Initial release: ParamTatva RLM Small v1 — Phonetically-Grounded Language Model

LICENSE

@@ -0,0 +1,98 @@
+ParamTatva Restricted Use License
+Version 1.0, February 2026
+
+Copyright (c) 2025-2026 ParamTatva.org
+All rights reserved.
+
+TERMS AND CONDITIONS
+
+1. DEFINITIONS
+
+   "Software" means the model weights, configuration files, and
+   associated documentation distributed under this License.
+
+   "Licensor" means ParamTatva.org.
+
+   "You" means the individual or entity exercising permissions
+   granted by this License.
+
+   "Commercial Use" means any use intended for or directed toward
+   commercial advantage or monetary compensation.
+
+   "Derivative Work" means any work that is based on or derived
+   from the Software, including but not limited to fine-tuned models,
+   distilled models, merged models, or quantized versions.
+
+2. GRANT OF RIGHTS
+
+   Subject to the terms of this License, the Licensor grants You a
+   worldwide, non-exclusive, non-transferable, revocable license to:
+
+   (a) Use the Software for research and academic purposes;
+   (b) Use the Software for personal, non-commercial applications;
+   (c) Create Derivative Works for research purposes only;
+   (d) Publish research results obtained using the Software,
+       provided proper attribution is given.
+
+3. RESTRICTIONS
+
+   You may NOT:
+
+   (a) Use the Software or any Derivative Work for Commercial Use
+       without a separate written agreement with the Licensor;
+   (b) Distribute Derivative Works without:
+       (i)  including this License;
+       (ii) providing clear attribution to ParamTatva.org;
+   (c) Attempt to reverse engineer, decompile, or derive the
+       training methodology, training data, or proprietary
+       components of the Licensor's systems from the Software;
+   (d) Remove or alter any copyright, trademark, or attribution
+       notices contained in the Software;
+   (e) Use the Software to train competing commercial products
+       without written permission;
+   (f) Sublicense the Software.
+
+4. ATTRIBUTION
+
+   Any use of the Software must include the following attribution:
+
+   "This work uses the ParamTatva Resonance Language Model (RLM),
+    developed by ParamTatva.org."
+
+   Academic citations should reference:
+   @misc{paramtatva2026rlm,
+     title={ParamTatva RLM: A Phonetically-Grounded Language Model},
+     author={ParamTatva.org},
+     year={2026},
+     url={https://huggingface.co/paramtatva/rlm-small-v1}
+   }
+
+5. COMMERCIAL LICENSING
+
+   For commercial use, contact: licensing@paramtatva.org
+
+   Commercial licenses are available for:
+   - Integration into commercial products
+   - Commercial API services
+   - Enterprise deployments
+
+6. NO WARRANTY
+
+   THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+   EXPRESS OR IMPLIED. IN NO EVENT SHALL THE LICENSOR BE LIABLE FOR
+   ANY CLAIM, DAMAGES, OR OTHER LIABILITY ARISING FROM THE SOFTWARE.
+
+7. TERMINATION
+
+   This License automatically terminates if You violate any of its
+   terms. Upon termination, You must destroy all copies of the
+   Software in Your possession.
+
+8. GOVERNING LAW
+
+   This License shall be governed by the laws of India.
+
+---
+
+For questions: legal@paramtatva.org
+Website: https://paramtatva.org

