Upload all models and assets for bdr (20251001)

.gitattributes

@@ -33,3 +33,8 @@ 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/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 +1,709 @@
+---
+language: bdr
+language_name: BDR
+language_family: austronesian_other
+tags:
+  - wikilangs
+  - nlp
+  - tokenizer
+  - embeddings
+  - n-gram
+  - markov
+  - wikipedia
+  - monolingual
+  - family-austronesian_other
+license: mit
+library_name: wikilangs
+pipeline_tag: feature-extraction
+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: 4.792
+  - name: best_isotropy
+    type: isotropy
+    value: 0.0482
+  - name: vocabulary_size
+    type: vocab
+    value: 0
+generated: 2026-01-03
+---
+
+# BDR - Wikilangs Models
+## Comprehensive Research Report & Full Ablation Study
+
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **BDR** 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)
+- [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** | 4.792x 🏆 | 4.81 | 0.1661% | 33,107 |
+
+### Tokenization Examples
+
+Below are sample sentences tokenized with each vocabulary size:
+
+**Sample 1:** `Nimbug iyono indu' manuk nuut ngentelo ta' keteraan manuk lain.`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁nimbug ▁iyono ▁indu ' ▁manuk ▁nuut ▁ngentelo ▁ta ' ▁keteraan ... (+3 more)` | 13 |
+
+**Sample 2:** `Raja iyo no' dangan jomo kuleh kuasa diom pemerintah dikau kerajaan.Endo rojo pi...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁raja ▁iyo ▁no ' ▁dangan ▁jomo ▁kuleh ▁kuasa ▁diom ▁pemerintah ... (+20 more)` | 30 |
+
+**Sample 3:** `Para-para iyo no tempat ngena segala barang enjata rak`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁para - para ▁iyo ▁no ▁tempat ▁ngena ▁segala ▁barang ▁enjata ... (+1 more)` | 11 |
+
+
+### Key Findings
+
+- **Best Compression:** 8k achieves 4.792x compression
+- **Lowest UNK Rate:** 8k with 0.1661% 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 | 287 | 8.16 | 401 | 53.3% | 100.0% |
+| **2-gram** | Subword | 181 🏆 | 7.50 | 597 | 77.1% | 100.0% |
+| **3-gram** | Word | 221 | 7.79 | 271 | 59.6% | 100.0% |
+| **3-gram** | Subword | 1,140 | 10.15 | 3,421 | 32.8% | 85.1% |
+| **4-gram** | Word | 273 | 8.09 | 346 | 51.1% | 100.0% |
+| **4-gram** | Subword | 4,426 | 12.11 | 11,413 | 17.0% | 52.6% |
+
+### Top 5 N-grams by Size
+
+**2-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `tungan metelak` | 162 |
+| 2 | `iyo no` | 138 |
+| 3 | `iyo noh` | 69 |
+| 4 | `iyo tu` | 68 |
+| 5 | `bioso ni` | 45 |
+
+**3-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `ma na ni` | 40 |
+| 2 | `dewan undangan negeri` | 26 |
+| 3 | `undangan negeri sabah` | 25 |
+| 4 | `iyo tu dangan` | 19 |
+| 5 | `tungan metelak dendo` | 18 |
+
+**4-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `dewan undangan negeri sabah` | 25 |
+| 2 | `tungan metelak dendo malaysia` | 18 |
+| 3 | `sama ma na ni` | 14 |
+| 4 | `iyo no endangan jomo` | 12 |
+| 5 | `no endangan jomo politik` | 12 |
+
+**2-grams (Subword):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `a n` | 5,437 |
+| 2 | `n _` | 3,734 |
+| 3 | `n g` | 3,473 |
+| 4 | `i _` | 3,019 |
+| 5 | `_ t` | 2,998 |
+
+**3-grams (Subword):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `a n _` | 2,443 |
+| 2 | `a n g` | 1,577 |
+| 3 | `n g _` | 1,357 |
+| 4 | `_ t a` | 1,076 |
+| 5 | `_ n i` | 987 |
+
+**4-grams (Subword):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `a n g _` | 910 |
+| 2 | `_ n i _` | 650 |
+| 3 | `_ i y o` | 643 |
+| 4 | `n g a n` | 619 |
+| 5 | `g a n _` | 579 |
+
+
+### Key Findings
+
+- **Best Perplexity:** 2-gram (subword) with 181
+- **Entropy Trend:** Decreases with larger n-grams (more predictable)
+- **Coverage:** Top-1000 patterns cover ~53% 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 | 0.8053 | 1.747 | 3.60 | 5,241 | 19.5% |
+| **1** | Subword | 1.4652 | 2.761 | 11.14 | 104 | 0.0% |
