> For the complete documentation index, see [llms.txt](https://infinitode-docs.gitbook.io/documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://infinitode-docs.gitbook.io/documentation/ai-documentation.md).

# AI Documentation

Infinitode’s AI resources are small, single-purpose pretrained models — mostly tabular classification and regression — plus open datasets and the OPEN-ARC research projects. Get them from Hugging Face

## AI resources

Use these resources for research, prototyping, and fine-tuning. Each entry links to its model or dataset and supporting implementation.

Start with a model when you need inference. Start with a dataset when you need training data. Use [OPEN-ARC](https://github.com/Infinitode/OPEN-ARC) for guided research projects.

Browse [Infinitode Spaces on Hugging Face](https://huggingface.co/spaces?author=InfinitodeLTD) for interactive demos and applications.

## :robot: AI Models

Our models are small and single-purpose: each one does one job for one kind of data. Most are **tabular classification** or **tabular regression** models from [OPEN-ARC](https://github.com/Infinitode/OPEN-ARC) — crop recommendation, plant stress, traffic accidents, mushrooms, personality traits, iris species, spam mail, housing prices. On the text side we publish a compact **name generator** rather than general language models, and we publish one **image upscaler**. GitHub links point at the notebook each model came from.

{% hint style="info" %}
**What we don't have:** open-ended chat, story or essay models, and no general text-generation API. If you need prose, bring your own language model — our models are tuned for short, structured output like a name, a label, or a number.
{% endhint %}

<table data-view="cards"><thead><tr><th>View all models</th><th data-card-target data-type="content-ref">Hugging Face</th></tr></thead><tbody><tr><td>Browse every Infinitode model on Hugging Face.</td><td><a href="https://huggingface.co/models?author=InfinitodeLTD">https://huggingface.co/models?author=InfinitodeLTD</a></td></tr></tbody></table>