README.md

@@ -0,0 +1,128 @@
+---
+license: other
+license_name: paramtatva-restricted-1.0
+license_link: LICENSE
+language:
+  - sa
+  - en
+library_name: transformers
+tags:
+  - paramtatva
+  - rlm
+  - resonance
+  - sanskrit
+  - maheshwara-sutras
+  - math
+  - phonetic-grounding
+pipeline_tag: text-generation
+---
+
+# ParamTatva RLM-Small-v1
+
+**Resonance Language Model** — A phonetically-grounded transformer trained with insights from the Maheshwara Sutras.
+
+## Model Description
+
+ParamTatva RLM is a novel language model architecture that replaces standard positional encodings with **phonetic graph embeddings** derived from the [Maheshwara Sutras](https://en.wikipedia.org/wiki/Shiva_Sutras), the foundational grammar rules of Sanskrit attributed to Pāṇini.
+
+### Key Innovations
+
+| Feature | Description |
+|---------|-------------|
+| **Paramtatva Graph Embeddings** | Token embeddings informed by phonetic proximity in the Maheshwara Sutras |
+| **Pratyāhāra Attention Bias** | Attention biases derived from Pāṇini's abbreviation system (pratyāhāra) |
+| **Mā-Bridge Normalization** | Layer normalization conditioned on phonetic group structure |
+
+### Architecture
+
+```
+ParamtatvaTransformer (Small)
+├── Embedding: ParamtatvaEmbedding (phonetic graph-aware)
+├── Layers: 6 × TransformerBlock
+│   ├── Attention: Multi-Head + Pratyāhāra Bias
+│   ├── FFN: GELU activation
+│   └── Norm: LayerNorm + Mā-Bridge
+├── Final LayerNorm
+└── LM Head
+```
+
+| Parameter | Value |
+|-----------|-------|
+| Parameters | ~10M |
+| Hidden dim | 256 |
+| Layers | 6 |
+| Attention heads | 8 |
+| Intermediate dim | 1024 |
+| Max sequence length | 1024 |
+| Activation | GELU |
+
+## Intended Use
+
+This model is released for **research and academic purposes**. It demonstrates the viability of phonetically-grounded language modeling using ancient linguistic frameworks.
+
+### Recommended Uses
+- Research into phonetic/linguistic priors for language models
+- Studies on Sanskrit computational linguistics
+- Mathematical reasoning experiments
+- Exploration of alternative positional encoding schemes
+
+### Out-of-Scope Uses
+- Production/commercial applications (requires separate license)
+- Safety-critical systems
+- Any use that violates the license terms
+
+## Training
+
+The model was trained using the ParamTatva training pipeline. The training methodology, loss functions, and data curation are proprietary. Only the resulting model weights are released.
+
+**Note**: The full Resonance Learning System (including the proprietary ResonanceEncoder) is NOT included in this release. This release contains only the standard ParamtatvaTransformer weights.
+
+## How to Use
+
+```python
+import torch
+from safetensors.torch import load_file
+
+# Load weights
+state_dict = load_file("model.safetensors")
+
+# The model uses a custom architecture — see paramtatva_transformer.py
+# for the full model class definition.
+print(f"Parameters: {sum(v.numel() for v in state_dict.values()):,}")
+```
+
+## Limitations
+
+- This is a **small** model (~10M parameters) — intended as a proof of concept
+- The model was trained on a limited dataset
+- Performance on downstream tasks has not been extensively benchmarked
+- The proprietary resonance components are not included
+
+## Citation
+
+```bibtex
+@misc{paramtatva2026rlm,
+  title={ParamTatva RLM: A Phonetically-Grounded Language Model
+         Based on the Maheshwara Sutras},
+  author={ParamTatva.org},
+  year={2026},
+  url={https://huggingface.co/paramtatva/rlm-small-v1}
+}
+```
+
+## License
+
+This model is released under the **ParamTatva Restricted Use License v1.0**:
+- ✅ Research and academic use
+- ✅ Non-commercial applications
+- ✅ Fine-tuning for research
+- ❌ Commercial use (requires written agreement)
+- ❌ Reverse engineering of training methodology
+
+See [LICENSE](LICENSE) for full terms.
+
+## Contact
+
+- **Commercial licensing**: licensing@paramtatva.org
+- **Research inquiries**: research@paramtatva.org
+- **Website**: [paramtatva.org](https://paramtatva.org)

config.json

@@ -0,0 +1,25 @@
+{
+  "architectures": [
+    "ParamtatvaTransformer"
+  ],
+  "model_type": "paramtatva-rlm",
+  "vocab_size": 1224,
+  "hidden_size": 128,
+  "num_hidden_layers": 4,
+  "num_attention_heads": 2,
+  "intermediate_size": 512,
+  "max_position_embeddings": 1024,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "attention_probs_dropout_prob": 0.1,
+  "layer_norm_eps": 1e-06,
+  "initializer_range": 0.02,
+  "torch_dtype": "float32",
+  "transformers_version": "4.40.0",
+  "paramtatva_config": {
+    "use_graph_embeddings": true,
+    "use_pratyahara_bias": true,
+    "use_ma_bridge": true,
+    "phonetic_basis": "maheshwara_sutras"
+  }
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e345e1eb32297a516fe4ce74bb5c93ffea5fc76338b19b72758511a07c35bce
+size 4121338