Upload all models and assets for bg (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 +1,569 @@
+---
+language: bg
+language_name: Bulgarian
+language_family: slavic_south
+tags:
+  - wikilangs
+  - nlp
+  - tokenizer
+  - embeddings
+  - n-gram
+  - markov
+  - wikipedia
+  - monolingual
+  - family-slavic_south
+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: 3.805
+  - name: best_isotropy
+    type: isotropy
+    value: 0.7912
+  - name: vocabulary_size
+    type: vocab
+    value: 960471
+generated: 2025-12-28
+---
+
+# Bulgarian - Wikilangs Models
+## Comprehensive Research Report & Full Ablation Study
+
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **Bulgarian** 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** | 3.140x | 3.10 | 0.0447% | 3,031,858 |
+| **16k** | 3.405x | 3.36 | 0.0485% | 2,795,587 |
+| **32k** | 3.631x | 3.59 | 0.0517% | 2,621,828 |
+| **64k** | 3.805x 🏆 | 3.76 | 0.0542% | 2,501,712 |
+
+### Tokenization Examples
+
+Below are sample sentences tokenized with each vocabulary size:
+
+**Sample 1:** `Събития 
+
+ Родени 
+
+ Починали 
+ 28 юни – Андрей I, велик княз на Владимир-Суздал`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁събития ▁родени ▁починали ▁ 2 8 ▁юни ▁– ▁андрей ▁i ... (+9 more)` | 19 |
+| 16k | `▁събития ▁родени ▁починали ▁ 2 8 ▁юни ▁– ▁андрей ▁i ... (+9 more)` | 19 |
+| 32k | `▁събития ▁родени ▁починали ▁ 2 8 ▁юни ▁– ▁андрей ▁i ... (+9 more)` | 19 |
+| 64k | `▁събития ▁родени ▁починали ▁ 2 8 ▁юни ▁– ▁андрей ▁i ... (+8 more)` | 18 |
+
+**Sample 2:** `Събития 
+
+ Родени 
+
+ Починали`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁събития ▁родени ▁починали` | 3 |
+| 16k | `▁събития ▁родени ▁починали` | 3 |
+| 32k | `▁събития ▁родени ▁починали` | 3 |
+| 64k | `▁събития ▁родени ▁починали` | 3 |
+
+**Sample 3:** `Хайд може да се отнася за:
+
+ Градове
+ Хайд, град в Англия
+
+ Окръзи в САЩ
+ Хайд (...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁ха йд ▁може ▁да ▁се ▁отнася ▁за : ▁градове ▁ха ... (+28 more)` | 38 |
+| 16k | `▁ха йд ▁може ▁да ▁се ▁отнася ▁за : ▁градове ▁ха ... (+25 more)` | 35 |
+| 32k | `▁хайд ▁може ▁да ▁се ▁отнася ▁за : ▁градове ▁хайд , ... (+20 more)` | 30 |
+| 64k | `▁хайд ▁може ▁да ▁се ▁отнася ▁за : ▁градове ▁хайд , ... (+20 more)` | 30 |
+
+
+### Key Findings
+
+- **Best Compression:** 64k achieves 3.805x compression
+- **Lowest UNK Rate:** 8k with 0.0447% 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** | 171,622 🏆 | 17.39 | 2,295,348 | 9.8% | 21.1% |
+| **2-gram** | 445 🏆 | 8.80 | 25,460 | 58.1% | 96.2% |
+| **3-gram** | 975,598 | 19.90 | 5,989,128 | 3.6% | 10.5% |
+| **3-gram** | 4,162 | 12.02 | 263,503 | 21.9% | 59.8% |
+| **4-gram** | 3,001,891 | 21.52 | 11,403,312 | 1.9% | 5.9% |
+| **4-gram** | 25,670 | 14.65 | 1,642,365 | 10.2% | 31.2% |
+
+### Top 5 N-grams by Size
+
+**2-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `г .` | 1,208,709 |
+| 2 | `категория :` | 853,964 |
+| 3 | `) ,` | 479,143 |
+| 4 | `) .` | 331,723 |
+| 5 | `. в` | 330,654 |
+
+**3-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `г . ,` | 116,060 |
+| 2 | `г . )` | 88,901 |
+| 3 | `( ) е` | 84,347 |
+| 4 | `източници категория :` | 82,359 |
+| 5 | `г . в` | 77,398 |
+
+**4-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `. източници категория :` | 51,218 |
+| 2 | `категория : родени в` | 43,839 |
+| 3 | `. н . е` | 40,096 |
+| 4 | `н . е .` | 40,004 |
+| 5 | `пр . н .` | 39,838 |
+
+
+### Key Findings
+
+- **Best Perplexity:** 2-gram with 445
+- **Entropy Trend:** Decreases with larger n-grams (more predictable)
+- **Coverage:** Top-1000 patterns cover ~31% 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** | 0.7433 | 1.674 | 8.73 | 2,274,140 | 25.7% |
+| **1** | 1.3957 | 2.631 | 10.79 | 7,585 | 0.0% |
+| **2** | 0.4514 | 1.367 | 2.92 | 19,851,824 | 54.9% |
+| **2** | 0.8817 | 1.843 | 6.63 | 81,858 | 11.8% |
+| **3** | 0.2135 | 1.159 | 1.55 | 58,038,366 | 78.7% |
+| **3** | 0.9116 | 1.881 | 5.23 | 542,599 | 8.8% |
+| **4** | 0.1027 🏆 | 1.074 | 1.21 | 90,112,776 | 89.7% |
+| **4** | 0.7326 🏆 | 1.662 | 3.61 | 2,836,397 | 26.7% |
+
+### Generated Text Samples
+
+Below are text samples generated from each Markov chain model:
+
+**Context Size 1:**
+
+1. `. през 1954 , и образование . “ , и на концертните може да отпият от`
+2. `, pomacentrus taeniometopon и определя характера , oktober . за купата на виенския университет на ол...`
+3. `на ветроходен спорт . структурата на непорочното зачатие на място по икономика “ . и цигулка`
+