<table data-view="cards"><thead><tr><th>Model Name</th><th>Task</th><th data-type="content-ref">HF Link</th><th data-type="content-ref">GitHub Link</th><th>Latest Version</th></tr></thead><tbody><tr><td><strong>terraria-weapon-name-generation-model (OPEN-ARC)</strong></td><td>Text generation (name generation)</td><td><a href="https://huggingface.co/InfinitodeLTD/TWNGM-OPEN-ARC">https://huggingface.co/InfinitodeLTD/TWNGM-OPEN-ARC</a></td><td><a href="https://github.com/Infinitode/OPEN-ARC/blob/main/Project-5-TWNG/project-5-twng.ipynb">https://github.com/Infinitode/OPEN-ARC/blob/main/Project-5-TWNG/project-5-twng.ipynb</a></td><td>v1.0.0</td></tr><tr><td><strong>crop-recommendation-model (OPEN-ARC)</strong></td><td>Tabular classification</td><td><a href="https://huggingface.co/InfinitodeLTD/CRM-OPEN-ARC">https://huggingface.co/InfinitodeLTD/CRM-OPEN-ARC</a></td><td><a href="https://github.com/Infinitode/OPEN-ARC/blob/main/Project-7-CR/project-7-cr.ipynb">https://github.com/Infinitode/OPEN-ARC/blob/main/Project-7-CR/project-7-cr.ipynb</a></td><td>v1.0.0</td></tr><tr><td><strong>plant-stress-prediction-model (OPEN-ARC)</strong></td><td>Tabular classification</td><td><a href="https://huggingface.co/InfinitodeLTD/PSPM-OPEN-ARC">https://huggingface.co/InfinitodeLTD/PSPM-OPEN-ARC</a></td><td><a href="https://github.com/Infinitode/OPEN-ARC/blob/main/Project-8-PSPM/project-8-pspm.ipynb">https://github.com/Infinitode/OPEN-ARC/blob/main/Project-8-PSPM/project-8-pspm.ipynb</a></td><td>v1.0.0</td></tr><tr><td><strong>traffic-accident-prediction-model (OPEN-ARC)</strong></td><td>Tabular classification</td><td><a href="https://huggingface.co/InfinitodeLTD/TAPM-OPEN-ARC">https://huggingface.co/InfinitodeLTD/TAPM-OPEN-ARC</a></td><td><a href="https://github.com/Infinitode/OPEN-ARC/blob/main/Project-9-TAPM/project-9-tapm.ipynb">https://github.com/Infinitode/OPEN-ARC/blob/main/Project-9-TAPM/project-9-tapm.ipynb</a></td><td>v1.0.0</td></tr><tr><td><strong>mushroom-classification-model (OPEN-ARC)</strong></td><td>Tabular classification</td><td><a href="https://huggingface.co/InfinitodeLTD/MCM-OPEN-ARC">https://huggingface.co/InfinitodeLTD/MCM-OPEN-ARC</a></td><td><a href="https://github.com/Infinitode/OPEN-ARC/blob/main/Project-10-MCM/project-10-mcm.ipynb">https://github.com/Infinitode/OPEN-ARC/blob/main/Project-10-MCM/project-10-mcm.ipynb</a></td><td>v1.0.0</td></tr><tr><td><strong>basic-personality-prediction-model (OPEN-ARC)</strong></td><td>Tabular classification</td><td><a href="https://huggingface.co/InfinitodeLTD/BPPM-OPEN-ARC">https://huggingface.co/InfinitodeLTD/BPPM-OPEN-ARC</a></td><td><a href="https://github.com/Infinitode/OPEN-ARC/blob/main/Project-11-BPPM/project-11-bppm.ipynb">https://github.com/Infinitode/OPEN-ARC/blob/main/Project-11-BPPM/project-11-bppm.ipynb</a></td><td>v1.0.0</td></tr><tr><td><strong>spam-mail-classification-model (OPEN-ARC)</strong></td><td>Text classification</td><td><a href="https://huggingface.co/InfinitodeLTD/SMCM-OPEN-ARC">https://huggingface.co/InfinitodeLTD/SMCM-OPEN-ARC</a></td><td><a href="https://github.com/Infinitode/OPEN-ARC/blob/main/Project-12/notebook.ipynb">https://github.com/Infinitode/OPEN-ARC/blob/main/Project-12/notebook.ipynb</a></td><td>v1.0.0</td></tr><tr><td><strong>india-housing-price-prediction-model (OPEN-ARC)</strong></td><td>Tabular regression</td><td><a href="https://huggingface.co/InfinitodeLTD/IHPPM-OPEN-ARC">https://huggingface.co/InfinitodeLTD/IHPPM-OPEN-ARC</a></td><td><a href="https://github.com/Infinitode/OPEN-ARC/blob/main/Project-13/notebook.ipynb">https://github.com/Infinitode/OPEN-ARC/blob/main/Project-13/notebook.ipynb</a></td><td>v1.0.0</td></tr><tr><td><strong>iris-flower-species-classification-model (OPEN-ARC)</strong></td><td>Tabular classification</td><td><a href="https://huggingface.co/InfinitodeLTD/IFSCM-OPEN-ARC">https://huggingface.co/InfinitodeLTD/IFSCM-OPEN-ARC</a></td><td><a href="https://github.com/Infinitode/OPEN-ARC/blob/main/Project-15/notebook.ipynb">https://github.com/Infinitode/OPEN-ARC/blob/main/Project-15/notebook.ipynb</a></td><td>—</td></tr><tr><td><strong>SRNet-x4-Upscaler</strong></td><td>Image to image (x4 super-resolution)</td><td><a href="https://huggingface.co/InfinitodeLTD/SRNet-x4-Upscaler">https://huggingface.co/InfinitodeLTD/SRNet-x4-Upscaler</a></td><td></td><td>—</td></tr></tbody></table>

### Use a model

{% hint style="info" %}
Open the model card for usage details. Clone its linked repository to inspect or reproduce the notebook implementation.
{% endhint %}

The tabular models load with `joblib` or scikit-learn, so inference is a single call once you have the artifact. The upscaler ships as PyTorch and ONNX; an ONNX build of it runs entirely in the browser as one of our [Experiments](https://infinitode.netlify.app/experiments), with nothing uploaded to a server.

The newest OPEN-ARC artifact is [`InfinitodeLTD/IFSCM-OPEN-ARC`](https://huggingface.co/InfinitodeLTD/IFSCM-OPEN-ARC), a scikit-learn `RandomForestClassifier` for Iris species classification trained in [Project 15](https://github.com/Infinitode/OPEN-ARC/tree/main/Project-15). It ships `iris_classifier.joblib` plus a small `config.json`:

```python
import joblib
import pandas as pd
from huggingface_hub import hf_hub_download

repo_id = "InfinitodeLTD/IFSCM-OPEN-ARC"
hf_hub_download(repo_id=repo_id, filename="config.json")
model_path = hf_hub_download(repo_id=repo_id, filename="iris_classifier.joblib")
model = joblib.load(model_path)

flowers = pd.DataFrame(
    [[5.1, 3.5, 1.4, 0.2], [5.9, 3.0, 4.2, 1.5]],
    columns=["SepalLengthCm", "SepalWidthCm", "PetalLengthCm", "PetalWidthCm"],
)
print(model.predict(flowers))
```

