Upload all models and assets for be (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 +1,560 @@
+---
+language: be
+language_name: BE
+language_family: slavic_east
+tags:
+  - wikilangs
+  - nlp
+  - tokenizer
+  - embeddings
+  - n-gram
+  - markov
+  - wikipedia
+  - monolingual
+  - family-slavic_east
+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: 3.609
+  - name: best_isotropy
+    type: isotropy
+    value: 0.6652
+  - name: vocabulary_size
+    type: vocab
+    value: 834514
+generated: 2025-12-28
+---
+
+# BE - Wikilangs Models
+## Comprehensive Research Report & Full Ablation Study
+
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **BE** 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** | 2.945x | 2.88 | 0.0099% | 395,619 |
+| **16k** | 3.198x | 3.13 | 0.0107% | 364,322 |
+| **32k** | 3.434x | 3.36 | 0.0115% | 339,332 |
+| **64k** | 3.609x 🏆 | 3.53 | 0.0121% | 322,873 |
+
+### Tokenization Examples
+
+Below are sample sentences tokenized with each vocabulary size:
+
+**Sample 1:** `Адаі () — вёска ў Акнянскім раёне Адэскай вобласці Украіны.
+
+ Крыніцы 
+
+Катэгоры...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁ада і ▁() ▁— ▁вёска ▁ў ▁ак ня нскім ▁раёне ... (+14 more)` | 24 |
+| 16k | `▁ада і ▁() ▁— ▁вёска ▁ў ▁ак ня нскім ▁раёне ... (+13 more)` | 23 |
+| 32k | `▁ада і ▁() ▁— ▁вёска ▁ў ▁ак ня нскім ▁раёне ... (+12 more)` | 22 |
+| 64k | `▁ада і ▁() ▁— ▁вёска ▁ў ▁ак нянскім ▁раёне ▁адэскай ... (+11 more)` | 21 |
+
+**Sample 2:** `Назву Асманіе маюць: горад і правінцыя ў Турцыі.`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁назву ▁ас мані е ▁маюць : ▁горад ▁і ▁правінцыя ▁ў ... (+2 more)` | 12 |
+| 16k | `▁назву ▁ас мані е ▁маюць : ▁горад ▁і ▁правінцыя ▁ў ... (+2 more)` | 12 |
+| 32k | `▁назву ▁ас мані е ▁маюць : ▁горад ▁і ▁правінцыя ▁ў ... (+2 more)` | 12 |
+| 64k | `▁назву ▁ас мані е ▁маюць : ▁горад ▁і ▁правінцыя ▁ў ... (+2 more)` | 12 |
+
+**Sample 3:** `M21 (каталог Месье) — рассеянае скопішча ў сузор'і Стральца.
+
+Катэгорыя:Астранам...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁m 2 1 ▁( ката лог ▁месье ) ▁— ▁рас ... (+39 more)` | 49 |
+| 16k | `▁m 2 1 ▁( ката лог ▁месье ) ▁— ▁рассея ... (+36 more)` | 46 |
+| 32k | `▁m 2 1 ▁( ката лог ▁месье ) ▁— ▁рассеянае ... (+35 more)` | 45 |
+| 64k | `▁m 2 1 ▁( каталог ▁месье ) ▁— ▁рассеянае ▁скопішча ... (+34 more)` | 44 |
+
+
+### Key Findings
+
+- **Best Compression:** 64k achieves 3.609x compression
+- **Lowest UNK Rate:** 8k with 0.0099% 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** | 79,086 🏆 | 16.27 | 1,374,832 | 15.1% | 29.5% |
+| **2-gram** | 534 🏆 | 9.06 | 19,108 | 52.5% | 95.6% |
+| **3-gram** | 294,087 | 18.17 | 2,802,568 | 8.0% | 20.4% |
+| **3-gram** | 5,046 | 12.30 | 199,361 | 17.9% | 56.0% |
+| **4-gram** | 713,543 | 19.44 | 5,150,781 | 6.7% | 16.8% |
+| **4-gram** | 30,738 | 14.91 | 1,228,443 | 8.4% | 28.2% |
+
+### Top 5 N-grams by Size
+
+**2-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `0 ,` | 1,884,848 |
+| 2 | `катэгорыя :` | 589,578 |
+| 3 | `. у` | 390,424 |
+| 4 | `) .` | 239,852 |
+| 5 | `) —` | 224,324 |
+
+**3-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `0 , 10` | 188,226 |
+| 2 | `0 , 09` | 178,159 |
+| 3 | `0 , 11` | 136,658 |
+| 4 | `0 , 08` | 133,746 |
+| 5 | `0 , 07` | 96,108 |
+
+**4-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `катэгорыя : населеныя пункты` | 53,581 |
+| 2 | `) — вёска ў` | 38,538 |
+| 3 | `. уваходзіць у склад` | 36,002 |
+| 4 | `. — с .` | 29,359 |
+| 5 | `, 44 0 ,` | 29,322 |
+
+
+### Key Findings
+
+- **Best Perplexity:** 2-gram with 534
+- **Entropy Trend:** Decreases with larger n-grams (more predictable)
+- **Coverage:** Top-1000 patterns cover ~28% 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** | 0.5334 | 1.447 | 6.21 | 2,760,646 | 46.7% |
+| **1** | 0.5275 | 1.441 | 4.42 | 17,006 | 47.2% |
+| **2** | 0.3318 | 1.259 | 2.22 | 17,120,073 | 66.8% |
