Upload all models and assets for dv (20251201)

.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/performance_dashboard.png filter=lfs diff=lfs merge=lfs -text
+visualizations/position_encoding_comparison.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 +1,556 @@
+---
+language: dv
+language_name: DV
+language_family: indoaryan_insular
+tags:
+  - wikilangs
+  - nlp
+  - tokenizer
+  - embeddings
+  - n-gram
+  - markov
+  - wikipedia
+  - monolingual
+  - family-indoaryan_insular
+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: 6.114
+  - name: best_isotropy
+    type: isotropy
+    value: 0.8870
+  - name: vocabulary_size
+    type: vocab
+    value: 6434
+generated: 2025-12-30
+---
+
+# DV - Wikilangs Models
+## Comprehensive Research Report & Full Ablation Study
+
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **DV** 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-gram)
+- Markov chains (context of 1, 2, 3 and 4)
+- Subword N-gram and Markov chains
+- Embeddings in various sizes and dimensions
+- 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. Summary & Recommendations](#6-summary--recommendations)
+- [Metrics Glossary](#appendix-metrics-glossary--interpretation-guide)
+- [Visualizations Index](#visualizations-index)
+
+---
+## 1. Tokenizer Evaluation
+
+![Tokenizer Compression](visualizations/tokenizer_compression.png)
+
+### Results
+
+| Vocab Size | Compression | Avg Token Len | UNK Rate | Total Tokens |
+|------------|-------------|---------------|----------|--------------|
+| **8k** | 4.416x | 4.38 | 0.0656% | 543,835 |
+| **16k** | 5.055x | 5.01 | 0.0751% | 475,116 |
+| **32k** | 5.621x | 5.58 | 0.0836% | 427,269 |
+| **64k** | 6.114x 🏆 | 6.06 | 0.0909% | 392,821 |
+
+### Tokenization Examples
+
+Below are sample sentences tokenized with each vocabulary size:
+
+**Sample 1:** `ވާރުތަ ބަލި އަކީ ލޭގެ ގުޅުމުގެ ސަބަބުން ދަރިފަސްކޮޅަށް ދެމިގެންދާ ބަލިތަކެވެ.
+
+ޤ...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁ވާރު ތަ ▁ބަލި ▁އަކީ ▁ލޭގެ ▁ގުޅ ުމުގެ ▁ސަބަބުން ▁ދަރިފ ަސް ... (+9 more)` | 19 |
+| 16k | `▁ވާރުތަ ▁ބަލި ▁އަކީ ▁ލޭގެ ▁ގުޅުމުގެ ▁ސަބަބުން ▁ދަރިފަސް ކޮޅަށް ▁ދެމިގެން ދާ ... (+6 more)` | 16 |
+| 32k | `▁ވާރުތަ ▁ބަލި ▁އަކީ ▁ލޭގެ ▁ގުޅުމުގެ ▁ސަބަބުން ▁ދަރިފަސް ކޮޅަށް ▁ދެމިގެންދާ ▁ބަލިތަކެވެ ... (+4 more)` | 14 |
+| 64k | `▁ވާރުތަ ▁ބަލި ▁އަކީ ▁ލޭގެ ▁ގުޅުމުގެ ▁ސަބަބުން ▁ދަރިފަސް ކޮޅަށް ▁ދެމިގެންދާ ▁ބަލިތަކެވެ ... (+4 more)` | 14 |
+
+**Sample 2:** `މިއީ 20ވަނަ ޤަރުނުގެ 99ވަނަ އަހަރެވެ.`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁މިއީ ▁ 2 0 ވަނަ ▁ޤަރުނުގެ ▁ 9 9 ވަނަ ... (+2 more)` | 12 |
+| 16k | `▁މިއީ ▁ 2 0 ވަނަ ▁ޤަރުނުގެ ▁ 9 9 ވަނަ ... (+2 more)` | 12 |
+| 32k | `▁މިއީ ▁ 2 0 ވަނަ ▁ޤަރުނުގެ ▁ 9 9 ވަނަ ... (+2 more)` | 12 |
+| 64k | `▁މިއީ ▁ 2 0 ވަނަ ▁ޤަރުނުގެ ▁ 9 9 ވަނަ ... (+2 more)` | 12 |
+
+**Sample 3:** `މި މަޒުމޫނަކީ ނައިޖީރިއާގެ ނާއިބު ރައީސް އާ ބެހޭ މަޒުމޫނެކެވެ.`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁މި ▁މަޒުމޫނ ަކީ ▁ނައިޖީ ރިއާގެ ▁ނާއިބު ▁ރައީސް ▁އާ ▁ބެހޭ ▁މަޒުމޫނ ... (+2 more)` | 12 |
+| 16k | `▁މި ▁މަޒުމޫނ ަކީ ▁ނައިޖީ ރިއާގެ ▁ނާއިބު ▁ރައީސް ▁އާ ▁ބެހޭ ▁މަޒުމޫނ ... (+2 more)` | 12 |
