Upload all models and assets for pnt (latest)

.gitattributes

@@ -33,3 +33,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+visualizations/embedding_similarity.png filter=lfs diff=lfs merge=lfs -text
+visualizations/embedding_tsne_multilingual.png filter=lfs diff=lfs merge=lfs -text
+visualizations/performance_dashboard.png filter=lfs diff=lfs merge=lfs -text
+visualizations/tsne_sentences.png filter=lfs diff=lfs merge=lfs -text
+visualizations/tsne_words.png filter=lfs diff=lfs merge=lfs -text
+visualizations/zipf_law.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,762 @@
+---
+language: pnt
+language_name: Pontic
+language_family: greek
+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-greek
+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: 3.670
+  - name: best_isotropy
+    type: isotropy
+    value: 0.0523
+  - name: vocabulary_size
+    type: vocab
+    value: 0
+generated: 2026-01-10
+---
+
+# Pontic - Wikilangs Models
+## Comprehensive Research Report & Full Ablation Study
+
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **Pontic** Wikipedia data.
+We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and word embeddings.
+
+## 📋 Repository Contents
+
+### Models & Assets
+
+- Tokenizers (8k, 16k, 32k, 64k)
+- N-gram models (2, 3, 4, 5-gram)
+- Markov chains (context of 1, 2, 3, 4 and 5)
+- Subword N-gram and Markov chains
+- Embeddings in various sizes and dimensions (aligned and unaligned)
+- Language Vocabulary
+- Language Statistics
+
+![Performance Dashboard](visualizations/performance_dashboard.png)
+
+### Analysis and Evaluation
+
+- [1. Tokenizer Evaluation](#1-tokenizer-evaluation)
+- [2. N-gram Model Evaluation](#2-n-gram-model-evaluation)
+- [3. Markov Chain Evaluation](#3-markov-chain-evaluation)
+- [4. Vocabulary Analysis](#4-vocabulary-analysis)
+- [5. Word Embeddings Evaluation](#5-word-embeddings-evaluation)
+- [6. Morphological Analysis (Experimental)](#6--morphological-analysis-experimental)
+- [7. Summary & Recommendations](#7-summary--recommendations)
+- [Metrics Glossary](#appendix-metrics-glossary--interpretation-guide)
+- [Visualizations Index](#visualizations-index)
+
+---
+## 1. Tokenizer Evaluation
+
+![Tokenizer Compression](visualizations/tokenizer_compression.png)
+
+![Tokenizer Fertility](visualizations/tokenizer_fertility.png)
+
+![Tokenizer OOV](visualizations/tokenizer_oov.png)
+
+![Total Tokens](visualizations/tokenizer_total_tokens.png)
+
+### Results
+
+| Vocab Size | Compression | Avg Token Len | UNK Rate | Total Tokens |
+|------------|-------------|---------------|----------|--------------|
+| **8k** | 3.197x | 3.20 | 0.1329% | 100,820 |
+| **16k** | 3.540x | 3.55 | 0.1472% | 91,057 |
+| **32k** | 3.670x 🏆 | 3.68 | 0.1526% | 87,822 |
+
+### Tokenization Examples
+
+Below are sample sentences tokenized with each vocabulary size:
+
+**Sample 1:** `Η Οσάκα εν πολιτεία σην Ιαπωνίαν. Οσήμερον εχ' πληθυσμόν 2.668.586 ανθρώπ.`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁η ▁οσάκα ▁εν ▁πολιτεία ▁σην ▁ιαπωνίαν . ▁οσήμερον ▁εχ ' ... (+13 more)` | 23 |
+| 16k | `▁η ▁οσάκα ▁εν ▁πολιτεία ▁σην ▁ιαπωνίαν . ▁οσήμερον ▁εχ ' ... (+13 more)` | 23 |
+| 32k | `▁η ▁οσάκα ▁εν ▁πολιτεία ▁σην ▁ιαπωνίαν . ▁οσήμερον ▁εχ ' ... (+13 more)` | 23 |
+
+**Sample 2:** `Ο Βόλος εν πόλην τρανή (τρανότερη της Μαγνησίας ούλης) παρά την θάλασσαν ασο κέν...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁ο ▁βόλοσ ▁εν ▁πόλην ▁τραν ή ▁( τ ραν ότερη ... (+26 more)` | 36 |
+| 16k | `▁ο ▁βόλοσ ▁εν ▁πόλην ▁τρανή ▁( τρανότερη ▁τησ ▁μαγνησίασ ▁ούλησ ... (+20 more)` | 30 |
+| 32k | `▁ο ▁βόλοσ ▁εν ▁πόλην ▁τρανή ▁( τρανότερη ▁τησ ▁μαγνησίασ ▁ούλησ ... (+20 more)` | 30 |
+
+**Sample 3:** `Η Λα Ροσέλ εν πολιτεία σην Γαλλίαν. Οσήμερον εχ' πληθυσμόν 74.998 ανθρώπ.`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁η ▁λα ▁ρ οσ έλ ▁εν ▁πολιτεία ▁σην ▁γαλλίαν . ... (+13 more)` | 23 |
+| 16k | `▁η ▁λα ▁ροσέλ ▁εν ▁πολιτεία ▁σην ▁γαλλίαν . ▁οσήμερον ▁εχ ... (+11 more)` | 21 |
+| 32k | `▁η ▁λα ▁ροσέλ ▁εν ▁πολιτεία ▁σην ▁γαλλίαν . ▁οσήμερον ▁εχ ... (+11 more)` | 21 |
+
+
+### Key Findings
+
+- **Best Compression:** 32k achieves 3.670x compression
+- **Lowest UNK Rate:** 8k with 0.1329% unknown tokens
+- **Trade-off:** Larger vocabularies improve compression but increase model size
+- **Recommendation:** 32k vocabulary provides optimal balance for production use
+
+---
+## 2. N-gram Model Evaluation
+
+![N-gram Perplexity](visualizations/ngram_perplexity.png)
+
+![N-gram Unique](visualizations/ngram_unique.png)
+
+![N-gram Coverage](visualizations/ngram_coverage.png)
+
+### Results
+
+| N-gram | Variant | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
+|--------|---------|------------|---------|----------------|------------------|-------------------|
+| **2-gram** | Word | 378 | 8.56 | 778 | 59.1% | 100.0% |