+| **2** | 0.7155 | 1.642 | 5.63 | 75,115 | 28.5% |
+| **3** | 0.1509 | 1.110 | 1.38 | 37,905,234 | 84.9% |
+| **3** | 0.8770 | 1.837 | 5.02 | 422,960 | 12.3% |
+| **4** | 0.0724 🏆 | 1.051 | 1.16 | 52,452,226 | 92.8% |
+| **4** | 0.7380 🏆 | 1.668 | 3.59 | 2,124,122 | 26.2% |
+
+### Generated Text Samples
+
+Below are text samples generated from each Markov chain model:
+
+**Context Size 1:**
+
+1. `, аэрапорт знаходзіцца ў свеце вядома з сакрысціямі і іншыя матэрыялы , створаных самім фактам ,`
+2. `. 3 адкрыты чэмпіянат па 21 кастрычніка 1957 — супраціўнае звычайна была прысвоена званне « за`
+3. `0 , лячэбны факультэт кіравання версіямі , евангелле паводле ніпурскага , а ў сялянскай сям ’`
+
+**Context Size 2:**
+
+1. `0 , 41 0 , 11 1912897 0 , 47 0 , 44 0 , 81 0`
+2. `катэгорыя : лігатуры кірыліцы катэгорыя : паняцці індуізму катэгорыя : будынкі і збудаванні бабруйск...`
+3. `. у 1932 — 1982 ) , святы рпцз . 12 . 3 . игнатенко , в`
+
+**Context Size 3:**
+
+1. `0 , 10 47053 sd 1 , 20 0 , 09 21985 e - so 0 , 93`
+2. `0 , 09 637479 0 , 39 0 , 11 88688 0 , 60 0 , 08 310690`
+3. `0 , 11 348939 0 , 60 0 , 08 2587753 0 , 73 0 , 07 838578`
+
+**Context Size 4:**
+
+1. `катэгорыя : населеныя пункты гарадэнкіўскага раёна катэгорыя : населеныя пункты на дняпры катэгорыя ...`
+2. `) — вёска ў сосныцкім раёне чарнігаўскай вобласці украіны . крыніцы катэгорыя : населеныя пункты аст...`
+3. `. уваходзіць у склад сямлёўскага сельскага паселішча . гісторыя 25 сакавіка 1918 года згодна з трэця...`
+
+
+### Key Findings
+
+- **Best Predictability:** Context-4 with 92.8% predictability
+- **Branching Factor:** Decreases with context size (more deterministic)
+- **Memory Trade-off:** Larger contexts require more storage (2,124,122 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 | 834,514 |
+| Total Tokens | 59,867,674 |
+| Mean Frequency | 71.74 |
+| Median Frequency | 4 |
+| Frequency Std Dev | 3746.92 |
+
+### Most Common Words
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | 0 | 1,972,029 |
+| 2 | і | 1,334,509 |
+| 3 | у | 1,241,263 |
+| 4 | ў | 1,163,394 |
+| 5 | з | 869,929 |
+| 6 | на | 710,982 |
+| 7 | катэгорыя | 590,266 |
+| 8 | года | 368,272 |
+| 9 | да | 291,578 |
+| 10 | годзе | 258,701 |
+
+### Least Common Words (from vocabulary)
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | мурашава | 2 |
+| 2 | девятке | 2 |
+| 3 | дэкунаў | 2 |
+| 4 | iovine | 2 |
+| 5 | аёвіну | 2 |
+| 6 | джэніка | 2 |
+| 7 | мэрылінам | 2 |
+| 8 | сардэшная | 2 |
+| 9 | івасю | 2 |
+| 10 | стеценко | 2 |
+
+### Zipf's Law Analysis
+
+| Metric | Value |
+|--------|-------|
+| Zipf Coefficient | 0.9824 |
+| R² (Goodness of Fit) | 0.995868 |
+| Adherence Quality | **excellent** |
+
+### Coverage Analysis
+
+| Top N Words | Coverage |
+|-------------|----------|
+| Top 100 | 28.3% |
+| Top 1,000 | 50.1% |
+| Top 5,000 | 67.4% |
+| Top 10,000 | 74.4% |
+
+### Key Findings
+
+- **Zipf Compliance:** R²=0.9959 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 28.3% of corpus
+- **Long Tail:** 824,514 words needed for remaining 25.6% 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** | 563,209 | 32 | 4.064 | 1.812 | 0.6194 |
+| **mono_64d** | 563,209 | 64 | 4.475 | 1.623 | 0.6528 |
+| **mono_128d** | 563,209 | 128 | 4.940 | 1.374 | 0.6652 🏆 |
+| **embeddings_enhanced** | 0 | 0 | 0.000 | 0.000 | 0.0000 |
+
+### Key Findings
+
+- **Best Isotropy:** mono_128d with 0.6652 (more uniform distribution)
+- **Dimension Trade-off:** Higher dimensions capture more semantics but reduce isotropy
+- **Vocabulary Coverage:** All models cover 563,209 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 (3.61x) with low UNK rate |
+| N-gram | **5-gram** | Lowest perplexity (534) |
+| Markov | **Context-4** | Highest predictability (92.8%) |
+| 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: 2025-12-28 02:15:50*

