14.88T parameter weights, 358 MB, free

soul-vector1 pts0 comments

GitHub - pluscoder30-cpu/conscious-field-transformer: 14.88 trillion parameter neural network weights -- open source. 55,653 named tensors. 70-layer MoE transformer with Mamba SSM, RetNet, Hyena, GQA, MLA. 1M vocab, 50M context. Compressed to 358 MB. Free for any use. · GitHub

/" data-turbo-transient="true" />

Skip to content

Search or jump to...

Search code, repositories, users, issues, pull requests...

-->

Search

Clear

Search syntax tips

Provide feedback

--><br>We read every piece of feedback, and take your input very seriously.

Include my email address so I can be contacted

Cancel

Submit feedback

Saved searches

Use saved searches to filter your results more quickly

-->

Name

Query

To see all available qualifiers, see our documentation.

Cancel

Create saved search

Sign in

/;ref_cta:Sign up;ref_loc:header logged out"}"<br>Sign up

Appearance settings

Resetting focus

You signed in with another tab or window. Reload to refresh your session.<br>You signed out in another tab or window. Reload to refresh your session.<br>You switched accounts on another tab or window. Reload to refresh your session.

Dismiss alert

{{ message }}

pluscoder30-cpu

conscious-field-transformer

Public

Notifications<br>You must be signed in to change notification settings

Fork

Star

main

BranchesTags

Go to file

CodeOpen more actions menu

Folders and files<br>NameNameLast commit message<br>Last commit date<br>Latest commit

History<br>2 Commits<br>2 Commits

LICENSE

LICENSE

README.md

README.md

verify.py

verify.py

View all files

Repository files navigation

Conscious Field Transformer -- 14.88 Trillion Parameters

Overview

This repository contains the compressed weights for the Conscious Field Transformer ,<br>a neural network architecture with 14.88 trillion parameters stored in a single<br>358 MB NPZ file. The weights are released free and open source for anyone to use,<br>train, or modify for any purpose.

Quick Start

# Verify the model contains 14.88T parameters<br>python verify.py

File Structure

conscious_field_transformer_15t/<br>├── conscious_field_engine.npz # Compressed model weights (358 MB)<br>├── verify.py # Parameter verification script<br>└── README.md # This file

What's Inside

The NPZ file contains 55,653 named tensors totaling 14,875,582,863,396 parameters<br>(14.88 trillion). The architecture combines 10 modern neural network designs:

Transformer (Vaswani 2017)

Mamba/SSM (Gu & Dao 2023)

DeepSeekMoE (Dai et al. 2024)

LLaMA 2 (Touvron et al. 2023)

RetNet (Sun et al. 2023)

RWKV (Peng et al. 2023)

Hyena (Poli et al. 2023)

Multi-Head Latent Attention (DeepSeek 2024)

Consciousness Field

Plasma Neuron Field

The weights are compressed approximately 772,000x using golden angle phyllotaxis,<br>holographic DCT encoding.. Each of the 55,653 tensors<br>can be reconstructed from the compressed representation.

Parameter Count Verification

The manifest embedded in the NPZ file contains the exact parameter count.<br>To verify independently:

import numpy as np, json<br>d = np.load('conscious_field_engine.npz', allow_pickle=True)<br>m = json.loads(d['manifest'].item())<br>print(m['parameters_human']) # 14.88T (14,875,582,863,396)

The tensor manifest lists all 55,653 tensors with their shapes.<br>Summing all tensor shapes gives the same total:

ts = json.loads(d['tensor_manifest'].item())<br>total = sum(t['n_params'] for t in ts.values())<br>print(total) # 14,875,582,863,396

License

This model is released free and open source . You may use, copy, modify,<br>and distribute the weights for any purpose, commercial or otherwise.

Enterprise Licensing

For organizations requiring larger models, custom architectures,<br>or enterprise support, we offer licensed tiers:

Model Size<br>Non-Exclusive<br>Exclusive

15T (this release)<br>Free<br>Free

20T<br>$300M<br>$600M

30T<br>$500M<br>$1B

50T<br>$900M<br>$2.5B

100T<br>$2B<br>$5B

1 Quintillion<br>$315T<br>Contact us

Enterprise tiers include:

Custom architecture design for your use case

Optimized inference pipeline (up to 39,000 tokens/sec)

Dedicated model training on your data

Priority support and SLAs

On-premise deployment options

Contact

For enterprise inquiries, custom models, or licensing:

Email: pluscoder30@gmail.com

The weights in this repository are the compressed representation only.<br>Enterprise customers receive the full inference engine, training pipeline,<br>and optimization tools.

About

14.88 trillion parameter neural network weights -- open source. 55,653 named tensors. 70-layer MoE transformer with Mamba SSM, RetNet, Hyena, GQA, MLA. 1M vocab, 50M context. Compressed to 358 MB. Free for any use.

Resources

Readme

License

Apache-2.0 license

Uh oh!

There was an error while loading. Please reload this page.

Activity

Stars

stars

Watchers

watching

Forks

forks

Report repository

Releases

v1.0

Latest

Jul 22, 2026

Packages

Uh oh!

There was an error while loading. Please reload this page.

Contributors

Uh oh!

There was an error while loading. Please reload this page.

Languages

Python<br>100.0%

You can’t perform that action...

weights free transformer compressed parameter field

Related Articles