+| **2-gram** | Subword | 411 | 8.68 | 1,785 | 57.2% | 97.4% |
+| **3-gram** | Word | 302 🏆 | 8.24 | 829 | 68.5% | 100.0% |
+| **3-gram** | Subword | 2,487 | 11.28 | 9,341 | 25.0% | 68.3% |
+| **4-gram** | Word | 459 | 8.84 | 1,665 | 62.9% | 89.3% |
+| **4-gram** | Subword | 7,482 | 12.87 | 24,840 | 13.4% | 46.9% |
+| **5-gram** | Word | 321 | 8.33 | 1,188 | 70.5% | 96.3% |
+| **5-gram** | Subword | 11,940 | 13.54 | 32,544 | 9.4% | 38.6% |
+
+### Top 5 N-grams by Size
+
+**2-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `τη χρονίας` | 207 |
+| 2 | `για να` | 153 |
+| 3 | `το γρηγοριανόν` | 134 |
+| 4 | `γρηγοριανόν ημερολόγιον` | 134 |
+| 5 | `2 3` | 133 |
+
+**3-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `το γρηγοριανόν ημερολόγιον` | 133 |
+| 2 | `2 3 4` | 132 |
+| 3 | `15 16 17` | 131 |
+| 4 | `17 18 19` | 131 |
+| 5 | `9 10 11` | 131 |
+
+**4-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `25 26 27 28` | 131 |
+| 2 | `9 10 11 12` | 131 |
+| 3 | `10 11 12 13` | 131 |
+| 4 | `3 4 5 6` | 131 |
+| 5 | `11 12 13 14` | 131 |
+
+**5-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `8 9 10 11 12` | 131 |
+| 2 | `10 11 12 13 14` | 131 |
+| 3 | `11 12 13 14 15` | 131 |
+| 4 | `12 13 14 15 16` | 131 |
+| 5 | `13 14 15 16 17` | 131 |
+
+**2-grams (Subword):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `ν _` | 10,498 |
+| 2 | `_ τ` | 7,042 |
+| 3 | `α ν` | 5,192 |
+| 4 | `α _` | 4,454 |
+| 5 | `ς _` | 4,288 |
+
+**3-grams (Subword):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `ο ν _` | 2,539 |
+| 2 | `α ν _` | 2,368 |
+| 3 | `_ τ η` | 2,160 |
+| 4 | `_ κ α` | 1,899 |
+| 5 | `_ τ ο` | 1,888 |
+
+**4-grams (Subword):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `_ τ η _` | 1,677 |
+| 2 | `_ κ α ι` | 1,214 |
+| 3 | `κ α ι _` | 1,179 |
+| 4 | `_ τ ο _` | 1,174 |
+| 5 | `_ ε ν _` | 752 |
+
+**5-grams (Subword):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `_ κ α ι _` | 1,176 |
+| 2 | `ν _ τ η _` | 594 |
+| 3 | `_ χ ρ ο ν` | 536 |
+| 4 | `ι κ ό ν _` | 527 |
+| 5 | `χ ρ ο ν ί` | 523 |
+
+
+### Key Findings
+
+- **Best Perplexity:** 3-gram (word) with 302
+- **Entropy Trend:** Decreases with larger n-grams (more predictable)
+- **Coverage:** Top-1000 patterns cover ~39% of corpus
+- **Recommendation:** 4-gram or 5-gram for best predictive performance
+
+---
+## 3. Markov Chain Evaluation
+
+![Markov Entropy](visualizations/markov_entropy.png)
+
+![Markov Contexts](visualizations/markov_contexts.png)
+
+![Markov Branching](visualizations/markov_branching.png)
+
+### Results
+
+| Context | Variant | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
+|---------|---------|-------------|------------|------------------|-----------------|----------------|
+| **1** | Word | 0.4938 | 1.408 | 2.69 | 11,667 | 50.6% |
+| **1** | Subword | 1.2400 | 2.362 | 8.79 | 439 | 0.0% |
+| **2** | Word | 0.1458 | 1.106 | 1.25 | 31,022 | 85.4% |
+| **2** | Subword | 1.0962 | 2.138 | 5.25 | 3,856 | 0.0% |
+| **3** | Word | 0.0419 | 1.029 | 1.07 | 38,349 | 95.8% |
+| **3** | Subword | 0.6896 | 1.613 | 2.67 | 20,216 | 31.0% |
+| **4** | Word | 0.0219 🏆 | 1.015 | 1.04 | 40,551 | 97.8% |
+| **4** | Subword | 0.3670 | 1.290 | 1.70 | 53,958 | 63.3% |
+
+### Generated Text Samples (Word-based)
+
+Below are text samples generated from each word-based Markov chain model:
+
+**Context Size 1:**
+
+1. `τη γαλλίαν οσήμερον εχ πληθυσμόν τη λέξην encyclopaedia αέτς άμον ντο εποίκαν χάταλα δίδυμα σα commo...`
+2. `το ποτάμ λέχκουνταν doğu karadeniz dağları άμα ευρίσκουμε και φιλόλογον γιόχαν βόλφγκανγκ μπέριςς έβ...`
+3. `και ο πληθυσμόν σα δυτικά πίσω σην πολωνίαν την ρώμην επέμνεν έναν συνεργασίαν ντο ευρίεται σο`
+
+**Context Size 2:**
+
+1. `για να τελούτεν η χρονία 363 ημέρας για να εγροικάτεν κι εσείς αούτο εν το γράψιμον το`
+2. `τη χρονίας ίστε λεει μας το γρηγοριανόν ημερολόγιον επέμναν άλλα 360 ημέρας σο δίσεκτον τη χρονία ατ...`
+3. `το γρηγοριανόν ημερολόγιον κι εχ 31 ημέρας σ αβούτον το κράτος έχεις ως επίσημον λαλίαν την καζακική...`
+
+**Context Size 3:**
+
+1. `το γρηγοριανόν ημερολόγιον επέμναν άλλα 348 ημέρας για να τελούτεν η χρονία ατά ντ εγένταν εγεννέθαν...`
+2. `2 3 4 5 6 u 7 u 8 9 10 11 12 13 14 15 16 17`
+3. `21 22 23 24 25 26 27 28 29 30 31 28 τρυγομηνά εν 301ον ημέρα τη χρονίας`
+
+**Context Size 4:**
+
+1. `10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28`
+2. `25 26 27 28 29 30 31 η 1 τη καλανταρί εν 1ον ημέρα τη χρονίας άμον ντο λεει`
+3. `19 20 21 22 23 24 25 26 27 28 29 30 31 η 7 τη καλανταρί εν 7ον`
+
+
+### Generated Text Samples (Subword-based)
+
+Below are text samples generated from each subword-based Markov chain model:
+
+**Context Size 1:**
+
+1. `_σ'_25_τοκ_σία):`
+2. `αυν_ην_ουμβοξυρό`
+3. `ν_η_ρίαι_δαιο_να`
+
+**Context Size 2:**
+