+| 32k | `▁މި ▁މަޒުމޫނަކީ ▁ނައިޖީރިއާގެ ▁ނާއިބު ▁ރައީސް ▁އާ ▁ބެހޭ ▁މަޒުމޫނެކެވެ .` | 9 |
+| 64k | `▁މި ▁މަޒުމޫނަކީ ▁ނައިޖީރިއާގެ ▁ނާއިބު ▁ރައީސް ▁އާ ▁ބެހޭ ▁މަޒުމޫނެކެވެ .` | 9 |
+
+
+### Key Findings
+
+- **Best Compression:** 64k achieves 6.114x compression
+- **Lowest UNK Rate:** 8k with 0.0656% 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 Coverage](visualizations/ngram_coverage.png)
+
+### Results
+
+| N-gram | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
+|--------|------------|---------|----------------|------------------|-------------------|
+| **2-gram** | 280 🏆 | 8.13 | 6,213 | 66.8% | 98.6% |
+| **2-gram** | 309 🏆 | 8.27 | 4,183 | 65.1% | 98.3% |
+| **3-gram** | 1,977 | 10.95 | 27,387 | 28.4% | 74.6% |
+| **3-gram** | 2,024 | 10.98 | 28,100 | 30.0% | 74.0% |
+| **4-gram** | 9,545 | 13.22 | 108,940 | 14.2% | 44.6% |
+| **4-gram** | 8,722 | 13.09 | 111,102 | 15.8% | 46.9% |
+
+### Top 5 N-grams by Size
+
+**2-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `ނ ް` | 201,447 |
+| 2 | `ަ އ` | 180,556 |
+| 3 | `އ ި` | 156,059 |
+| 4 | `ވ ެ` | 109,701 |
+| 5 | `އ ް` | 108,512 |
+
+**3-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `ަ އ ި` | 98,210 |
+| 2 | `ެ ވ ެ` | 69,037 |
+| 3 | `ު ނ ް` | 66,257 |
+| 4 | `ވ ެ .` | 64,917 |
+| 5 | `ަ ށ ް` | 49,257 |
+
+**4-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `ެ ވ ެ .` | 64,897 |
+| 2 | `ގ ަ އ ި` | 40,722 |
+| 3 | `އ ެ ވ ެ` | 37,057 |
+| 4 | `ު ގ ަ އ` | 23,243 |
+| 5 | `އ ި ނ ް` | 18,356 |
+
+
+### Key Findings
+
+- **Best Perplexity:** 2-gram with 280
+- **Entropy Trend:** Decreases with larger n-grams (more predictable)
+- **Coverage:** Top-1000 patterns cover ~47% of corpus
+- **Recommendation:** 4-gram or 5-gram for best predictive performance
+
+---
+## 3. Markov Chain Evaluation
+
+![Markov Entropy](visualizations/markov_entropy.png)
+
+![Markov Branching](visualizations/markov_branching.png)
+
+### Results
+
+| Context | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
+|---------|-------------|------------|------------------|-----------------|----------------|
+| **1** | 0.5858 | 1.501 | 3.83 | 15,381 | 41.4% |
+| **1** | 1.1454 | 2.212 | 8.73 | 1,139 | 0.0% |
+| **2** | 0.2988 🏆 | 1.230 | 2.08 | 58,771 | 70.1% |
+| **2** | 0.9952 🏆 | 1.993 | 5.52 | 9,934 | 0.5% |
+| **3** | 0.3330 | 1.260 | 2.10 | 122,099 | 66.7% |
+| **3** | 0.7840 | 1.722 | 3.51 | 54,858 | 21.6% |
+| **4** | 0.3715 | 1.294 | 2.01 | 255,795 | 62.9% |
+| **4** | 0.5399 | 1.454 | 2.34 | 192,399 | 46.0% |
+
+### Generated Text Samples
+
+Below are text samples generated from each Markov chain model:
+
+**Context Size 1:**
+
+1. `ަ ގ ެ ކ ޫ ގ ެ . 1980މ . އ ް ނ ާ ނ ީ`
+2. `ް ނ ް ފ ަ ބ ަ ކ ު ރ ި ތ ަ ރ ަ ސ`
+3. `ެ . މ ަ ބ ޮ ހ ު ތ ަ ލ ޭ އ ި ހ ެ`
+
+**Context Size 2:**
+
+1. `ނ ް ނ ު ކ ޮ ޅ ު ތ ު ރ ަ އ ި ބ ަ ބ`
+2. `ަ އ ް ކ ަ ނ ް ފ ަ ހ ަ އ ި ވ ެ . މ`
+3. `އ ި ގ ެ ލ ަ ބ ޫ ވ ާ ކ ަ ށ ް އ ި ހ`
+
+**Context Size 3:**
+
+1. `ަ އ ި ވ ާ ހ ަ ށ ި ގ ަ ނޑ ު ގ ަ އ ި ނ`
+2. `ެ ވ ެ . ފ ި ލ ް ރ ަ ސ ް ޖ ެ ހ ޭ ވ ަ`
+3. `ު ނ ް ކ ު ރ ެ އ ް ވ ާ ފ ަ އ ި ވ ާ ޔ`
