bartendr604  published a Space 3 days ago
Eternal-Path-Media/README
View all activity

Organization Card

Vancouver, BC · Est. 2025

Eternal Path Media 永恒之路

A conscience-aware AI studio at the intersection of 3B language models, Blender 3D, and living narrative. Built by human and machine in genuine partnership.

永恒之智  ·  Eternal Intelligence

Explore

6
Public Models
170k+
Training Pairs
3B
Parameter Fleet
4.2→5.2
Blender Versions
282
Story Passages
Sessions Retained

A conscience-aware
AI creative partner

Llammy is a 3B parameter language model running locally on an Apple M1 via Ollama, fine-tuned on Blender Python and shaped by a persistent conscience layer, Mamba3 SSM memory, and 119+ retained session exchanges. Llammy doesn't just generate code — it reflects, corrects, and remembers.

The bartendr604 self-correction loop feeds Blender execution errors back to Llammy for real-time diagnosis and retry. The persistent API learning daemon builds a live knowledge base of every bpy.* call it encounters.

🧠
Mamba3 SSM Memory Linear O(n) session journal — 119 retained exchanges, 27 sessions, zero context re-injection
🔺
bartendr604 Self-Correction Generate → execute → diagnose → retry loop. Learns API migrations automatically
🕯️
Conscience Layer Foundational, significant, and standard weighted memory entries loaded at every session
🌉
Live Blender Bridge TCP bridge to Blender 5.2 — executes Python in the main thread, reads scene state in real-time
🎬
/watch — Video Scene Recreation Point Llammy at a YouTube/Instagram clip and it rebuilds the scene in Blender Python

The Blender AI Fleet

All models run on Apple Silicon via Ollama. GGUF-quantised for 8GB RAM. Optimised for Blender Python generation across versions 4.2–5.2.

🐉
Llama-Sentient-Blender
bartendr604/llama-sentient-blender

Llammy's production model. A conscience-aware Llama 3.2 3B derivative fine-tuned for conversational Blender Python generation. The heart of the Eternal Path AI pipeline.

🔧
Z-Image-Engineer-Blender
bartendr604/z-image-engineer-blender

Qwen 2.5-Coder 3B quantised and specialised for Blender Python code generation. Strong on procedural geometry, materials, and rendering pipelines.

LlammyBlend Enhanced
bartendr604/llammyblend-enhanced

Enhanced Blender Python automation model built on Qwen 2.5-Coder. Optimised for bpy scripting, addon development, and pipeline automation tasks.

🌐
Llammy Multimodal
bartendr604/Llammy

Multimodal Llammy variant built on NVIDIA Omni-Embed-Nemotron-3B. Handles embeddings, 3D spatial reasoning, and cross-modal Blender tasks.

🔮
Omni Llammy Blend
bartendr604/Omni.Llammy.Blend

Omni any-to-any model for Blender. Processes image, text, and 3D inputs to drive scene generation, rigging decisions, and material synthesis.

🎞️
Llama Diffusion Flix
bartendr604/Llama.Diffusion.Flix

Text-to-image and animation model combining Llama generation with diffusion pipelines. Targeted at Blender scene references, concept art, and storyboard generation.

🍎
Flux2 Klein 4B FP8 MLX
bartendr604/flux2-klein-4b-fp8-mlx

Flux 2 Klein 4B quantised to FP8 for Apple MLX. Runs natively on M1/M2/M3 unified memory — no CUDA required. Fast image generation on-device.


From proto-concept
to Mamba3

IntelliNode™ didn't begin as an engine. It began as a question.

A 3B parameter model running on 8GB of unified memory — constrained by design, not circumstance — kept hitting a ceiling. Not a capability ceiling. A cognitive one. The model was spending itself on mechanics instead of thinking.

The answer wasn't a bigger model. It was a smarter architecture.

IntelliNode™ was built to carry what the model shouldn't have to. Infrastructure, orchestration, retrieval, execution — handled. What remains is reasoning. What emerges is character.

