Upload all models and assets for ang (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: ang
+language_name: ANG
+language_family: germanic_historical
+tags:
+  - wikilangs
+  - nlp
+  - tokenizer
+  - embeddings
+  - n-gram
+  - markov
+  - wikipedia
+  - monolingual
+  - family-germanic_historical
+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: 4.001
+  - name: best_isotropy
+    type: isotropy
+    value: 0.7980
+  - name: vocabulary_size
+    type: vocab
+    value: 32745
+generated: 2025-12-27
+---
+
+# ANG - Wikilangs Models
+## Comprehensive Research Report & Full Ablation Study
+
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **ANG** 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** | 3.091x | 3.04 | 0.0788% | 272,719 |
+| **16k** | 3.414x | 3.36 | 0.0871% | 246,960 |
+| **32k** | 3.716x | 3.66 | 0.0948% | 226,874 |
+| **64k** | 4.001x 🏆 | 3.94 | 0.1020% | 210,690 |
+
+### Tokenization Examples
+
+Below are sample sentences tokenized with each vocabulary size:
+
+**Sample 1:** `Yun Chi-ho (Corēanisc: 윤치호, 26 Gēolmōnaþ, 1864 – 9 Gēolmōnaþ, 1945) ƿæs Corēanis...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁y un ▁ch i - ho ▁( cor ēan isc ... (+37 more)` | 47 |
+| 16k | `▁y un ▁chi - ho ▁( cor ēan isc : ... (+35 more)` | 45 |
+| 32k | `▁y un ▁chi - ho ▁( corēan isc : ▁ ... (+33 more)` | 43 |
+| 64k | `▁yun ▁chi - ho ▁( corēanisc : ▁ 윤치호 , ... (+31 more)` | 41 |
+
+**Sample 2:** `Þega  ƿæs se cyning þāra Ēastgotena fram þǣm 552. gēare oþ þæt læt 552. gēare oþ...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁þe ga ▁ƿæs ▁se ▁cyning ▁þāra ▁ēast gotena ▁fram ▁þǣm ... (+32 more)` | 42 |
+| 16k | `▁þe ga ▁ƿæs ▁se ▁cyning ▁þāra ▁ēastgotena ▁fram ▁þǣm ▁ ... (+30 more)` | 40 |
+| 32k | `▁þe ga ▁ƿæs ▁se ▁cyning ▁þāra ▁ēastgotena ▁fram ▁þǣm ▁ ... (+30 more)` | 40 |
+| 64k | `▁þe ga ▁ƿæs ▁se ▁cyning ▁þāra ▁ēastgotena ▁fram ▁þǣm ▁ ... (+30 more)` | 40 |
+
+**Sample 3:** `Ælbūrcerrce () oþþe Ælbūrccerrcke is sēo mǣsteburg on Nīƿemexico.
+
+Flocc:Byrig o...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁æl bū r cer r ce ▁() ▁oþþe ▁æl bū ... (+27 more)` | 37 |
+| 16k | `▁æl būr cer r ce ▁() ▁oþþe ▁æl būr ccer ... (+24 more)` | 34 |
+| 32k | `▁æl būr cer r ce ▁() ▁oþþe ▁æl būr ccer ... (+23 more)` | 33 |
+| 64k | `▁ælbūr cer r ce ▁() ▁oþþe ▁ælbūr ccer rc ke ... (+19 more)` | 29 |
+
+
+### Key Findings
+
+- **Best Compression:** 64k achieves 4.001x compression
+- **Lowest UNK Rate:** 8k with 0.0788% 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** | 4,036 🏆 | 11.98 | 12,127 | 24.9% | 51.6% |
+| **2-gram** | 432 🏆 | 8.76 | 3,589 | 56.9% | 97.1% |
+| **3-gram** | 5,292 | 12.37 | 13,432 | 21.9% | 44.6% |
+| **3-gram** | 3,950 | 11.95 | 28,365 | 20.9% | 59.7% |
+| **4-gram** | 10,291 | 13.33 | 22,900 | 17.1% | 35.0% |
+| **4-gram** | 21,387 | 14.38 | 125,189 | 10.5% | 31.5% |
+
+### Top 5 N-grams by Size
+
+**2-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `flocc :` | 7,126 |
+| 2 | `, and` | 3,048 |
+| 3 | `. flocc` | 2,814 |
+| 4 | `) is` | 1,674 |
+| 5 | `: byrig` | 1,552 |
+
+**3-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `. flocc :` | 2,813 |
+| 2 | `flocc : byrig` | 1,551 |
+| 3 | `: byrig on` | 1,319 |
+| 4 | `( ) is` | 943 |
+| 5 | `< td valign` | 629 |
+
+**4-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `flocc : byrig on` | 1,319 |
+| 2 | `. flocc : byrig` | 999 |
+| 3 | `< td valign =` | 629 |
+| 4 | `td valign = top` | 627 |
+| 5 | `valign = top >` | 615 |
+
+
+### Key Findings
+
+- **Best Perplexity:** 2-gram with 432
+- **Entropy Trend:** Decreases with larger n-grams (more predictable)
+- **Coverage:** Top-1000 patterns cover ~32% 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** | 0.5832 | 1.498 | 3.71 | 92,131 | 41.7% |
+| **1** | 1.0106 | 2.015 | 7.57 | 1,214 | 0.0% |
+| **2** | 0.1973 | 1.147 | 1.44 | 340,526 | 80.3% |