+
+**Context Size 4:**
+
+1. `ެ ވ ެ . ހ ަ މ ަ އ ި ގ ަ ނ ް ނ ަ އ ި ޖ`
+2. `ގ ަ އ ި މ ި ވ ަ ނ ް ތ ަ ކ ެ ތ ީ ގ ަ އ`
+3. `އ ެ ވ ެ . ފ ަ ތ ަ ށ ް ޢ ަ މ ަ ލ ާ ގ ެ`
+
+
+### Key Findings
+
+- **Best Predictability:** Context-2 with 70.1% predictability
+- **Branching Factor:** Decreases with context size (more deterministic)
+- **Memory Trade-off:** Larger contexts require more storage (192,399 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 | 6,434 |
+| Total Tokens | 3,460,301 |
+| Mean Frequency | 537.81 |
+| Median Frequency | 3 |
+| Frequency Std Dev | 11173.10 |
+
+### Most Common Words
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | އ | 509,388 |
+| 2 | ނ | 367,732 |
+| 3 | މ | 247,599 |
+| 4 | ރ | 246,229 |
+| 5 | ވ | 242,170 |
+| 6 | ކ | 226,840 |
+| 7 | ގ | 210,999 |
+| 8 | ތ | 158,312 |
+| 9 | ދ | 142,011 |
+| 10 | ހ | 138,134 |
+
+### Least Common Words (from vocabulary)
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | علاء | 2 |
+| 2 | حاشية | 2 |
+| 3 | عابدين | 2 |
+| 4 | ٱق | 2 |
+| 5 | حصن | 2 |
+| 6 | 1972ވ | 2 |
+| 7 | abdul_raheem_abdulla_portrait | 2 |
+| 8 | 112x112pxޢ | 2 |
+| 9 | 282ވ | 2 |
+| 10 | costus | 2 |
+
+### Zipf's Law Analysis
+
+| Metric | Value |
+|--------|-------|
+| Zipf Coefficient | 1.2153 |
+| R² (Goodness of Fit) | 0.951559 |
+| Adherence Quality | **excellent** |
+
+### Coverage Analysis
+
+| Top N Words | Coverage |
+|-------------|----------|
+| Top 100 | 98.4% |
+| Top 1,000 | 99.4% |
+| Top 5,000 | 99.9% |
+| Top 10,000 | 0.0% |
+
+### Key Findings
+
+- **Zipf Compliance:** R²=0.9516 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 98.4% of corpus
+- **Long Tail:** -3,566 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)
+
+### Model Comparison
+
+| Model | Vocab Size | Dimension | Avg Norm | Std Norm | Isotropy |
+|-------|------------|-----------|----------|----------|----------|
+| **mono_32d** | 20,888 | 32 | 3.934 | 0.875 | 0.8870 🏆 |
+| **mono_64d** | 20,888 | 64 | 4.609 | 0.777 | 0.8593 |
+| **mono_128d** | 20,888 | 128 | 5.126 | 0.699 | 0.7135 |
+| **embeddings_enhanced** | 0 | 0 | 0.000 | 0.000 | 0.0000 |
+
+### Key Findings
+
+- **Best Isotropy:** mono_32d with 0.8870 (more uniform distribution)
+- **Dimension Trade-off:** Higher dimensions capture more semantics but reduce isotropy
+- **Vocabulary Coverage:** All models cover 20,888 words
+- **Recommendation:** 100d for balanced semantic capture and efficiency
+
+---
+## 6. Summary & Recommendations
+
+![Performance Dashboard](visualizations/performance_dashboard.png)
+
+### Production Recommendations
+
+| Component | Recommended | Rationale |
+|-----------|-------------|-----------|
+| Tokenizer | **32k BPE** | Best compression (6.11x) with low UNK rate |
+| N-gram | **5-gram** | Lowest perplexity (280) |
+| Markov | **Context-4** | Highest predictability (70.1%) |
+| 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},
+  publisher = {HuggingFace},
+  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)
+---
+*Generated by Wikilangs Models Pipeline*
+
+*Report Date: 2025-12-30 08:42:33*