models/embeddings/monolingual/be_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:348b8d47c3adfcab8cf7ed5c5c126bac20a90632d48e0a96d1ac58cf34bdd206
+size 1615310423

models/embeddings/monolingual/be_128d.meta.json

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

models/embeddings/monolingual/be_128d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "be",
+  "dimension": 128,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 128,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 563209
+}

models/embeddings/monolingual/be_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:95ba41adfe852e9124f1f4e2a999e53493f8ac5a72620a903103616d33e1d3f8
+size 414765911

models/embeddings/monolingual/be_32d.meta.json

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

models/embeddings/monolingual/be_32d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "be",
+  "dimension": 32,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 32,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 563209
+}

models/embeddings/monolingual/be_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0aae5fa847ee5bb9a765d41f00f80662622108ba7c2da83c491b6ee1aaff3cbc
+size 814947415

models/embeddings/monolingual/be_64d.meta.json

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

models/embeddings/monolingual/be_64d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "be",
+  "dimension": 64,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 64,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 563209
+}

models/subword_markov/be_markov_ctx1_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:613af873a5c8ee531d4fd55c96b5d2024b23df806ba9c46820017e322c5c5213
+size 581483

models/subword_markov/be_markov_ctx1_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "subword",
+  "language": "be",
+  "unique_contexts": 17006,
+  "total_transitions": 433862151
+}

models/subword_markov/be_markov_ctx2_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e0c1db4d2e461008750e32dfa8de1dc9c3409c516afaa54674aef8e0c2ab8830
+size 3387673

models/subword_markov/be_markov_ctx2_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "subword",
+  "language": "be",
+  "unique_contexts": 75115,
+  "total_transitions": 433604630
+}

models/subword_markov/be_markov_ctx3_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab73b13dfa84faefbb17dd279eb97e6291463168a9832e9f2d357b36483bf97b
+size 16387590

models/subword_markov/be_markov_ctx3_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "subword",
+  "language": "be",
+  "unique_contexts": 422960,
+  "total_transitions": 433347109
+}

models/subword_markov/be_markov_ctx4_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85fb9c5f2e85e087f9fd863b630a9776994de89274e1d9581cb75947969ee1f6
+size 64453185

models/subword_markov/be_markov_ctx4_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "subword",
+  "language": "be",
+  "unique_contexts": 2124122,
+  "total_transitions": 433089588
+}

models/subword_ngram/be_2gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c0e30a48df55f96d4bf2b8dc208c62d144ebf34b1a70cfcf18baab44a7ed51fd
+size 267023