+| **2** | 1.0352 | 2.049 | 6.24 | 9,184 | 0.0% |
+| **3** | 0.0600 | 1.042 | 1.10 | 488,936 | 94.0% |
+| **3** | 0.8839 | 1.845 | 4.02 | 57,270 | 11.6% |
+| **4** | 0.0221 🏆 | 1.015 | 1.03 | 535,590 | 97.8% |
+| **4** | 0.6030 🏆 | 1.519 | 2.46 | 230,176 | 39.7% |
+
+### Generated Text Samples
+
+Below are text samples generated from each Markov chain model:
+
+**Context Size 1:**
+
+1. `. magnus hēold brytenƿealdan sƿā þæt land æt 77 . 00 solidarity tax rates vary in`
+2. `, 568 , canadian golfer 1960 / haː / 1743 flocc : erica durance sƿā sumum`
+3. `and scotlandes heallum , se mǣsta luh onmiddan þām þēgnum þæs australisca ƿerungþrēates , cent and`
+
+**Context Size 2:**
+
+1. `flocc : sċīrbyrig on colorado flocc : crabbas`
+2. `, and is hēofodburg paranā þæs rīces and valetta þæs cynelican hired þæs spræce gesprōcen on west`
+3. `. flocc : byrig on þeodsclande . flocc : byrig on eoferwicscīre flocc : ceastra þæs geānedan`
+
+**Context Size 3:**
+
+1. `. flocc : geboren in 1989 flocc : ƿīf flocc : angelseaxisc englaland flocc : stǣr flocc :`
+2. `flocc : byrig on eoferwicscīre flocc : ceastra þæs geānedan cynerīces ‎ flocc : sūþcorēa`
+3. `: byrig on ġeolurēadsċīr ( californie ) flocc : sċīrbyrig on cǣnsasum flocc : byrig and þorpas on`
+
+**Context Size 4:**
+
+1. `flocc : byrig on orlēanascīre flocc : sċīrbyrig on nīwe eoforwīc flocc : sċīrbyrig on miscegan`
+2. `. flocc : byrig and þorpas on sorie`
+3. `< td valign = top > < small > 24 sēremōnaþ 79 oþ 13 hærfestmōnaþ 81 < td valign`
+
+
+### Key Findings
+
+- **Best Predictability:** Context-4 with 97.8% predictability
+- **Branching Factor:** Decreases with context size (more deterministic)
+- **Memory Trade-off:** Larger contexts require more storage (230,176 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 | 32,745 |
+| Total Tokens | 440,987 |
+| Mean Frequency | 13.47 |
+| Median Frequency | 3 |
+| Frequency Std Dev | 159.89 |
+
+### Most Common Words
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | and | 14,426 |
+| 2 | on | 10,769 |
+| 3 | is | 10,434 |
+| 4 | in | 10,245 |
+| 5 | flocc | 7,161 |
+| 6 | of | 6,202 |
+| 7 | se | 4,329 |
+| 8 | the | 4,026 |
+| 9 | þǣm | 3,673 |
+| 10 | þæs | 3,617 |
+
+### Least Common Words (from vocabulary)
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | ƿīleacstede | 2 |
+| 2 | cōcsċīre | 2 |
+| 3 | winnebagsċīre | 2 |
+| 4 | ælfrēdingtūn | 2 |
+| 5 | irfung | 2 |
+| 6 | dællassċīr | 2 |
+| 7 | lubbecsċīr | 2 |
+| 8 | larēodo | 2 |
+| 9 | grœndā | 2 |
+| 10 | dǣlungs | 2 |
+
+### Zipf's Law Analysis
+
+| Metric | Value |
+|--------|-------|
+| Zipf Coefficient | 0.9423 |
+| R² (Goodness of Fit) | 0.997151 |
+| Adherence Quality | **excellent** |
+
+### Coverage Analysis
+
+| Top N Words | Coverage |
+|-------------|----------|
+| Top 100 | 37.1% |
+| Top 1,000 | 58.8% |
+| Top 5,000 | 77.7% |
+| Top 10,000 | 86.0% |
+
+### Key Findings
+
+- **Zipf Compliance:** R²=0.9972 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 37.1% of corpus
+- **Long Tail:** 22,745 words needed for remaining 14.0% 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** | 10,885 | 32 | 3.878 | 0.872 | 0.7980 🏆 |
+| **mono_64d** | 10,885 | 64 | 4.075 | 0.834 | 0.4885 |
+| **mono_128d** | 10,885 | 128 | 4.131 | 0.840 | 0.1418 |
+| **embeddings_enhanced** | 0 | 0 | 0.000 | 0.000 | 0.0000 |
+
+### Key Findings
+
+- **Best Isotropy:** mono_32d with 0.7980 (more uniform distribution)
+- **Dimension Trade-off:** Higher dimensions capture more semantics but reduce isotropy
+- **Vocabulary Coverage:** All models cover 10,885 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 (4.00x) with low UNK rate |
+| N-gram | **5-gram** | Lowest perplexity (432) |
+| 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},
+  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: 2025-12-27 06:04:51*