+| **2** | Word | 0.1666 | 1.122 | 1.26 | 18,592 | 83.3% |
+| **2** | Subword | 1.1951 | 2.290 | 5.74 | 1,154 | 0.0% |
+| **3** | Word | 0.0377 | 1.026 | 1.05 | 22,996 | 96.2% |
+| **3** | Subword | 0.7985 | 1.739 | 3.15 | 6,603 | 20.2% |
+| **4** | Word | 0.0104 🏆 | 1.007 | 1.01 | 23,590 | 99.0% |
+| **4** | Subword | 0.5443 | 1.458 | 2.09 | 20,699 | 45.6% |
+
+### Generated Text Samples (Word-based)
+
+Below are text samples generated from each word-based Markov chain model:
+
+**Context Size 1:**
+
+1. `ni ta lok kuah engko tangsi selegubdi tu terhasil moko dangan pelego dendo malaysia beliau tu`
+2. `tu tungan ni un duo ni mediam tepung buas tak sekul tena tana amun pinapi enggo`
+3. `iyo boi nilego oleg ni jomo yang bok ni pan akan buan raya kota belud tu`
+
+**Context Size 2:**
+
+1. `iyo no nyaun preskripsi toos bineli ta farmasi atau mediam kadai yang nyaun sebarang halangan engko ...`
+2. `iyo noh kui tradisional jomo mitu sabah kui tu bentuk ni dokon indung jari engko binuat lua`
+3. `iyo tu boi ni urus le ni gua a masi un sampai betiru terutama ni sembiang pardu`
+
+**Context Size 3:**
+
+1. `ma na ni teko ta tampat tungan setemu tapi jomo tenemuan ai no lumaan`
+2. `dewan undangan negeri sabah ta kewasan tempasuk lua tungan metelak politik malaysia di pertua laat a...`
+3. `undangan negeri sabah betiru`
+
+**Context Size 4:**
+
+1. `dewan undangan negeri sabah dun lua september tu anggota pertubuhan kebangsaan melayu bersatu malays...`
+2. `sama ma na ni ai ngemban matai`
+3. `ahli dewan undangan negeri sabah dewan undangan negeri sabah dewan undangan negeri sabah ta kewasan ...`
+
+
+### Generated Text Samples (Subword-based)
+
+Below are text samples generated from each subword-based Markov chain model:
+
+**Context Size 1:**
+
+1. `_njano-bo_cseria`
+2. `anegal_8_t_bu"_b`
+3. `ngim_nd_bo_isaup`
+
+**Context Size 2:**
+
+1. `an_kain_tamuḥamal`
+2. `n_jom_no_turi_mud`
+3. `ngko_ta_tang_boi_`
+
+**Context Size 3:**
+
+1. `an_ni_ana'_nakasal`
+2. `ang_jomo_untuan_ta`
+3. `ng_teali_pulo_ko'_`
+
+**Context Size 4:**
+
+1. `ang_sefalopod_lua'_`
+2. `_ni_denga_septembag`
+3. `_iyo_no_telia_punya`
+
+
+### Key Findings
+
+- **Best Predictability:** Context-4 (word) with 99.0% predictability
+- **Branching Factor:** Decreases with context size (more deterministic)
+- **Memory Trade-off:** Larger contexts require more storage (20,699 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 | 2,342 |
+| Total Tokens | 23,366 |
+| Mean Frequency | 9.98 |
+| Median Frequency | 3 |
+| Frequency Std Dev | 33.27 |
+
+### Most Common Words
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | ni | 760 |
+| 2 | tu | 584 |
+| 3 | iyo | 549 |
+| 4 | ta | 455 |
+| 5 | yang | 382 |
+| 6 | boi | 354 |
+| 7 | pan | 303 |
+| 8 | kok | 280 |
+| 9 | jomo | 275 |
+| 10 | tungan | 250 |
+
+### Least Common Words (from vocabulary)
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | pelikat | 2 |
+| 2 | avi | 2 |
+| 3 | me | 2 |
+| 4 | jewatan | 2 |
+| 5 | michael | 2 |
+| 6 | joseph | 2 |
+| 7 | ho | 2 |
+| 8 | ny | 2 |
+| 9 | pembunuh | 2 |
+| 10 | mundu | 2 |
+
+### Zipf's Law Analysis
+
+| Metric | Value |
+|--------|-------|
+| Zipf Coefficient | 0.9532 |
+| R² (Goodness of Fit) | 0.984280 |
+| Adherence Quality | **excellent** |
+
+### Coverage Analysis
+
+| Top N Words | Coverage |
+|-------------|----------|
+| Top 100 | 45.5% |
+| Top 1,000 | 85.6% |
+| Top 5,000 | 0.0% |
+| Top 10,000 | 0.0% |
+
+### Key Findings
+
+- **Zipf Compliance:** R²=0.9843 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 45.5% of corpus
+- **Long Tail:** -7,658 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
+
+> *Note: Multilingual alignment visualization not available for this language.*
+
+
+### 5.2 Model Comparison
+
+| Model | Dimension | Isotropy | Semantic Density | Alignment R@1 | Alignment R@10 |