+1. `ν_απερ,_ο_γιαφέρα`
+2. `_τηλεία_βικος_ατ'`
+3. `αν_τη_μορί_|_4_5_`
+
+**Context Size 3:**
+
+1. `ον_ο_άλλα_20_21_22`
+2. `αν_τη_βερολόγιον,_`
+3. `_τη_ασίαν_η_χώρας_`
+
+**Context Size 4:**
+
+1. `_τη_χρονίαν_σην_και`
+2. `_και_διαδικτυακήν_σ`
+3. `και_κάτω_από_τιμόρ.`
+
+
+### Key Findings
+
+- **Best Predictability:** Context-4 (word) with 97.8% predictability
+- **Branching Factor:** Decreases with context size (more deterministic)
+- **Memory Trade-off:** Larger contexts require more storage (53,958 contexts)
+- **Recommendation:** Context-3 or Context-4 for text generation
+
+---
+## 4. Vocabulary Analysis
+
+![Zipf's Law](visualizations/zipf_law.png)
+
+![Top Words](visualizations/top20_words.png)
+
+![Coverage Curve](visualizations/vocab_coverage.png)
+
+### Statistics
+
+| Metric | Value |
+|--------|-------|
+| Vocabulary Size | 3,936 |
+| Total Tokens | 45,584 |
+| Mean Frequency | 11.58 |
+| Median Frequency | 3 |
+| Frequency Std Dev | 56.78 |
+
+### Most Common Words
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | τη | 1,685 |
+| 2 | το | 1,240 |
+| 3 | και | 1,182 |
+| 4 | η | 1,115 |
+| 5 | ο | 813 |
+| 6 | εν | 783 |
+| 7 | τ | 746 |
+| 8 | σα | 652 |
+| 9 | τα | 572 |
+| 10 | σο | 475 |
+
+### Least Common Words (from vocabulary)
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | μπασάν | 2 |
+| 2 | επάτησεν | 2 |
+| 3 | υόρκην | 2 |
+| 4 | τραγωδός | 2 |
+| 5 | πρόσωπα | 2 |
+| 6 | μούζικας | 2 |
+| 7 | born | 2 |
+| 8 | rolling | 2 |
+| 9 | stone | 2 |
+| 10 | δείχν | 2 |
+
+### Zipf's Law Analysis
+
+| Metric | Value |
+|--------|-------|
+| Zipf Coefficient | 0.9722 |
+| R² (Goodness of Fit) | 0.971817 |
+| Adherence Quality | **excellent** |
+
+### Coverage Analysis
+
+| Top N Words | Coverage |
+|-------------|----------|
+| Top 100 | 54.0% |
+| Top 1,000 | 83.1% |
+| Top 5,000 | 0.0% |
+| Top 10,000 | 0.0% |
+
+### Key Findings
+
+- **Zipf Compliance:** R²=0.9718 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 54.0% of corpus
+- **Long Tail:** -6,064 words needed for remaining 100.0% coverage
+
+---
+## 5. Word Embeddings Evaluation
+
+![Embedding Isotropy](visualizations/embedding_isotropy.png)
+
+![Similarity Matrix](visualizations/embedding_similarity.png)
+
+![t-SNE Words](visualizations/tsne_words.png)
+
+![t-SNE Sentences](visualizations/tsne_sentences.png)
+
+
+### 5.1 Cross-Lingual Alignment
+
+![Alignment Quality](visualizations/embedding_alignment_quality.png)
+
+![Multilingual t-SNE](visualizations/embedding_tsne_multilingual.png)
+
+
+### 5.2 Model Comparison
+
+| Model | Dimension | Isotropy | Semantic Density | Alignment R@1 | Alignment R@10 |
+|-------|-----------|----------|------------------|---------------|----------------|
+| **mono_32d** | 32 | 0.0523 | 0.6492 | N/A | N/A |
+| **mono_64d** | 64 | 0.0087 | 0.6629 | N/A | N/A |
+| **mono_128d** | 128 | 0.0013 | 0.6851 | N/A | N/A |
+| **aligned_32d** | 32 | 0.0523 🏆 | 0.6519 | 0.0429 | 0.3286 |
+| **aligned_64d** | 64 | 0.0087 | 0.6348 | 0.0357 | 0.3571 |
+| **aligned_128d** | 128 | 0.0013 | 0.6853 | 0.0500 | 0.4143 |
+
+### Key Findings
+
+- **Best Isotropy:** aligned_32d with 0.0523 (more uniform distribution)
+- **Semantic Density:** Average pairwise similarity of 0.6615. Lower values indicate better semantic separation.
+- **Alignment Quality:** Aligned models achieve up to 5.0% R@1 in cross-lingual retrieval.
+- **Recommendation:** 128d aligned for best cross-lingual performance
+
+---
+## 6.  Morphological Analysis (Experimental)
+
+This section presents an automated morphological analysis derived from the statistical divergence between word-level and subword-level models. By analyzing where subword predictability spikes and where word-level coverage fails, we can infer linguistic structures without supervised data.
+
+### 6.1 Productivity & Complexity
+
+| Metric | Value | Interpretation | Recommendation |
+|--------|-------|----------------|----------------|
+| Productivity Index | **5.000** | High morphological productivity | Reliable analysis |
+| Idiomaticity Gap | **1.493** | High formulaic/idiomatic content | - |
+
+### 6.2 Affix Inventory (Productive Units)
+
+These are the most productive prefixes and suffixes identified by sampling the vocabulary for global substitutability patterns. A unit is considered an affix if stripping it leaves a valid stem that appears in other contexts.
+
+#### Productive Prefixes
+| Prefix | Examples |
+|--------|----------|
+| `-ε` | εύκρατο, εγκυκλοπαίδεια, εταιρία |
+| `-α` | αλβανίας, αρχηγοί, αρχαία |
+| `-κ` | κόκκινον, κοσμικοί, καλατσεύ |
+| `-π` | πρώτον, περσικόν, προξενείον |
+| `-σ` | συστέματα, σάββατον, συγκεκριμένα |
+| `-κα` | καλατσεύ, κατάληξη, καζίνο |
+| `-πα` | παπάδες, πατέρα, παράδειγμα |
+| `-επ` | επιτροπή, εποίν, επίσης |
+
+#### Productive Suffixes
