MARA is an open-weight model from Korea Research Institute of Chemical Technology, released under cc-by-nc-nd-4.0. Its published files total 2.4 GB.
Parameters—
Context—
Weights249.3 MB
Licensecc-by-nc-nd-4.0
AccessOpen weights
Monthly Downloads—
Model Card
The publisher has not written a card for this model.
Identity and Version
- Repository
- KRICT-AI/MARA
- Publisher
- Korea Research Institute of Chemical Technology
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 7f7f7df97596d1bc5d9686181053eb1b2a4be770
- First published
- 2026-06-23
- Last updated
- 2026-10-05
Files and Weights
29 files, 2.4 GB in total. The weights are 1 file totalling 249.3 MB in pt.
Weights1 file · 249.3 MB
Configuration16 files · 626.3 KB
Documentation1 file · 33 B
Other10 files · 2.2 GB
Repository1 file · 1.6 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| save/dataset/nist.pt | Weights | 249.3 MB | c7ebee48aacf |
| exec_mara.py | Configuration | 4.8 KB | — |
| exec_mara_nist.py | Configuration | 3.9 KB | — |
| exec_nist.py | Configuration | 3.3 KB | — |
| exec_qme14s.py | Configuration | 3.0 KB | — |
| mara/llm.py | Configuration | 5.5 KB | — |
| mara/model.py | Configuration | 5.4 KB | — |
| mara/train_util.py | Configuration | 1.5 KB | — |
| ml/cross_modality_retriever.py | Configuration | 4.4 KB | — |
| ml/nn.py | Configuration | 1.2 KB | — |
| res/cgcnn-embedding.json | Configuration | 28.4 KB | — |
| res/matscholar-embedding.json | Configuration | 449.2 KB | — |
| res/megnet16-embedding.json | Configuration | 34.4 KB | — |
| res/onehot-embedding.json | Configuration | 63.6 KB | — |
| util/chem.py | Configuration | 6.4 KB | — |
| util/data.py | Configuration | 9.6 KB | — |
| util/json_util.py | Configuration | 1.8 KB | — |
| README.md | Documentation | 33 B | — |
| mara/__pycache__/llm.cpython-312.pyc | Other | 7.3 KB | — |
| mara/__pycache__/model.cpython-312.pyc | Other | 8.9 KB | — |
| mara/__pycache__/train_util.cpython-312.pyc | Other | 2.8 KB | — |
| mara_generated_reports.zip | Other | 1.5 GB | db320697a9c2 |
| ml/__pycache__/cross_modality_retriever.cpython-312.pyc | Other | 6.8 KB | — |
| ml/__pycache__/nn.cpython-312.pyc | Other | 2.5 KB | — |
| save/explainability_evalation_results.zip | Other | 739.8 MB | 984e72492a04 |
| util/__pycache__/chem.cpython-312.pyc | Other | 10.9 KB | — |
| util/__pycache__/data.cpython-312.pyc | Other | 14.8 KB | — |
| util/__pycache__/json_util.cpython-312.pyc | Other | 2.2 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
License and Download
- License
- cc-by-nc-nd-4.0
- Access
- Open weights, no gate
- Download size
- 249.3 MB
Download from Korea Research Institute of Chemical Technology
Released by Korea Research Institute of Chemical Technology through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 249.3 MB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About MARA
What license is MARA released under?
cc-by-nc-nd-4.0, as its publisher declares it. Read the license text before commercial use.