+|-------|-----------|----------|------------------|---------------|----------------|
+| **mono_32d** | 32 | 0.0482 🏆 | 0.8825 | N/A | N/A |
+| **mono_64d** | 64 | 0.0132 | 0.9050 | N/A | N/A |
+| **mono_128d** | 128 | 0.0053 | 0.9273 | N/A | N/A |
+
+### Key Findings
+
+- **Best Isotropy:** mono_32d with 0.0482 (more uniform distribution)
+- **Semantic Density:** Average pairwise similarity of 0.9049. Lower values indicate better semantic separation.
+- **Alignment Quality:** No aligned models evaluated in this run.
+- **Recommendation:** 128d aligned for best cross-lingual performance
+
+---
+## 6.  Morphological Analysis (Experimental)
+
+> ⚠️ **Warning:** This language shows low morphological productivity. The statistical signals used for this analysis may be noisy or less reliable than for morphologically rich languages.
+
+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 | **0.000** | Low morphological productivity | ⚠️ Likely unreliable |
+| Idiomaticity Gap | **-1.000** | Low formulaic 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 |
+|--------|----------|
+| `-pe` | petaling, pekakas, peketa |
+| `-se` | sejak, seniram, sejati |
+| `-ke` | kenangan, kerita, keratas |
+| `-te` | tehe, tempoh, tetiak |
+| `-me` | meruma, menurut, melioro |
+| `-be` | berukuran, berfikir, benua |
+
+#### Productive Suffixes
+| Suffix | Examples |
+|--------|----------|
+| `-n` | kumpulan, regisin, haiwan |
+| `-an` | kumpulan, haiwan, berukuran |
+| `-ng` | petaling, kantung, ngulang |
+| `-ang` | ngulang, manang, sayang |
+| `-ah` | tah, fatimah, umrah |
+
+### 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.
+
+*No significant bound stems detected.*
+
+
+### 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 |
+|--------|--------|-----------|----------|
+| `-pe` | `-n` | 55 words | pelan, pentaran |
+| `-pe` | `-an` | 49 words | pelan, pentaran |
+| `-ke` | `-n` | 42 words | kenangan, keteraan |
+| `-ke` | `-an` | 36 words | kenangan, keteraan |
+| `-se` | `-n` | 11 words | sebahagian, selain |
+| `-te` | `-n` | 9 words | temban, tenomon |
+| `-se` | `-ng` | 9 words | sedong, sepanjang |
+| `-me` | `-n` | 9 words | mesimpon, meluman |
+| `-pe` | `-ng` | 7 words | petaling, pelancong |
+| `-be` | `-n` | 7 words | berukuran, been |
+
+### 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 |
+|------|-----------------|------------|------|
+| kebenyakan | **`ke-be-nyak-an`** | 7.5 | `nyak` |
+| kebangsaan | **`ke-bangsa-an`** | 6.0 | `bangsa` |
+| kelebihan | **`ke-lebih-an`** | 6.0 | `lebih` |
+| kelahiran | **`ke-lahir-an`** | 6.0 | `lahir` |
+| keramaian | **`ke-ramai-an`** | 6.0 | `ramai` |
+| kepulauan | **`ke-pulau-an`** | 6.0 | `pulau` |
+| kebudayaan | **`ke-budaya-an`** | 6.0 | `budaya` |
+| keputeraan | **`ke-putera-an`** | 6.0 | `putera` |
+| sedembila | **`se-dembila`** | 4.5 | `dembila` |
+| perpisahan | **`pe-rpis-ah-an`** | 4.5 | `rpis` |
+| keselamatan | **`ke-se-lamat-an`** | 4.5 | `lamat` |
+| pernikahan | **`pe-rnik-ah-an`** | 4.5 | `rnik` |
+| perjuangan | **`pe-rjua-ng-an`** | 4.5 | `rjua` |
+| kemerdekaan | **`ke-me-rdeka-an`** | 4.5 | `rdeka` |
+| kepelbagaian | **`ke-pe-lbagai-an`** | 4.5 | `lbagai` |
+
+### 6.6 Linguistic Interpretation
+
+> **Automated Insight:**
+The language BDR appears to be more isolating or has a highly fixed vocabulary. Word-level models perform nearly as well as subword models, indicating fewer productive morphological processes.
+
+---
+## 7. Summary & Recommendations
+
+![Performance Dashboard](visualizations/performance_dashboard.png)
+
+### Production Recommendations
+
+| Component | Recommended | Rationale |
+|-----------|-------------|-----------|
+| Tokenizer | **8k BPE** | Best compression (4.79x) |
+| N-gram | **2-gram** | Lowest perplexity (181) |
+| Markov | **Context-4** | Highest predictability (99.0%) |
+| 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: 2026-01-03 06:44:23*