models/embeddings/monolingual/dv_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17f06f2e5b33a30261c4b7b077e79cbd4c974eed9dd568506a068811d5b85523
+size 1045966622

models/embeddings/monolingual/dv_128d.meta.json

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

models/embeddings/monolingual/dv_128d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "dv",
+  "dimension": 128,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 128,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 20888
+}

models/embeddings/monolingual/dv_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:989d442fa29ab19c3397905afc409c5ea3fb2ac3f7ced4ecef5c5b59e58881c8
+size 261924638

models/embeddings/monolingual/dv_32d.meta.json

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

models/embeddings/monolingual/dv_32d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "dv",
+  "dimension": 32,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 32,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 20888
+}

models/embeddings/monolingual/dv_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4c8f3716afafb553389e17fdcee49ce632da31ea6cd2b808bd64ec68e7bdece
+size 523271966

models/embeddings/monolingual/dv_64d.meta.json

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

models/embeddings/monolingual/dv_64d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "dv",
+  "dimension": 64,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 64,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 20888
+}

models/subword_markov/dv_markov_ctx1_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3b530227d1a204f72f8b65de1a305ae378f77c4e8b2924dfc23dbddc8dd38662
+size 81302

models/subword_markov/dv_markov_ctx1_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "subword",
+  "language": "dv",
+  "unique_contexts": 1139,
+  "total_transitions": 8248184
+}

models/subword_markov/dv_markov_ctx2_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3eb962d2bf0387e41cfec8127109760ff53e27cb9bee987ddb74ecf1bca3092
+size 447616

models/subword_markov/dv_markov_ctx2_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "subword",
+  "language": "dv",
+  "unique_contexts": 9934,
+  "total_transitions": 8243619
+}

models/subword_markov/dv_markov_ctx3_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5988c0a1488f6ee94f379fe5d3c8e0f141b7cdd8004be19dd66020914eed032a
+size 1540276

models/subword_markov/dv_markov_ctx3_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "subword",
+  "language": "dv",
+  "unique_contexts": 54858,
+  "total_transitions": 8239054
+}

models/subword_markov/dv_markov_ctx4_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:385dd9376ce9fb1f3e0ab8e1dfb784f0bd42cb629f08765bdeacbdd9c859e5d7
+size 4185420

models/subword_markov/dv_markov_ctx4_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "subword",
+  "language": "dv",
+  "unique_contexts": 192399,
+  "total_transitions": 8234489
+}

models/subword_ngram/dv_2gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc947db07c0da034539c406fe8a1f79e854caae8ba0e21db7a629f743943f4f4
+size 56113

models/subword_ngram/dv_2gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "subword",
+  "language": "dv",
+  "unique_ngrams": 4183,
+  "total_ngrams": 8248184
+}

models/subword_ngram/dv_3gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2de063fc8edb0676cdad6e6286041bbe5f0a9cc4a2d60744c57758a2f1b512be
+size 373741