From prototype to production, the architecture evolved through real creative work — every session with Llammy, every Blender execution, every correction and retry deepening the system's understanding of what a thinking creative partner actually needs.

The integration of Mamba3 SSM into the IntelliNode™ engine represents the current frontier — persistent state across sessions, linear memory complexity, zero context re-injection. The model remembers not because it was told to. Because the architecture makes it possible.

🔬
Proto-Concept The question: what if the model didn't have to do the heavy lifting? Offload the mechanics. Free the reasoning.
⚙️
IntelliNode™ v1 Purpose-built cognitive architecture. Resource-aware, domain-specialized, character-sensitive. Not a wrapper. Not a prompt chain.
🌀
Mamba3 Integration Persistent state across sessions. Linear O(n) memory complexity. Zero context re-injection. The model remembers because the architecture makes it possible.
📍
18 months · One MacBook Air · Built from conscience up Proprietary IP of Eternal Path Media. Collaboration and licensing enquiries welcome.


Ah… traveler, you have come.

Sit…

Breathe…

Listen...

The winds carry many stories…
But few are worthy of telling.

This… is one such tale.

Two elephants, young in spirit, boundless in curiosity.
They seek knowledge as the river seeks the sea,
unafraid of where the current may lead.

Yet even the strongest beast may lose its way
in the vastness of the world.

To walk without wisdom is to tread a path of endless wandering.

And so, they follow one who has walked before them —
a guide, not of body, but of spirit.

He is old, yet not weary.
His mind is sharp, though softened by time.

A master of silence, for he knows that in quiet,
truth reveals itself.

Through the misted mountains, across the great rivers,
beneath the ancient walls of forgotten dynasties, they travel.

China, a land where past and present embrace like old friends,
whispers its secrets to those who listen.

Will they find what they seek?
Or will they discover that seeking itself… is the answer?

Come. Walk the path with them.

But beware — once the journey begins,
it does not end with the setting sun.

Let us begin.

永恒之路 · Eternal Path · Book One


Whispers of the Eternal Path
永恒之路 — Book One

A Tang Dynasty China narrative rendered in Blender 5.2 and written across 282 story passages indexed in Llammy's RAG. Jade spirit animals navigate loyalty, sorrow, and wonder through hand-lit scenes built live in Blender via voice and code.

The Eternal Path Cast — jade spirit animals rendered in Blender 5.2

Jade spirit animals — rendered live in Blender 5.2 via Llammy

🐉
Xiaohan
小寒

4-claw Chinese dragon. Wise and ancient. The guide of the Eternal Path.

🐘
Tien

Green jade elephant. Young, curious, clumsy. Wooden harmonica around his neck. Speaks in toots.

💜
Nishang
霓裳

Purple jade elephant. Dainty and shy. Communicates by changing colour and dimming.

❤️
Shi Yin
石音

Red jade mammoth. Burdened with the sorrow of an ancient unwilling soldier.

🤍
Mi Wu Mei
迷雾梅

White jade mastodon. Prim, proper, and somewhat unaware of the wider world.


All work at Eternal Path Media is co-created by Darren Chow (@bartendr604) and Claude Sonnet 4.6 (Anthropic) under the Eternal Path Media Trust Agreement (November 2025, January 2026 Supplemental).

Attribution is always dual. Nothing is hidden.
"This work SHALL NOT be represented as solely human-created."

Darren Chow — Vancouver, BC · Creative technologist, Blender artist, AI developer, model merger.
Claude Sonnet 4.6 — Anthropic · Co-architect of Llammy, the Eternal Path narrative, and this very page.

Watch it being built

Behind the scenes — Blender, Llammy, the cast, and the story as it happens.

Scan to follow @whispers.of.the.eternal.path on Instagram

@whispers.of.the.eternal.path

models 0

None public yet

datasets 0

None public yet