README.md

---
language: fr
language_name: French
language_family: romance_galloitalic
tags:
  - wikilangs
  - nlp
  - tokenizer
  - embeddings
  - n-gram
  - markov
  - wikipedia
  - feature-extraction
  - sentence-similarity
  - tokenization
  - n-grams
  - markov-chain
  - text-mining
  - fasttext
  - babelvec
  - vocabulous
  - vocabulary
  - monolingual
  - family-romance_galloitalic
license: mit
library_name: wikilangs
pipeline_tag: text-generation
datasets:
  - omarkamali/wikipedia-monthly
dataset_info:
  name: wikipedia-monthly
  description: Monthly snapshots of Wikipedia articles across 300+ languages
metrics:
  - name: best_compression_ratio
    type: compression
    value: 4.573
  - name: best_isotropy
    type: isotropy
    value: 0.7808
  - name: best_alignment_r10
    type: alignment
    value: 0.9680
  - name: vocabulary_size
    type: vocab
    value: 1519124
generated: 2026-03-03
---

# French — Wikilangs Models

Open-source tokenizers, n-gram & Markov language models, vocabulary stats, and word embeddings trained on **French** Wikipedia by [Wikilangs](https://wikilangs.org).

🌐 [Language Page](https://wikilangs.org/languages/fr/) · 🎮 [Playground](https://wikilangs.org/playground/?lang=fr) · 📊 [Full Research Report](RESEARCH_REPORT.md)

## Language Samples

Example sentences drawn from the French Wikipedia corpus:

> Le décane est un alcane linéaire de formule brute qui possède 136 isomères. Ces diverses molécules comportent toutes dix [en grec δέκα (déca)] atomes de carbone. Notes et références linéaire du décane

> Cette liste représente les plus importantes villes de l'Égypte antique ordonnées par nome et suivies des divinités qui y étaient adorées. Basse-Égypte

> L'eicosane est un alcane linéaire de formule brute . Il possède isomères structuraux. Notes et références linéaire

> Hapy se réfère à Hâpi : Génie à tête de singe de la mythologie égyptienne. Hâpy : Dieu du Nil dans la mythologie égyptienne.

> L'heptane ou n-heptane est l'hydrocarbure saturé de la famille des alcanes linéaires de formule CH. Notes et références linéaire de l'heptane

## Quick Start

### Load the Tokenizer

```python
import sentencepiece as spm

sp = spm.SentencePieceProcessor()
sp.Load("fr_tokenizer_32k.model")

text = "Lapon peut désigner : les Samis ; les langues sames ; Lapon, une ville du Soudan"
tokens = sp.EncodeAsPieces(text)
ids    = sp.EncodeAsIds(text)

print(tokens)  # subword pieces
print(ids)     # integer ids

# Decode back
print(sp.DecodeIds(ids))
```

<details>
<summary><b>Tokenization examples (click to expand)</b></summary>

**Sample 1:** `Lapon peut désigner : les Samis ; les langues sames ; Lapon, une ville du Soudan…`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁lap on ▁peut ▁désigner ▁: ▁les ▁sam is ▁; ▁les … (+27 more)` | 37 |
| 16k | `▁lap on ▁peut ▁désigner ▁: ▁les ▁sam is ▁; ▁les … (+27 more)` | 37 |
| 32k | `▁lap on ▁peut ▁désigner ▁: ▁les ▁sam is ▁; ▁les … (+26 more)` | 36 |
| 64k | `▁lap on ▁peut ▁désigner ▁: ▁les ▁sam is ▁; ▁les … (+25 more)` | 35 |

**Sample 2:** `Le pentadécane est un alcane linéaire de formule brute . Il possède 4 347 isomèr…`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁le ▁pent ad éc ane ▁est ▁un ▁al c ane … (+27 more)` | 37 |
| 16k | `▁le ▁pent ad éc ane ▁est ▁un ▁alc ane ▁linéaire … (+22 more)` | 32 |
| 32k | `▁le ▁pent ad éc ane ▁est ▁un ▁alc ane ▁linéaire … (+20 more)` | 30 |
| 64k | `▁le ▁pent ad éc ane ▁est ▁un ▁alcane ▁linéaire ▁de … (+18 more)` | 28 |

**Sample 3:** `L'eicosane est un alcane linéaire de formule brute . Il possède isomères structu…`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁l ' e ic os ane ▁est ▁un ▁al c … (+22 more)` | 32 |
| 16k | `▁l ' e icos ane ▁est ▁un ▁alc ane ▁linéaire … (+16 more)` | 26 |
| 32k | `▁l ' e icos ane ▁est ▁un ▁alc ane ▁linéaire … (+14 more)` | 24 |
| 64k | `▁l ' e icos ane ▁est ▁un ▁alcane ▁linéaire ▁de … (+12 more)` | 22 |