+| Suffix | Examples |
+|--------|----------|
+| `-ν` | οποίον, έρτεν, τιράνων |
+| `-ς` | χερσόνησος, αλβανίας, λέξεις |
+| `-αν` | ρωμανίαν, ίαν, μάναν |
+| `-α` | συστέματα, ρωμαίικα, εγκυκλοπαίδεια |
+| `-ον` | οποίον, κόκκινον, πρώτον |
+| `-ας` | αλβανίας, εργασίας, γαλλίας |
+| `-όν` | περσικόν, σκανδιναβικόν, εθνικόν |
+| `-ος` | χερσόνησος, τέλος, μύρος |
+
+### 6.3 Bound Stems (Lexical Roots)
+
+Bound stems are high-frequency subword units that are semantically cohesive but rarely appear as standalone words. These often correspond to the 'core' of a word that requires inflection or derivation to be valid.
+
+| Stem | Cohesion | Substitutability | Examples |
+|------|----------|------------------|----------|
+| `ικόν` | 1.32x | 17 contexts | υλικόν, εικόνα, ενικόν |
+| `ματα` | 1.41x | 10 contexts | θέματα, βήματα, ρήματα |
+| `εται` | 1.37x | 5 contexts | ψέεται, ίνεται, λύεται |
+
+### 6.4 Affix Compatibility (Co-occurrence)
+
+This table shows which prefixes and suffixes most frequently co-occur on the same stems, revealing the 'stacking' rules of the language's morphology.
+
+| Prefix | Suffix | Frequency | Examples |
+|--------|--------|-----------|----------|
+| `-ε` | `-ν` | 127 words | εποίν, εγεννέθαν |
+| `-α` | `-ν` | 79 words | αν, ασίαν |
+| `-π` | `-ν` | 65 words | πρώτον, περσικόν |
+| `-α` | `-ς` | 60 words | αλβανίας, αχουλής |
+| `-κ` | `-ν` | 56 words | κόκκινον, κρύον |
+| `-ε` | `-εν` | 54 words | εδέβεν, εφτάτεν |
+| `-σ` | `-ν` | 49 words | σάββατον, σκανδιναβικόν |
+| `-π` | `-ς` | 44 words | ποδοσφαιριστής, παπάδες |
+| `-κ` | `-ς` | 41 words | καραμανλής, κερτς |
+| `-ε` | `-αν` | 40 words | εγεννέθαν, επηρεάσταν |
+
+### 6.5 Recursive Morpheme Segmentation
+
+Using **Recursive Hierarchical Substitutability**, we decompose complex words into their constituent morphemes. This approach handles nested affixes (e.g., `prefix-prefix-root-suffix`).
+
+| Word | Suggested Split | Confidence | Stem |
+|------|-----------------|------------|------|
+| γράψιματα | **`γράψιμ-α-τα`** | 7.5 | `α` |
+| εποίκενατα | **`εποίκεν-α-τα`** | 7.5 | `α` |
+| λιθουανία | **`λιθου-αν-ία`** | 7.5 | `αν` |
+| ανθρώπους | **`ανθρώπ-ου-ς`** | 6.0 | `ανθρώπ` |
+| καλαντάρτς | **`καλαντάρτ-ς`** | 4.5 | `καλαντάρτ` |
+| βιογραφικόν | **`βιογραφικό-ν`** | 4.5 | `βιογραφικό` |
+| υπολογιστήν | **`υπολογιστή-ν`** | 4.5 | `υπολογιστή` |
+| σημαντικόν | **`σημαντικό-ν`** | 4.5 | `σημαντικό` |
+| καλομηνάς | **`καλομηνά-ς`** | 4.5 | `καλομηνά` |
+| ξεχωριστόν | **`ξεχωριστό-ν`** | 4.5 | `ξεχωριστό` |
+| περιοδικόν | **`περιοδικό-ν`** | 4.5 | `περιοδικό` |
+| θεσσαλονίκης | **`θεσσαλονίκη-ς`** | 4.5 | `θεσσαλονίκη` |
+| ερχίνεσεν | **`ερχίνεσε-ν`** | 4.5 | `ερχίνεσε` |
+| συνορεύνε | **`συνορεύ-νε`** | 4.5 | `συνορεύ` |
+| συνδέζμαι | **`συνδέζμ-αι`** | 4.5 | `συνδέζμ` |
+
+### 6.6 Linguistic Interpretation
+
+> **Automated Insight:**
+The language Pontic shows high morphological productivity. The subword models are significantly more efficient than word models, suggesting a rich system of affixation or compounding.
+
+> **Note on Idiomaticity:** The high Idiomaticity Gap suggests a large number of frequent multi-word expressions or formulaic sequences that are statistically distinct from their component parts.
+
+---
+## 7. Summary & Recommendations
+
+![Performance Dashboard](visualizations/performance_dashboard.png)
+
+### Production Recommendations
+
+| Component | Recommended | Rationale |
+|-----------|-------------|-----------|
+| Tokenizer | **32k BPE** | Best compression (3.67x) |
+| N-gram | **3-gram** | Lowest perplexity (302) |
+| Markov | **Context-4** | Highest predictability (97.8%) |
+| Embeddings | **100d** | Balanced semantic capture and isotropy |
+
+
+---
+## Appendix: Metrics Glossary & Interpretation Guide
+
+This section provides definitions, intuitions, and guidance for interpreting the metrics used throughout this report.
+
+### Tokenizer Metrics
+
+**Compression Ratio**
+> *Definition:* The ratio of characters to tokens (chars/token). Measures how efficiently the tokenizer represents text.
+>
+> *Intuition:* Higher compression means fewer tokens needed to represent the same text, reducing sequence lengths for downstream models. A 3x compression means ~3 characters per token on average.
+>
+> *What to seek:* Higher is generally better for efficiency, but extremely high compression may indicate overly aggressive merging that loses morphological information.
+
+**Average Token Length (Fertility)**
+> *Definition:* Mean number of characters per token produced by the tokenizer.
+>
+> *Intuition:* Reflects the granularity of tokenization. Longer tokens capture more context but may struggle with rare words; shorter tokens are more flexible but increase sequence length.
+>