{% hint style="warning" %}
Only load joblib files from sources you trust — deserialization can execute arbitrary code. The IFSCM artifact was serialized with scikit-learn 1.9.1, and cross-version loading is unsupported. Keep the input column order exactly as above.
{% endhint %}

{% hint style="warning" %}
**Nothing filters content inside these files.** ValX-based checks are applied by individual products at their own boundaries — Dungen screens its seed text, Dungen Dev and KaosGen screen their prompts — and the classification and regression models don't screen anything, because they return labels and numbers, not prose. If you host a model yourself, bring your own moderation. ValX is open source and free to run: `pip install valx`.
{% endhint %}

{% hint style="danger" %}
Use AI resources responsibly. Do not use them for illegal activity or harmful content. Review each repository's license before distributing derived work.
{% endhint %}

***

## :package: Datasets

Our curated datasets support research and fine-tuning. Hugging Face hosts the dataset releases. GitHub provides repository mirrors.

<table data-view="cards"><thead><tr><th>View all datasets</th><th data-card-target data-type="content-ref">Hugging Face</th></tr></thead><tbody><tr><td>Browse every Infinitode dataset on Hugging Face.</td><td><a href="https://huggingface.co/datasets?author=InfinitodeLTD">https://huggingface.co/datasets?author=InfinitodeLTD</a></td></tr></tbody></table>

<table data-view="cards"><thead><tr><th>Dataset Name</th><th>Description</th><th data-type="content-ref">HF Link</th><th data-type="content-ref">GitHub Repo</th></tr></thead><tbody><tr><td><strong>PWLDS</strong></td><td>10M+ passwords with strength levels</td><td><a href="https://huggingface.co/datasets/InfinitodeLTD/PWLDS">https://huggingface.co/datasets/InfinitodeLTD/PWLDS</a></td><td><a href="https://github.com/Infinitode/PWLDS">https://github.com/Infinitode/PWLDS</a></td></tr><tr><td><strong>CRSD</strong></td><td>10K+ synthetic customer reviews with sentiment labels</td><td><a href="https://huggingface.co/datasets/InfinitodeLTD/CRSD">https://huggingface.co/datasets/InfinitodeLTD/CRSD</a></td><td><a href="https://github.com/Infinitode/CRSD">https://github.com/Infinitode/CRSD</a></td></tr></tbody></table>

***

## OPEN-ARC

**OPEN-ARC** stands for *Open-source Platform for Engineering Neural Architectures and Research Collaboration*. It is our collaborative AI research initiative: each project pairs a public dataset and a challenge with a commented reference notebook, a leaderboard, and sometimes a pre-trained base model.

* [OPEN-ARC repository](https://github.com/Infinitode/OPEN-ARC) — projects, notebooks, datasets, and artifacts.
* [LEADERBOARD.md](https://github.com/Infinitode/OPEN-ARC/blob/main/LEADERBOARD.md) — per-project rankings.
* [OPEN-ARC website](https://open-arc.netlify.app) — project overview.
* [OPEN-ARC Open Models](https://huggingface.co/collections/InfinitodeLTD/open-arc-open-models) — the Hugging Face collection of published base models.

### Projects

Fifteen numbered challenges are available, covering tabular classification and regression, computer vision, and text generation. Scores below are from the reference implementation's notebook on its own test split.