models/embeddings/monolingual/ang_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc194d4a31942e2e32c3a8dae780c49b26b5b2dcf2d1a0743c00d42c18909362
+size 1035329826

models/embeddings/monolingual/ang_128d.meta.json

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

models/embeddings/monolingual/ang_128d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "ang",
+  "dimension": 128,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 128,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 10885
+}

models/embeddings/monolingual/ang_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6df9faaec5697096804c52251169b69509f4b58ecec42b9c97f06f9b99d3a2d6
+size 258970146

models/embeddings/monolingual/ang_32d.meta.json

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

models/embeddings/monolingual/ang_32d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "ang",
+  "dimension": 32,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 32,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 10885
+}

models/embeddings/monolingual/ang_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07df11aa5c3e43134f3a3f21f7f1109a62605b73ffd0840c3a062fefcaac8477
+size 517756706

models/embeddings/monolingual/ang_64d.meta.json

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

models/embeddings/monolingual/ang_64d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "ang",
+  "dimension": 64,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 64,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 10885
+}

models/subword_markov/ang_markov_ctx1_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96efa4fd32f8723ecce729acbd1c362fe3222b904974ce6779fa687e76c388d0
+size 73394

models/subword_markov/ang_markov_ctx1_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "subword",
+  "language": "ang",
+  "unique_contexts": 1214,
+  "total_transitions": 3248126
+}

models/subword_markov/ang_markov_ctx2_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:baf874239e5f656387a4b53b454dad05437a34b8038d1d1b3d564c5cba24be6a
+size 453198

models/subword_markov/ang_markov_ctx2_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "subword",
+  "language": "ang",
+  "unique_contexts": 9184,
+  "total_transitions": 3242991
+}

models/subword_markov/ang_markov_ctx3_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43326969d8e8254044d5345d200b3dc66db5c4bc7c296648dd4075f37a900d29
+size 1646395

models/subword_markov/ang_markov_ctx3_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "subword",
+  "language": "ang",
+  "unique_contexts": 57270,
+  "total_transitions": 3237856
+}

models/subword_markov/ang_markov_ctx4_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2300760c01bd695ee195c7c4bbe9527831968d70b9e8a499bbdf3dbb2acc8094
+size 4538468

models/subword_markov/ang_markov_ctx4_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "subword",
+  "language": "ang",
+  "unique_contexts": 230176,
+  "total_transitions": 3232721
+}

models/subword_ngram/ang_2gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d0aef14dece44db6119149b33b24ba911374e76abaa8d1c9930045cc8cd84a9
+size 47002

models/subword_ngram/ang_2gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "subword",
+  "language": "ang",
+  "unique_ngrams": 3589,
+  "total_ngrams": 3248126
+}