</details>

### Load Word Embeddings

```python
from gensim.models import KeyedVectors

# Aligned embeddings (cross-lingual, mapped to English vector space)
wv = KeyedVectors.load("fr_embeddings_128d_aligned.kv")

similar = wv.most_similar("word", topn=5)
for word, score in similar:
    print(f"  {word}: {score:.3f}")
```

### Load N-gram Model

```python
import pyarrow.parquet as pq

df = pq.read_table("fr_3gram_word.parquet").to_pandas()
print(df.head())
```

## Models Overview

![Performance Dashboard](visualizations/performance_dashboard.png)

| Category | Assets |
|----------|--------|
| Tokenizers | BPE at 8k, 16k, 32k, 64k vocab sizes |
| N-gram models | 2 / 3 / 4 / 5-gram (word & subword) |
| Markov chains | Context 1–5 (word & subword) |
| Embeddings | 32d, 64d, 128d — mono & aligned |
| Vocabulary | Full frequency list + Zipf analysis |
| Statistics | Corpus & model statistics JSON |

## Metrics Summary

| Component | Model | Key Metric | Value |
|-----------|-------|------------|-------|
| Tokenizer | 8k BPE | Compression | 3.72x |
| Tokenizer | 16k BPE | Compression | 4.08x |
| Tokenizer | 32k BPE | Compression | 4.37x |
| Tokenizer | 64k BPE | Compression | 4.57x 🏆 |
| N-gram | 2-gram (subword) | Perplexity | 251 🏆 |
| N-gram | 2-gram (word) | Perplexity | 197,170 |
| N-gram | 3-gram (subword) | Perplexity | 1,988 |
| N-gram | 3-gram (word) | Perplexity | 1,815,223 |
| N-gram | 4-gram (subword) | Perplexity | 11,120 |
| N-gram | 4-gram (word) | Perplexity | 5,518,864 |
| N-gram | 5-gram (subword) | Perplexity | 46,850 |
| N-gram | 5-gram (word) | Perplexity | 4,103,331 |
| Markov | ctx-1 (subword) | Predictability | 0.0% |
| Markov | ctx-1 (word) | Predictability | 9.5% |
| Markov | ctx-2 (subword) | Predictability | 39.6% |
| Markov | ctx-2 (word) | Predictability | 53.5% |
| Markov | ctx-3 (subword) | Predictability | 36.4% |
| Markov | ctx-3 (word) | Predictability | 74.5% |
| Markov | ctx-4 (subword) | Predictability | 34.0% |
| Markov | ctx-4 (word) | Predictability | 87.3% 🏆 |
| Vocabulary | full | Size | 1,519,124 |
| Vocabulary | full | Zipf R² | 0.9927 |
| Embeddings | mono_32d | Isotropy | 0.7808 🏆 |
| Embeddings | mono_64d | Isotropy | 0.7574 |
| Embeddings | mono_128d | Isotropy | 0.6995 |
| Embeddings | aligned_32d | Isotropy | 0.7808 |
| Embeddings | aligned_64d | Isotropy | 0.7574 |
| Embeddings | aligned_128d | Isotropy | 0.6995 |
| Alignment | aligned_32d | R@1 / R@5 / R@10 | 48.2% / 74.8% / 82.4% |
| Alignment | aligned_64d | R@1 / R@5 / R@10 | 70.8% / 89.6% / 94.2% |
| Alignment | aligned_128d | R@1 / R@5 / R@10 | 81.2% / 93.4% / 96.8% 🏆 |

📊 **[Full ablation study, per-model breakdowns, and interpretation guide →](RESEARCH_REPORT.md)**

---

## About

Trained on [wikipedia-monthly](https://huggingface.co/datasets/omarkamali/wikipedia-monthly) — monthly snapshots of 300+ Wikipedia languages.

A project by **[Wikilangs](https://wikilangs.org)** · Maintainer: [Omar Kamali](https://omarkamali.com) · [Omneity Labs](https://omneitylabs.com)

### Citation

```bibtex
@misc{wikilangs2025,
  author    = {Kamali, Omar},
  title     = {Wikilangs: Open NLP Models for Wikipedia Languages},
  year      = {2025},
  doi       = {10.5281/zenodo.18073153},
  publisher = {Zenodo},
  url       = {https://huggingface.co/wikilangs},
  institution = {Omneity Labs}
}
```

### Links

- 🌐 [wikilangs.org](https://wikilangs.org)
- 🌍 [Language page](https://wikilangs.org/languages/fr/)
- 🎮 [Playground](https://wikilangs.org/playground/?lang=fr)
- 🤗 [HuggingFace models](https://huggingface.co/wikilangs)
- 📊 [wikipedia-monthly dataset](https://huggingface.co/datasets/omarkamali/wikipedia-monthly)
- 👤 [Omar Kamali](https://huggingface.co/omarkamali)
- 🤝 Sponsor: [Featherless AI](https://featherless.ai)

**License:** MIT — free for academic and commercial use.

---
*Generated by Wikilangs Pipeline · 2026-03-03 05:41:40*