| #  | Project                                     | Task                   | Reference architecture     | Score               | Notebook                                                                                           |
| -- | ------------------------------------------- | ---------------------- | -------------------------- | ------------------- | -------------------------------------------------------------------------------------------------- |
| 1  | Liver Cirrhosis Stage Classification (LCSC) | Tabular classification | RandomForestClassifier     | 95.6% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-1-LCSC/project-1-lcsc.ipynb)   |
| 2  | Weather Type Classification (WTC)           | Tabular classification | RandomForestClassifier     | 91.2% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-2-WTC/project-2-wtc.ipynb)     |
| 3  | Potato Plant Disease Classification (PPDC)  | Image classification   | Custom CNN                 | 95.1% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-3-PPDC/project-3-ppdc.ipynb)   |
| 4  | Red Wine Quality Classification (RWQC)      | Tabular classification | GradientBoostingClassifier | 72.8% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-4-RWQC/project-4-rwqc.ipynb)   |
| 5  | Terraria Weapon Name Generation (TWNG)      | Text generation        | SimpleRNN                  | 78.6% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-5-TWNG/project-5-twng.ipynb)   |
| 6  | News Headline Generation (NHG)              | Text generation        | DistilBART                 | 52.8 BLEU           | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-6-NHG/project-6-nhg.ipynb)     |
| 7  | Crop Recommendation (CR)                    | Tabular classification | XGBClassifier              | 98.6% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-7-CR/project-7-cr.ipynb)       |
| 8  | Plant Stress Prediction (PSPM)              | Tabular classification | XGBClassifier              | 99.1% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-8-PSPM/project-8-pspm.ipynb)   |
| 9  | Traffic Accident Prediction (TAPM)          | Tabular classification | XGBClassifier              | 85.2% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-9-TAPM/project-9-tapm.ipynb)   |
| 10 | Mushroom Classification (MCM)               | Tabular classification | RandomForestClassifier     | 91.1% accuracy (CV) | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-10-MCM/project-10-mcm.ipynb)   |
| 11 | Basic Personality Prediction (BPPM)         | Tabular classification | XGBClassifier              | 92% accuracy        | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-11-BPPM/project-11-bppm.ipynb) |
| 12 | Spam Mail Classification (SMCM)             | Text classification    | MultinomialNB              | 98.4% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-12/notebook.ipynb)             |
| 13 | India Housing Price Prediction (IHPPM)      | Tabular regression     | CatBoostRegressor          | 3.86k MAE           | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-13/notebook.ipynb)             |
| 14 | Real vs Fake Face Classification            | Image classification   | ResNet18                   | 98.8% accuracy      | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-14/notebook.ipynb)             |
| 15 | Iris Flower Species Classification (IFSCM)  | Tabular classification | RandomForestClassifier     | 100% accuracy       | [Notebook](https://github.com/Infinitode/OPEN-ARC/blob/main/Project-15/notebook.ipynb)             |

Each project README links its Kaggle dataset and includes **Open in Kaggle** and **Open in Colab** buttons, so a notebook can be run without a local environment. Several projects also commit their trained artifact — for example `Project-15/iris_classifier.joblib`, `Project-13/model.cbm`, `Project-10-MCM/mushroom_classifier.pkl`, and `Project-14/novel_fast_face_detector_resnet18.pth`.

### Beginner projects

Three guided notebooks sit alongside the numbered challenges and assume no prior machine-learning experience:

* [Data Cleaning & Text Filtering](https://github.com/Infinitode/OPEN-ARC/blob/main/Beginner-Projects/Project-1-Data-Cleaning-Text-Filtering/notebook.ipynb) — data-quality checks, whitespace cleaning, and moderation with [`valx`](https://github.com/Infinitode/ValX).
* [Exploratory Data Analysis & Visualization](https://github.com/Infinitode/OPEN-ARC/blob/main/Beginner-Projects/Project-2-Exploratory-Data-Analysis/notebook.ipynb) — summary statistics, bar charts, histograms, and box plots.
* [Data Manipulation & Aggregation](https://github.com/Infinitode/OPEN-ARC/blob/main/Beginner-Projects/Project-3-Data-Manipulation-Aggregation/notebook.ipynb) — feature engineering, multi-condition filtering, group-by aggregations, and pivot tables.

### How to participate

1. Pick a project and open its notebook, or start from scratch.
2. Build an implementation, or improve the reference one.
3. Fork the repository and add a row to `LEADERBOARD.md` with your name or alias, architecture type, platform (Kaggle, Colab, or local), and a public link to your notebook or repository.
4. Open a pull request.

{% hint style="success" %}
OPEN-ARC is beginner-friendly. Notebooks are commented, may include pre-trained models, and you do not need to change anything except `LEADERBOARD.md` to take part.
{% endhint %}

### Published models and terms

The OPEN-ARC Hugging Face collection mirrors the project base models, including the newest addition, `IFSCM-OPEN-ARC` from Project 15. Project 14's face classifier and Project 15's joblib artifact also live in the repository itself.

* The repository code and our implementations are MIT licensed, with an additional clause reserving the OPEN-ARC name and branding. See [LICENSE.md](https://github.com/Infinitode/OPEN-ARC/blob/main/LICENSE.md).
* Community contributors' models and implementations remain under their own licenses.
* Every project describes its models as learning, research, and experimental resources, **not intended for commercial use**. Some trained artifacts, including `IFSCM-OPEN-ARC`, have no separate Hugging Face license declared, so check the project README and contact us if you need clarification.