models/embeddings/monolingual/bdr_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14901cc49f6d41fceff30e160e6f74a2d3e4c31128383c0eb0c581df8f55667d
+size 1024837728

models/embeddings/monolingual/bdr_128d.meta.json

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

models/embeddings/monolingual/bdr_128d_metadata.json

@@ -0,0 +1,15 @@
+{
+  "language": "bdr",
+  "dimension": 128,
+  "version": "monolingual",
+  "training_params": {
+    "algorithm": "skipgram",
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 128
+  },
+  "vocab_size": 806
+}

models/embeddings/monolingual/bdr_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa045fafcb5ac078ce6dc442fa1420ea920b282300bd6a2013469a40763dfb5e
+size 256218720

models/embeddings/monolingual/bdr_32d.meta.json

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

models/embeddings/monolingual/bdr_32d_metadata.json

@@ -0,0 +1,15 @@
+{
+  "language": "bdr",
+  "dimension": 32,
+  "version": "monolingual",
+  "training_params": {
+    "algorithm": "skipgram",
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 32
+  },
+  "vocab_size": 806
+}

models/embeddings/monolingual/bdr_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f57494e6d97c810c174b928ec3c054ba096b896edd93f5bb2951c51ea1469fb
+size 512425056

models/embeddings/monolingual/bdr_64d.meta.json

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

models/embeddings/monolingual/bdr_64d_metadata.json

@@ -0,0 +1,15 @@
+{
+  "language": "bdr",
+  "dimension": 64,
+  "version": "monolingual",
+  "training_params": {
+    "algorithm": "skipgram",
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 64
+  },
+  "vocab_size": 806
+}

models/subword_markov/bdr_markov_ctx1_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90dd99baf68b20403d569b8e5f8d37366e4ddf2960b74590e45c40e1931d3cf1
+size 12465

models/subword_markov/bdr_markov_ctx1_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "subword",
+  "language": "bdr",
+  "unique_contexts": 104,
+  "total_transitions": 158102
+}

models/subword_markov/bdr_markov_ctx2_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de0458067f06b19cd92069895f072cb47f6cdac7b1aec47edc3f54439b9dc8fb
+size 51058

models/subword_markov/bdr_markov_ctx2_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "subword",
+  "language": "bdr",
+  "unique_contexts": 1154,
+  "total_transitions": 157564
+}

models/subword_markov/bdr_markov_ctx3_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a603938c76585d0058fba20159b63f48cc8de1a081d60c9e812d5ef4cc856c2
+size 147977

models/subword_markov/bdr_markov_ctx3_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "subword",
+  "language": "bdr",
+  "unique_contexts": 6603,
+  "total_transitions": 157026
+}

models/subword_markov/bdr_markov_ctx4_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40b65aa668a1cff884dd7d1981f09619017109f7d40edc49b7916cb9d523100b
+size 355053