models/subword_ngram/ang_3gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4356d4e74991c0b2f9daec42ea9d0f193e6e25fc974deb779a45134d6258b09
+size 345647

models/subword_ngram/ang_3gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "subword",
+  "language": "ang",
+  "unique_ngrams": 28365,
+  "total_ngrams": 3242991
+}

models/subword_ngram/ang_4gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4867afacfc08d62f62d898a90eed09e96af5d562b3a54b91b0acde18a8c8c96
+size 1444787

models/subword_ngram/ang_4gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "subword",
+  "language": "ang",
+  "unique_ngrams": 125189,
+  "total_ngrams": 3237856
+}

models/tokenizer/ang_tokenizer_16k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10e4d0807a40b872f29528b5c23f9841503ebf5aab866002da9ea38b52f50d27
+size 506547

models/tokenizer/ang_tokenizer_32k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:334a8b8b3a0ebc409c568a256e9614276ad9185b0da9661339bec148fc9baa34
+size 779127

models/tokenizer/ang_tokenizer_64k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b03ece4ccfeeb1e095675628468e8394e4b1b8f9b3dc73660c628eabfb345361
+size 1372424

models/tokenizer/ang_tokenizer_8k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:951f11d514320c9af6f06489cd33c223767ce8110d61ae10acfd10c2f4c1d819
+size 370408

models/vocabulary/ang_vocabulary.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8293a2194f8ee9bccc9ba1e789d27affd0a5326a35adb253b8aa12dd20a95bee
+size 558369

models/vocabulary/ang_vocabulary_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "ang",
+  "vocabulary_size": 32745,
+  "statistics": {
+    "type_token_ratio": 0.18429092362255856,
+    "coverage": {
+      "top_100": 0.3267196163644538,
+      "top_1000": 0.5182676457365503,
+      "top_5000": 0.6850711823767421,
+      "top_10000": 0.7579399570408112
+    },
+    "hapax_count": 59488,
+    "hapax_ratio": 0.6449752257868658,
+    "total_documents": 5135
+  }
+}

models/word_markov/ang_markov_ctx1_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a00910d0170554920791ac4446cf494c25bbb7bd8699047f94b43700084dc3c
+size 3148814

models/word_markov/ang_markov_ctx1_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "word",
+  "language": "ang",
+  "unique_contexts": 92131,
+  "total_transitions": 614610
+}

models/word_markov/ang_markov_ctx2_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d2a28e8e4c8f3183b5a563700bef62bed372f66fd54af666b1adc514bcaee1a
+size 6460657

models/word_markov/ang_markov_ctx2_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "word",
+  "language": "ang",
+  "unique_contexts": 340526,
+  "total_transitions": 609476
+}

models/word_markov/ang_markov_ctx3_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4013cfaf6a7e5fb5b05fd16c6fb82a6b07a3986ca2ba084cb4249f829ba48fe6
+size 8659399

models/word_markov/ang_markov_ctx3_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "word",
+  "language": "ang",
+  "unique_contexts": 488936,
+  "total_transitions": 604342
+}

models/word_markov/ang_markov_ctx4_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:248fdd8a9842bf29f9fb3c6c2994608e1182bf86996d9b007f01871b5c8c6fb8
+size 9896185

models/word_markov/ang_markov_ctx4_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "word",
+  "language": "ang",
+  "unique_contexts": 535590,
+  "total_transitions": 599209
+}

models/word_ngram/ang_2gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77969c802023316308f47b559b4654b98f131e91a3e596b4349461e965a4d7a0
+size 171492

models/word_ngram/ang_2gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "word",
+  "language": "ang",
+  "unique_ngrams": 12127,
+  "total_ngrams": 614610
+}

models/word_ngram/ang_3gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddb4a50ee6552832977f0d2b46d56ea53cfa83bf4d67e4c1cc23feaf8a715561
+size 216617

models/word_ngram/ang_3gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "word",
+  "language": "ang",
+  "unique_ngrams": 13432,
+  "total_ngrams": 609476
+}

models/word_ngram/ang_4gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c0cefee4edc601a3bdad8fbf3b8f6e175478c16df0bfcb72ac4db6fe7981f8e
+size 402550

models/word_ngram/ang_4gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "word",
+  "language": "ang",
+  "unique_ngrams": 22900,
+  "total_ngrams": 604342
+}