models/subword_ngram/be_2gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "subword",
+  "language": "be",
+  "unique_ngrams": 19108,
+  "total_ngrams": 433862151
+}

models/subword_ngram/be_3gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f8fca4862a13251eca0d62c746599da6a1c83bee3b4fd07ef8d26fd3782087b
+size 2504735

models/subword_ngram/be_3gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "subword",
+  "language": "be",
+  "unique_ngrams": 199361,
+  "total_ngrams": 433604630
+}

models/subword_ngram/be_4gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:258fc772791e206747c2feee7088b7f28d6c39225074e25abce4217fc6aa0cba
+size 15989176

models/subword_ngram/be_4gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "subword",
+  "language": "be",
+  "unique_ngrams": 1228443,
+  "total_ngrams": 433347109
+}

models/tokenizer/be_tokenizer_16k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1402065668cb58d0465980ced26b6d3d48c3faaebd6dda93b82b5cbe1aabbe99
+size 589454

models/tokenizer/be_tokenizer_32k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e0a60bc4e1c3914f6172171b41c29fd097713bfa2666961c50cd45a0a57c495
+size 968380

models/tokenizer/be_tokenizer_64k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7dd70bc8b3492061d8b4f5bfbcfdc3a51c6d45fd46e0c5ca413a9b0977d1bbb1
+size 1750083

models/tokenizer/be_tokenizer_8k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:335477bcdfc1cb61bc835c34d31b30091119b1b7473fa74c622aca874b26dab0
+size 408070

models/vocabulary/be_vocabulary.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67ae10394b92721eb663679625db202a64a880e70109d0b21040a910e8804efb
+size 13628498

models/vocabulary/be_vocabulary_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "be",
+  "vocabulary_size": 834514,
+  "statistics": {
+    "type_token_ratio": 0.044671476336453776,
+    "coverage": {
+      "top_100": 0.2744355925705539,
+      "top_1000": 0.48551658368338324,
+      "top_5000": 0.6526621297329156,
+      "top_10000": 0.7207621763882552
+    },
+    "hapax_count": 1925896,
+    "hapax_ratio": 0.697684764219808,
+    "total_documents": 257521
+  }
+}

models/word_markov/be_markov_ctx1_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f573302999a72f7aa3e9d142ae089b7b678ffd15700999d05203009d925512c0
+size 196984346

models/word_markov/be_markov_ctx1_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "word",
+  "language": "be",
+  "unique_contexts": 2760646,
+  "total_transitions": 81548782
+}

models/word_markov/be_markov_ctx2_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:63dffe71cfd86f4ff61c37367e2dbee74fa2f7c50e05a4825b7739595e1b2155
+size 696994958

models/word_markov/be_markov_ctx2_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "word",
+  "language": "be",
+  "unique_contexts": 17120073,
+  "total_transitions": 81291262
+}

models/word_markov/be_markov_ctx3_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd8f08d3be91c4c054ae6f0aaa136416885527a2d29113415ab337c834789d0e
+size 1226716369

models/word_markov/be_markov_ctx3_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "word",
+  "language": "be",
+  "unique_contexts": 37905234,
+  "total_transitions": 81033747
+}

models/word_markov/be_markov_ctx4_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33587a71fb18a5c4cc838d154494581511cee2776f0934c16659de4c95e12078
+size 1617896422

models/word_markov/be_markov_ctx4_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "word",
+  "language": "be",
+  "unique_contexts": 52452226,
+  "total_transitions": 80776256
+}

models/word_ngram/be_2gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e4fc961aa5fc25059f776fbf54bffac241e518ff469fdaf22b2abdea5a86e4d
+size 33153735

models/word_ngram/be_2gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "word",
+  "language": "be",
+  "unique_ngrams": 1374832,
+  "total_ngrams": 81548782
+}

models/word_ngram/be_3gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58da42f46883e63cb423ab5d1f9a73b92e98b59bc8da341ae4aaed966d43beef
+size 75522602

models/word_ngram/be_3gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "word",
+  "language": "be",
+  "unique_ngrams": 2802568,
+  "total_ngrams": 81291262
+}

models/word_ngram/be_4gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5e3e3cc455ebf7ad153584a4edb334f24017f1636a5e52b898c56b9a3a24755
+size 147804215

models/word_ngram/be_4gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "word",
+  "language": "be",
+  "unique_ngrams": 5150781,
+  "total_ngrams": 81033747
+}