models/subword_markov/bdr_markov_ctx4_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "subword",
+  "language": "bdr",
+  "unique_contexts": 20699,
+  "total_transitions": 156488
+}

models/subword_ngram/bdr_2gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f89d4cd1bab1bda41e8fb6e1bcbdcfb402307503bc759166623f23f6684cb7d8
+size 9761

models/subword_ngram/bdr_2gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "subword",
+  "language": "bdr",
+  "unique_ngrams": 597,
+  "total_ngrams": 158102
+}

models/subword_ngram/bdr_3gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82c353ab6740086eface298a08842c1494816358218a9e1924c16d5cc9298c3d
+size 37547

models/subword_ngram/bdr_3gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "subword",
+  "language": "bdr",
+  "unique_ngrams": 3421,
+  "total_ngrams": 157564
+}

models/subword_ngram/bdr_4gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00c9461926041d02b2401848dda26b82b5b87974371b657a4eeb9a268d495f8c
+size 128743

models/subword_ngram/bdr_4gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "subword",
+  "language": "bdr",
+  "unique_ngrams": 11413,
+  "total_ngrams": 157026
+}

models/tokenizer/bdr_tokenizer_8k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60470ba72369b5d6feb2a6ddd8122b4eb757065597901f8d33f1f47db9d61152
+size 373260

models/vocabulary/bdr_vocabulary.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a58092d4262f6aa43f6194c3af31f93b2c60e2e1f1457610992d6848f4156246
+size 38048

models/vocabulary/bdr_vocabulary_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "bdr",
+  "vocabulary_size": 2342,
+  "variant": "full",
+  "statistics": {
+    "type_token_ratio": 0.20221606648199447,
+    "coverage": {
+      "top_100": 0.4034075816795052,
+      "top_1000": 0.7593442871779305,
+      "top_5000": 0.9875156528668463
+    },
+    "hapax_count": 2987,
+    "hapax_ratio": 0.5605179208106587,
+    "total_documents": 538
+  }
+}

models/word_markov/bdr_markov_ctx1_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50da15bb2d596e55b1af74da418be234ae20d7994f359665a73457c179ace2c4
+size 149999

models/word_markov/bdr_markov_ctx1_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "word",
+  "language": "bdr",
+  "unique_contexts": 5241,
+  "total_transitions": 25815
+}

models/word_markov/bdr_markov_ctx2_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a70c0745d64f93f951581ccf1e770bc0198416279a90b1c76af931bc397bb65f
+size 308032

models/word_markov/bdr_markov_ctx2_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "word",
+  "language": "bdr",
+  "unique_contexts": 18592,
+  "total_transitions": 25277
+}

models/word_markov/bdr_markov_ctx3_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:223447078bb309ba5d758145f41c4571806689ffaf1efe8aef015c68c49c7526
+size 388681

models/word_markov/bdr_markov_ctx3_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "word",
+  "language": "bdr",
+  "unique_contexts": 22996,
+  "total_transitions": 24739
+}

models/word_markov/bdr_markov_ctx4_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:66c6e307074a65e79883d656ec156325dab1be18c7c2082eab0c9665ad6e3133
+size 429090

models/word_markov/bdr_markov_ctx4_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "word",
+  "language": "bdr",
+  "unique_contexts": 23590,
+  "total_transitions": 24201
+}

models/word_ngram/bdr_2gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f7dbbf4075f6494f4e948bc0417ec1b327b33cb3676c74c44c5cf7b15c76e93
+size 7711

models/word_ngram/bdr_2gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "word",
+  "language": "bdr",
+  "unique_ngrams": 401,
+  "total_ngrams": 25815
+}

models/word_ngram/bdr_3gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03582a17ce01397bddd0d583a4eb8301b1d9ad4a0ebaa6b95bafe86a66321950
+size 6657

models/word_ngram/bdr_3gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "word",
+  "language": "bdr",
+  "unique_ngrams": 271,
+  "total_ngrams": 25277
+}

models/word_ngram/bdr_4gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b39ba29b04354ddf98ba4fefae173d2708834a6a0814a4f9cd3970bc844751a
+size 8746

models/word_ngram/bdr_4gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "word",
+  "language": "bdr",
+  "unique_ngrams": 346,
+  "total_ngrams": 24739
+}