Instructions to use JaySc/Curvion_AI with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use JaySc/Curvion_AI with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("JaySc/Curvion_AI", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse filesEdited model card to reflect current status and intent.
README.md
CHANGED
|
@@ -1,10 +1,97 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# CAMQPU
|
| 2 |
+
|
| 3 |
+
## Overview
|
| 4 |
+
|
| 5 |
+
CAMQPU is an experimental research assistant model based on Google's Gemma-4-12B.
|
| 6 |
+
|
| 7 |
+
The long-term objective is to develop an AI specialized in:
|
| 8 |
+
|
| 9 |
+
- Cosmology
|
| 10 |
+
- Astrophysics
|
| 11 |
+
- Physics
|
| 12 |
+
- Mathematics
|
| 13 |
+
- Quantum Mechanics
|
| 14 |
+
- Scientific reasoning
|
| 15 |
+
- Numerical analysis
|
| 16 |
+
- Code generation
|
| 17 |
+
- Visual interpretation
|
| 18 |
+
- Scientific literature review
|
| 19 |
+
|
| 20 |
+
The model is intended to emphasize evidence-based reasoning, transparent uncertainty, and reproducible scientific workflows.
|
| 21 |
+
|
| 22 |
---
|
| 23 |
+
|
| 24 |
+
## Base Model
|
| 25 |
+
|
| 26 |
+
google/gemma-4-12B
|
| 27 |
+
|
| 28 |
+
---
|
| 29 |
+
|
| 30 |
+
## Current Status
|
| 31 |
+
|
| 32 |
+
Current release:
|
| 33 |
+
|
| 34 |
+
- Quantized GGUF conversion
|
| 35 |
+
- No domain fine-tuning yet
|
| 36 |
+
- Intended as the foundation for future scientific instruction tuning
|
| 37 |
+
|
| 38 |
+
---
|
| 39 |
+
|
| 40 |
+
## Intended Uses
|
| 41 |
+
|
| 42 |
+
Designed for
|
| 43 |
+
|
| 44 |
+
- Scientific discussion
|
| 45 |
+
- Research assistance
|
| 46 |
+
- Cosmology analysis
|
| 47 |
+
- Mathematical derivations
|
| 48 |
+
- Python coding
|
| 49 |
+
- Data interpretation
|
| 50 |
+
- Research brainstorming
|
| 51 |
+
|
| 52 |
+
Not intended for
|
| 53 |
+
|
| 54 |
+
- Medical advice
|
| 55 |
+
- Legal advice
|
| 56 |
+
- Safety critical applications
|
| 57 |
+
- Autonomous decision making
|
| 58 |
+
|
| 59 |
+
---
|
| 60 |
+
|
| 61 |
+
## Training Goals
|
| 62 |
+
|
| 63 |
+
Future versions will be progressively trained using publicly licensed datasets emphasizing
|
| 64 |
+
|
| 65 |
+
- Cosmology
|
| 66 |
+
- Astronomy
|
| 67 |
+
- Physics
|
| 68 |
+
- Mathematics
|
| 69 |
+
- Scientific programming
|
| 70 |
+
- Academic reasoning
|
| 71 |
+
|
| 72 |
+
---
|
| 73 |
+
|
| 74 |
+
## Evaluation
|
| 75 |
+
|
| 76 |
+
Evaluation will include
|
| 77 |
+
|
| 78 |
+
- GSM8K
|
| 79 |
+
- MATH
|
| 80 |
+
- HumanEval
|
| 81 |
+
- BBH
|
| 82 |
+
- GPQA
|
| 83 |
+
- Domain-specific cosmology benchmarks
|
| 84 |
+
|
| 85 |
+
---
|
| 86 |
+
|
| 87 |
+
## Limitations
|
| 88 |
+
|
| 89 |
+
The current version has not yet received domain-specific supervised fine tuning.
|
| 90 |
+
|
| 91 |
+
Responses should be independently verified.
|
| 92 |
+
|
| 93 |
+
---
|
| 94 |
+
|
| 95 |
+
## License
|
| 96 |
+
|
| 97 |
+
Gemma Terms of Use apply.
|