+> *What to seek:* Balance between 2-5 characters for most languages. Arabic/morphologically-rich languages may benefit from slightly longer tokens.
+
+**Unknown Token Rate (OOV Rate)**
+> *Definition:* Percentage of tokens that map to the unknown/UNK token, indicating words the tokenizer cannot represent.
+>
+> *Intuition:* Lower OOV means better vocabulary coverage. High OOV indicates the tokenizer encounters many unseen character sequences.
+>
+> *What to seek:* Below 1% is excellent; below 5% is acceptable. BPE tokenizers typically achieve very low OOV due to subword fallback.
+
+### N-gram Model Metrics
+
+**Perplexity**
+> *Definition:* Measures how "surprised" the model is by test data. Mathematically: 2^(cross-entropy). Lower values indicate better prediction.
+>
+> *Intuition:* If perplexity is 100, the model is as uncertain as if choosing uniformly among 100 options at each step. A perplexity of 10 means effectively choosing among 10 equally likely options.
+>
+> *What to seek:* Lower is better. Perplexity decreases with larger n-grams (more context). Values vary widely by language and corpus size.
+
+**Entropy**
+> *Definition:* Average information content (in bits) needed to encode the next token given the context. Related to perplexity: perplexity = 2^entropy.
+>
+> *Intuition:* High entropy means high uncertainty/randomness; low entropy means predictable patterns. Natural language typically has entropy between 1-4 bits per character.
+>
+> *What to seek:* Lower entropy indicates more predictable text patterns. Entropy should decrease as n-gram size increases.
+
+**Coverage (Top-K)**
+> *Definition:* Percentage of corpus occurrences explained by the top K most frequent n-grams.
+>
+> *Intuition:* High coverage with few patterns indicates repetitive/formulaic text; low coverage suggests diverse vocabulary usage.
+>
+> *What to seek:* Depends on use case. For language modeling, moderate coverage (40-60% with top-1000) is typical for natural text.
+
+### Markov Chain Metrics
+
+**Average Entropy**
+> *Definition:* Mean entropy across all contexts, measuring average uncertainty in next-word prediction.
+>
+> *Intuition:* Lower entropy means the model is more confident about what comes next. Context-1 has high entropy (many possible next words); Context-4 has low entropy (few likely continuations).
+>
+> *What to seek:* Decreasing entropy with larger context sizes. Very low entropy (<0.1) indicates highly deterministic transitions.
+
+**Branching Factor**
+> *Definition:* Average number of unique next tokens observed for each context.
+>
+> *Intuition:* High branching = many possible continuations (flexible but uncertain); low branching = few options (predictable but potentially repetitive).
+>
+> *What to seek:* Branching factor should decrease with context size. Values near 1.0 indicate nearly deterministic chains.
+
+**Predictability**
+> *Definition:* Derived metric: (1 - normalized_entropy) × 100%. Indicates how deterministic the model's predictions are.
+>
+> *Intuition:* 100% predictability means the next word is always certain; 0% means completely random. Real text falls between these extremes.
+>
+> *What to seek:* Higher predictability for text generation quality, but too high (>98%) may produce repetitive output.
+
+### Vocabulary & Zipf's Law Metrics
+
+**Zipf's Coefficient**
+> *Definition:* The slope of the log-log plot of word frequency vs. rank. Zipf's law predicts this should be approximately -1.
+>
+> *Intuition:* A coefficient near -1 indicates the corpus follows natural language patterns where a few words are very common and most words are rare.
+>
+> *What to seek:* Values between -0.8 and -1.2 indicate healthy natural language distribution. Deviations may suggest domain-specific or artificial text.
+
+**R² (Coefficient of Determination)**
+> *Definition:* Measures how well the linear fit explains the frequency-rank relationship. Ranges from 0 to 1.
+>
+> *Intuition:* R² near 1.0 means the data closely follows Zipf's law; lower values indicate deviation from expected word frequency patterns.
+>
+> *What to seek:* R² > 0.95 is excellent; > 0.99 indicates near-perfect Zipf adherence typical of large natural corpora.
+
+**Vocabulary Coverage**
+> *Definition:* Cumulative percentage of corpus tokens accounted for by the top N words.
+>
+> *Intuition:* Shows how concentrated word usage is. If top-100 words cover 50% of text, the corpus relies heavily on common words.
+>
+> *What to seek:* Top-100 covering 30-50% is typical. Higher coverage indicates more repetitive text; lower suggests richer vocabulary.
+
+### Word Embedding Metrics
+
+**Isotropy**
+> *Definition:* Measures how uniformly distributed vectors are in the embedding space. Computed as the ratio of minimum to maximum singular values.
+>
