Upload LICENSE with huggingface_hub
Browse files
LICENSE
CHANGED
|
@@ -10,32 +10,34 @@ ARE PROTECTED BY COPYRIGHT LAW, TRADE SECRET LAW, AND APPLICABLE INTELLECTUAL PR
|
|
| 10 |
PROPRIETARY ASSETS COVERED UNDER THIS LICENSE NOTICE INCLUDE, BUT ARE NOT LIMITED TO:
|
| 11 |
1. **Language-U Framework:** The compression taxonomy, semantic decomposition theory,
|
| 12 |
Shannon Bypass mathematical framework, and entropy decomposition equations.
|
| 13 |
-
2. **Cuneiform-U Semantic Hypercube System:** The 6-dimensional hypercube mapping along orthogonal axes
|
| 14 |
(Domain, Subdomain, Operation, Modality, Depth, Polarity) and coordinate radical representation
|
| 15 |
schemas (Classifier Radical R_C, Factor Radical R_F, and Active Radical R_A).
|
| 16 |
-
3. **
|
|
|
|
| 17 |
reconstruction, and dynamic layers streaming execution architecture.
|
| 18 |
-
|
| 19 |
encoding schemas, coordinate-packing matrices, and weights projection indexing.
|
| 20 |
-
|
| 21 |
and 28-chirp physical packetization formats), Qualia Seed configuration (0xE0), Huffman facts coding
|
| 22 |
mappings, and XOR-FEC (Forward Error Correction) recovery stream algorithms.
|
| 23 |
-
|
| 24 |
decomposition (SVD) projection targets, discrete cosine transform (DCT) spectral compression,
|
| 25 |
and grow_weight / shrink_weight reconstruction implementations.
|
| 26 |
-
|
| 27 |
and decoding mechanisms, dynamic entropy priors, and collapse signal scaling.
|
| 28 |
-
|
| 29 |
weight deltas/manifolds onto base-model shared word embedding matrices.
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
and edge-optimized memory management bindings (including C++, Rust, Go, Python, Swift,
|
| 34 |
Java, and TypeScript implementations).
|
| 35 |
-
|
| 36 |
multi-task SFT alignment recipes, loss balancing weights, and token-based on-device
|
| 37 |
reconstruction correction scripts (including decode_teleport.py).
|
| 38 |
-
|
|
|
|
| 39 |
artwork associated with Zymatica, Language-U, and TheAiCollective.art.
|
| 40 |
|
| 41 |
REPRODUCTION, DISSEMINATION, TRANSLATION, PORTING, REVERSE-ENGINEERING, OR MODIFICATION
|
|
|
|
| 10 |
PROPRIETARY ASSETS COVERED UNDER THIS LICENSE NOTICE INCLUDE, BUT ARE NOT LIMITED TO:
|
| 11 |
1. **Language-U Framework:** The compression taxonomy, semantic decomposition theory,
|
| 12 |
Shannon Bypass mathematical framework, and entropy decomposition equations.
|
| 13 |
+
2. **Cuneiform-U Semantic Hypercube System (Yin):** The 6-dimensional hypercube mapping along orthogonal axes
|
| 14 |
(Domain, Subdomain, Operation, Modality, Depth, Polarity) and coordinate radical representation
|
| 15 |
schemas (Classifier Radical R_C, Factor Radical R_F, and Active Radical R_A).
|
| 16 |
+
3. **Cuneiform-U Production Engine (Yang):** Edge-ready semantic range coder production engine.
|
| 17 |
+
4. **Genesis Protocol:** The multi-level procedural model transmission, sharded weights
|
| 18 |
reconstruction, and dynamic layers streaming execution architecture.
|
| 19 |
+
5. **ProceduralSeed File Format (.LLM / .genesis):** The custom binary seed format, layer
|
| 20 |
encoding schemas, coordinate-packing matrices, and weights projection indexing.
|
| 21 |
+
6. **Chirp Packetization & FEC Scheme:** The LoRA packet layout wrappers (including 78-chirp
|
| 22 |
and 28-chirp physical packetization formats), Qualia Seed configuration (0xE0), Huffman facts coding
|
| 23 |
mappings, and XOR-FEC (Forward Error Correction) recovery stream algorithms.
|
| 24 |
+
7. **SVD/DCT Compression & Reconstructor Pipeline:** The weights reduction algorithms, singular value
|
| 25 |
decomposition (SVD) projection targets, discrete cosine transform (DCT) spectral compression,
|
| 26 |
and grow_weight / shrink_weight reconstruction implementations.
|
| 27 |
+
8. **LLM-Logits-Driven Range Coding (LLD-AC):** The active logits-driven probability range coding
|
| 28 |
and decoding mechanisms, dynamic entropy priors, and collapse signal scaling.
|
| 29 |
+
9. **Embedding-Driven Weight Projection (E-PAUP / 1-PAUP):** The mathematical projection of target
|
| 30 |
weight deltas/manifolds onto base-model shared word embedding matrices.
|
| 31 |
+
10. **Tokenizer Prefix-Suffix Varint Differential Coding:** The lossless tokenizer serialization
|
| 32 |
+
storing tokens by ID order using variable-length prefix sharing indexes and suffix bytes.
|
| 33 |
+
11. **Multi-Language Runtimes & Ports (Yang):** Low-overhead execution runtimes, off-line tokenizers,
|
| 34 |
and edge-optimized memory management bindings (including C++, Rust, Go, Python, Swift,
|
| 35 |
Java, and TypeScript implementations).
|
| 36 |
+
12. **SFT Healing & Adaptive Cognitive Alignment:** The LoRA PEFT receiver-side healing loops,
|
| 37 |
multi-task SFT alignment recipes, loss balancing weights, and token-based on-device
|
| 38 |
reconstruction correction scripts (including decode_teleport.py).
|
| 39 |
+
13. **Zymatica Inference Engine (Class 27):** Multi-runtime execution inventory containing 27 language and target runtimes.
|
| 40 |
+
14. **Brand Assets & Artwork:** The official branding, visual representations, logos, and design
|
| 41 |
artwork associated with Zymatica, Language-U, and TheAiCollective.art.
|
| 42 |
|
| 43 |
REPRODUCTION, DISSEMINATION, TRANSLATION, PORTING, REVERSE-ENGINEERING, OR MODIFICATION
|