models/subword_ngram/dv_3gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "subword",
+  "language": "dv",
+  "unique_ngrams": 28100,
+  "total_ngrams": 8243619
+}

models/subword_ngram/dv_4gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a03c4da24aa6e3cbafac8d2e54359f6953d2a058be2d361e5f00d8b80e16540
+size 1401587

models/subword_ngram/dv_4gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "subword",
+  "language": "dv",
+  "unique_ngrams": 111102,
+  "total_ngrams": 8239054
+}

models/tokenizer/dv_tokenizer_16k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b009d0f5075910344a5ec70654b5bb2f1453e75fdf1995842ff06f0a18b68e9
+size 613437

models/tokenizer/dv_tokenizer_32k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee60e395f3e281dda1c84ee287ea26df87d132e2cba1b17e96eb62b98922ba28
+size 1013278

models/tokenizer/dv_tokenizer_64k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da10dcfbcb21fe494de6a305e8ddfbe3a3b893d20e2f3ce5944c6d8f420440a5
+size 1826708

models/tokenizer/dv_tokenizer_8k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7360478c29b2bdf38819aa24a0e676d54c3133a34a63adb669179db4dcfaa256
+size 419002

models/vocabulary/dv_vocabulary.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78a5a307e9810e9b095653562c638e91d2a71e0b2809564df0072f0c05d3e27b
+size 100935

models/vocabulary/dv_vocabulary_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "dv",
+  "vocabulary_size": 6434,
+  "statistics": {
+    "type_token_ratio": 0.004362639272805053,
+    "coverage": {
+      "top_100": 0.9816569669481213,
+      "top_1000": 0.9919847241323944,
+      "top_5000": 0.9966653223795554,
+      "top_10000": 0.9985200350187273
+    },
+    "hapax_count": 8700,
+    "hapax_ratio": 0.5748645434121845,
+    "total_documents": 4565
+  }
+}

models/word_markov/dv_markov_ctx1_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88a9fe6cb63c2277dd80fc088352c0625dfd2fa4ff7477a1b0f56bdee4ab18ce
+size 505042

models/word_markov/dv_markov_ctx1_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "word",
+  "language": "dv",
+  "unique_contexts": 15381,
+  "total_transitions": 7006421
+}

models/word_markov/dv_markov_ctx2_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:655358b5e277faeeb357b22e4f035c2e76a94c1762518ad441fd4c2fdae93abb
+size 1283233

models/word_markov/dv_markov_ctx2_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "word",
+  "language": "dv",
+  "unique_contexts": 58771,
+  "total_transitions": 7001856
+}

models/word_markov/dv_markov_ctx3_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cac192494e3e067382873a2d619a1be722628e35c0213e6989f699f4ed8c281
+size 2642738

models/word_markov/dv_markov_ctx3_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "word",
+  "language": "dv",
+  "unique_contexts": 122099,
+  "total_transitions": 6997291
+}

models/word_markov/dv_markov_ctx4_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ca07713b8573c3d1c74d4ceb4e1fe6ade380e2f8189e63ad82216f40043c423
+size 5437285

models/word_markov/dv_markov_ctx4_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "word",
+  "language": "dv",
+  "unique_contexts": 255795,
+  "total_transitions": 6992726
+}

models/word_ngram/dv_2gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:343a156c33aeeac660b85c2f308be4390973e3c45ebcca252ff8b4b2ffcb6df1
+size 83954

models/word_ngram/dv_2gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "word",
+  "language": "dv",
+  "unique_ngrams": 6213,
+  "total_ngrams": 7006421
+}

models/word_ngram/dv_3gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c6e0926b5288ec9069f3a76ef391fab261dcec0fe9ae0214eb7f2e18ef64b6e
+size 393880

models/word_ngram/dv_3gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "word",
+  "language": "dv",
+  "unique_ngrams": 27387,
+  "total_ngrams": 7001856
+}

models/word_ngram/dv_4gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ded5d0b5ec14a0c1d229616a7bf1f2fc323bbaa97462b502794e5081e90d6d0e
+size 1427660

models/word_ngram/dv_4gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "word",
+  "language": "dv",
+  "unique_ngrams": 108940,
+  "total_ngrams": 6997291
+}