+> *Intuition:* High isotropy (near 1.0) means vectors spread evenly in all directions; low isotropy means vectors cluster in certain directions, reducing expressiveness.
+>
+> *What to seek:* Higher isotropy generally indicates better-quality embeddings. Values > 0.1 are reasonable; > 0.3 is good. Lower-dimensional embeddings tend to have higher isotropy.
+
+**Average Norm**
+> *Definition:* Mean magnitude (L2 norm) of word vectors in the embedding space.
+>
+> *Intuition:* Indicates the typical "length" of vectors. Consistent norms suggest stable training; high variance may indicate some words are undertrained.
+>
+> *What to seek:* Relatively consistent norms across models. The absolute value matters less than consistency (low std deviation).
+
+**Cosine Similarity**
+> *Definition:* Measures angular similarity between vectors, ranging from -1 (opposite) to 1 (identical direction).
+>
+> *Intuition:* Words with similar meanings should have high cosine similarity. This is the standard metric for semantic relatedness in embeddings.
+>
+> *What to seek:* Semantically related words should score > 0.5; unrelated words should be near 0. Synonyms often score > 0.7.
+
+**t-SNE Visualization**
+> *Definition:* t-Distributed Stochastic Neighbor Embedding - a dimensionality reduction technique that preserves local structure for visualization.
+>
+> *Intuition:* Clusters in t-SNE plots indicate groups of semantically related words. Spread indicates vocabulary diversity; tight clusters suggest semantic coherence.
+>
+> *What to seek:* Meaningful clusters (e.g., numbers together, verbs together). Avoid over-interpreting distances - t-SNE preserves local, not global, structure.
+
+### General Interpretation Guidelines
+
+1. **Compare within model families:** Metrics are most meaningful when comparing models of the same type (e.g., 8k vs 64k tokenizer).
+2. **Consider trade-offs:** Better performance on one metric often comes at the cost of another (e.g., compression vs. OOV rate).
+3. **Context matters:** Optimal values depend on downstream tasks. Text generation may prioritize different metrics than classification.
+4. **Corpus influence:** All metrics are influenced by corpus characteristics. Wikipedia text differs from social media or literature.
+5. **Language-specific patterns:** Morphologically rich languages (like Arabic) may show different optimal ranges than analytic languages.
+
+
+### Visualizations Index
+
+| Visualization | Description |
+|---------------|-------------|
+| Tokenizer Compression | Compression ratios by vocabulary size |
+| Tokenizer Fertility | Average token length by vocabulary |
+| Tokenizer OOV | Unknown token rates |
+| Tokenizer Total Tokens | Total tokens by vocabulary |
+| N-gram Perplexity | Perplexity by n-gram size |
+| N-gram Entropy | Entropy by n-gram size |
+| N-gram Coverage | Top pattern coverage |
+| N-gram Unique | Unique n-gram counts |
+| Markov Entropy | Entropy by context size |
+| Markov Branching | Branching factor by context |
+| Markov Contexts | Unique context counts |
+| Zipf's Law | Frequency-rank distribution with fit |
+| Vocab Frequency | Word frequency distribution |
+| Top 20 Words | Most frequent words |
+| Vocab Coverage | Cumulative coverage curve |
+| Embedding Isotropy | Vector space uniformity |
+| Embedding Norms | Vector magnitude distribution |
+| Embedding Similarity | Word similarity heatmap |
+| Nearest Neighbors | Similar words for key terms |
+| t-SNE Words | 2D word embedding visualization |
+| t-SNE Sentences | 2D sentence embedding visualization |
+| Position Encoding | Encoding method comparison |
+| Model Sizes | Storage requirements |
+| Performance Dashboard | Comprehensive performance overview |
+
+---
+## About This Project
+
+### Data Source
+
+Models trained on [wikipedia-monthly](https://huggingface.co/datasets/omarkamali/wikipedia-monthly) - a monthly snapshot of Wikipedia articles across 300+ languages.
+
+### Project
+
+A project by **[Wikilangs](https://wikilangs.org)** - Open-source NLP models for every Wikipedia language.
+
+### Maintainer
+
+[Omar Kamali](https://omarkamali.com) - [Omneity Labs](https://omneitylabs.com)
+
+### Citation
+
+If you use these models in your research, please cite:
+
+```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}
+}
+```
+
+### License
+
+MIT License - Free for academic and commercial use.
+
+### Links
+
+- 🌐 Website: [wikilangs.org](https://wikilangs.org)
+- 🤗 Models: [huggingface.co/wikilangs](https://huggingface.co/wikilangs)
+- 📊 Data: [wikipedia-monthly](https://huggingface.co/datasets/omarkamali/wikipedia-monthly)
+- 👤 Author: [Omar Kamali](https://huggingface.co/omarkamali)
+- 🤝 Sponsor: [Featherless AI](https://featherless.ai)
+---
+*Generated by Wikilangs Models Pipeline*
+
+*Report Date: 2026-01-10 18:08:17*

models/embeddings/aligned/pnt_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5ae2c0f72a238311b78cfe7a7f0c2f653b1c6850bf745081c02f5181e757469
+size 1025048284

models/embeddings/aligned/pnt_128d.meta.json

@@ -0,0 +1 @@
+{"lang": "pnt", "dim": 128, "max_seq_len": 512, "is_aligned": true}

models/embeddings/aligned/pnt_128d.projection.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:226bb9d48701430ff4d830b6e2d75f27943809979b266e963b9822463030b05f
+size 65664

models/embeddings/aligned/pnt_128d_metadata.json

@@ -0,0 +1,8 @@
+{
+  "language": "pnt",
+  "dimension": 128,
+  "version": "aligned",
+  "hub_language": "en",
+  "seed_vocab_size": 140,
+  "vocab_size": 1003
+}

models/embeddings/aligned/pnt_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4b5fcf32db0f9fbaffda2a784180d7ab98d67453662b2903086bdb0e2e04199
+size 256277980

models/embeddings/aligned/pnt_32d.meta.json

@@ -0,0 +1 @@
+{"lang": "pnt", "dim": 32, "max_seq_len": 512, "is_aligned": true}

models/embeddings/aligned/pnt_32d.projection.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e43c7e28b167d79007acaca9acd12854c29dac3206cba4c37aaaa2f854bc5a03
+size 4224

models/embeddings/aligned/pnt_32d_metadata.json

@@ -0,0 +1,8 @@
+{
+  "language": "pnt",
+  "dimension": 32,
+  "version": "aligned",
+  "hub_language": "en",
+  "seed_vocab_size": 140,
+  "vocab_size": 1003
+}

models/embeddings/aligned/pnt_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29ce52adc07131d13adcf9304276cbe784c4d7f7415ef585d555e71299965485
+size 512534748

models/embeddings/aligned/pnt_64d.meta.json

@@ -0,0 +1 @@
+{"lang": "pnt", "dim": 64, "max_seq_len": 512, "is_aligned": true}

models/embeddings/aligned/pnt_64d.projection.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70ce77f965d9d7183055959123d1fd7d236bc9fc25f0ba72a12e0564a911c38b
+size 16512

models/embeddings/aligned/pnt_64d_metadata.json

@@ -0,0 +1,8 @@
+{
+  "language": "pnt",
+  "dimension": 64,
+  "version": "aligned",
+  "hub_language": "en",
+  "seed_vocab_size": 140,
+  "vocab_size": 1003
+}

models/embeddings/monolingual/pnt_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5ae2c0f72a238311b78cfe7a7f0c2f653b1c6850bf745081c02f5181e757469
+size 1025048284

models/embeddings/monolingual/pnt_128d.meta.json

@@ -0,0 +1 @@
+{"lang": "pnt", "dim": 128, "max_seq_len": 512, "is_aligned": false}

models/embeddings/monolingual/pnt_128d_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "pnt",
+  "dimension": 128,
+  "version": "monolingual",
+  "training_params": {
+    "algorithm": "skipgram",
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 128,
+    "threads": 5
+  },
+  "vocab_size": 1003
+}

models/embeddings/monolingual/pnt_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4b5fcf32db0f9fbaffda2a784180d7ab98d67453662b2903086bdb0e2e04199
+size 256277980

models/embeddings/monolingual/pnt_32d.meta.json

@@ -0,0 +1 @@
+{"lang": "pnt", "dim": 32, "max_seq_len": 512, "is_aligned": false}

models/embeddings/monolingual/pnt_32d_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "pnt",
+  "dimension": 32,
+  "version": "monolingual",
+  "training_params": {
+    "algorithm": "skipgram",
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 32,
+    "threads": 5
+  },
+  "vocab_size": 1003
+}

models/embeddings/monolingual/pnt_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29ce52adc07131d13adcf9304276cbe784c4d7f7415ef585d555e71299965485
+size 512534748

models/embeddings/monolingual/pnt_64d.meta.json

@@ -0,0 +1 @@
+{"lang": "pnt", "dim": 64, "max_seq_len": 512, "is_aligned": false}

models/embeddings/monolingual/pnt_64d_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "pnt",
+  "dimension": 64,
+  "version": "monolingual",
+  "training_params": {
+    "algorithm": "skipgram",
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 64,
+    "threads": 5
+  },
+  "vocab_size": 1003
+}

models/subword_markov/pnt_markov_ctx1_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:128329c9860d2f68e400f725438f08aa3102cb79ff77675e2e8c62ab83a6008c
+size 32858

models/subword_markov/pnt_markov_ctx1_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "subword",
+  "language": "pnt",
+  "unique_contexts": 439,
+  "total_transitions": 321623
+}

models/subword_markov/pnt_markov_ctx2_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0729f83125339b3437081bbc0e6c59f9ff88d8f7ff58c4f55ac983181208b03d
+size 145007

models/subword_markov/pnt_markov_ctx2_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "subword",
+  "language": "pnt",
+  "unique_contexts": 3856,
+  "total_transitions": 321089
+}

models/subword_markov/pnt_markov_ctx3_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1fd590e1976353b7a222866ea215301d2a966544b3e88b41056e6b9be949af1a
+size 441909

models/subword_markov/pnt_markov_ctx3_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "subword",
+  "language": "pnt",
+  "unique_contexts": 20216,
+  "total_transitions": 320555
+}

models/subword_markov/pnt_markov_ctx4_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1096308dc35224a1c6b9bc4e9e58af9594fc0d56a2feb46ba773f78aceed76d2
+size 891075

models/subword_markov/pnt_markov_ctx4_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "subword",
+  "language": "pnt",
+  "unique_contexts": 53958,
+  "total_transitions": 320021
+}

models/subword_ngram/pnt_2gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:852afb7985fe327a9265819b183ce6af310f1299dbe01a67851b3b7f9f1bfc66
+size 22919

models/subword_ngram/pnt_2gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "subword",
+  "language": "pnt",
+  "unique_ngrams": 1785,
+  "total_ngrams": 321623
+}

models/subword_ngram/pnt_3gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f558e06e0cea80afe06ddac1609c342dfd1b5b25466255e640730fedd2c65c5
+size 116138

models/subword_ngram/pnt_3gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "subword",
+  "language": "pnt",
+  "unique_ngrams": 9341,
+  "total_ngrams": 321089
+}

models/subword_ngram/pnt_4gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f790d0e2b5a62e169473b53878187ff891a8f15aa9f1a8fef650b092a176188b
+size 338158

models/subword_ngram/pnt_4gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "subword",
+  "language": "pnt",
+  "unique_ngrams": 24840,
+  "total_ngrams": 320555
+}

models/subword_ngram/pnt_5gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8c3c8fff90ffd3b4c4b412dd4c20bada73f10e3320f36db6b75cb881cbc42ea
+size 448242

models/subword_ngram/pnt_5gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 5,
+  "variant": "subword",
+  "language": "pnt",
+  "unique_ngrams": 32544,
+  "total_ngrams": 320021
+}

models/tokenizer/pnt_tokenizer_16k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ced00b763df107756349b412102baa6e09df7e4353e4db7d1dd830a7237a6a43
+size 573788

models/tokenizer/pnt_tokenizer_32k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9dcf2f6658ec47d4a472c3c0ab8aef1d750d9b4ac354521a1cb8994100f8f69b
+size 863303

models/tokenizer/pnt_tokenizer_8k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f142ffa0c252280961d968338131eb86cb27a718da13dba88d393b52de3724e5
+size 400741

models/vocabulary/pnt_vocabulary.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9204779bc124b56ba6b8a1398ef95d4765c9f32273595246713efdfa28985580
+size 70072

models/vocabulary/pnt_vocabulary_metadata.json

@@ -0,0 +1,17 @@
+{
+  "language": "pnt",
+  "vocabulary_size": 3936,
+  "variant": "full",
+  "statistics": {
+    "type_token_ratio": 0.22016252855484403,
+    "coverage": {
+      "top_100": 0.4610530651986668,
+      "top_1000": 0.7095831928996742,
+      "top_5000": 0.8734599108714377,
+      "top_10000": 0.9670823502977194
+    },
+    "hapax_count": 7822,
+    "hapax_ratio": 0.6652491920394625,
+    "total_documents": 534
+  }
+}

models/word_markov/pnt_markov_ctx1_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8583857210c1d89236298a3e7122e409e8048c5ae590f84f256ba9a00d24188b
+size 410132

models/word_markov/pnt_markov_ctx1_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "word",
+  "language": "pnt",
+  "unique_contexts": 11667,
+  "total_transitions": 52872
+}

models/word_markov/pnt_markov_ctx2_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a01e5ed9cb478084a16b4cd1fa08c227347010be0aea6337af280f4c53f516bc
+size 785051