+**Context Size 2:**
+
+1. `г . беше върната на сикст iv обявява конрад за маркиз акиле патерно , който по това`
+2. `категория : починали в тирана . след нашествието на унгарските интереси . правителството обявява наг...`
+3. `) , гръцки андартски деец , полковник станчов е български футболист 30 септември 1944 ) ташев ,`
+
+**Context Size 3:**
+
+1. `г . , ∞ 1309 за португалския крал афонсу v . той е и рекордьор за мъже в`
+2. `г . ) 1923 г . се завръща в пазарджик и председател на управителния съвет на rheinmetall са`
+3. `( ) е английски професионален футболист , който играе като вратар и се състезава в долните дивизии н...`
+
+**Context Size 4:**
+
+1. `. източници категория : литературни термини категория : научна фантастика категория : английски писа...`
+2. `категория : родени в софия категория : починали от рак категория : родени през 1710 година категория...`
+3. `. н . е . релефът изобразява човек на колесница с четири колела с четири коня и запаси за`
+
+
+### Key Findings
+
+- **Best Predictability:** Context-4 with 89.7% predictability
+- **Branching Factor:** Decreases with context size (more deterministic)
+- **Memory Trade-off:** Larger contexts require more storage (2,836,397 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 | 960,471 |
+| Total Tokens | 113,282,257 |
+| Mean Frequency | 117.94 |
+| Median Frequency | 4 |
+| Frequency Std Dev | 9016.46 |
+
+### Most Common Words
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | на | 6,000,585 |
+| 2 | в | 3,189,920 |
+| 3 | и | 3,170,304 |
+| 4 | е | 2,177,841 |
+| 5 | от | 2,156,538 |
+| 6 | за | 1,349,594 |
+| 7 | се | 1,262,248 |
+| 8 | г | 1,219,255 |
+| 9 | с | 1,091,067 |
+| 10 | категория | 861,853 |
+
+### Least Common Words (from vocabulary)
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | мъндън | 2 |
+| 2 | талиевия | 2 |
+| 3 | carbonato | 2 |
+| 4 | tallio | 2 |
+| 5 | tlhco3 | 2 |
+| 6 | разр | 2 |
+| 7 | mичман | 2 |
+| 8 | барутхана | 2 |
+| 9 | азадлу | 2 |
+| 10 | шталаг | 2 |
+
+### Zipf's Law Analysis
+
+| Metric | Value |
+|--------|-------|
+| Zipf Coefficient | 0.9535 |
+| R² (Goodness of Fit) | 0.996716 |
+| Adherence Quality | **excellent** |
+
+### Coverage Analysis
+
+| Top N Words | Coverage |
+|-------------|----------|
+| Top 100 | 33.9% |
+| Top 1,000 | 53.3% |
+| Top 5,000 | 70.0% |
+| Top 10,000 | 77.0% |
+
+### Key Findings
+
+- **Zipf Compliance:** R²=0.9967 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 33.9% of corpus
+- **Long Tail:** 950,471 words needed for remaining 23.0% 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** | 784,943 | 32 | 3.285 | 0.948 | 0.7912 🏆 |
+| **mono_64d** | 784,943 | 64 | 3.715 | 0.928 | 0.7726 |
+| **mono_128d** | 784,943 | 128 | 4.153 | 0.959 | 0.7213 |
+| **embeddings_enhanced** | 0 | 0 | 0.000 | 0.000 | 0.0000 |
+
+### Key Findings
+
+- **Best Isotropy:** mono_32d with 0.7912 (more uniform distribution)
+- **Dimension Trade-off:** Higher dimensions capture more semantics but reduce isotropy
+- **Vocabulary Coverage:** All models cover 784,943 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 (3.81x) with low UNK rate |
+| N-gram | **5-gram** | Lowest perplexity (445) |
+| Markov | **Context-4** | Highest predictability (89.7%) |
+| 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 05:10:25*

models/embeddings/monolingual/bg_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62cdb2428baa77bd04299975524021723a8ae2e1a65dd3825b63a90f909791a6
+size 1847208986

models/embeddings/monolingual/bg_128d.meta.json

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

models/embeddings/monolingual/bg_128d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "bg",
+  "dimension": 128,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 128,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 784943
+}

models/embeddings/monolingual/bg_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6694d4b5c6d5d590970d067a729784b32deae9005f9da64f4af63b62d72fd36
+size 476372762

models/embeddings/monolingual/bg_32d.meta.json

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

models/embeddings/monolingual/bg_32d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "bg",
+  "dimension": 32,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 32,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 784943
+}

models/embeddings/monolingual/bg_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c165692c56ff4e358f7d5166c0669be3db289fba438592d3be05f8b7c4c00096
+size 933318170

models/embeddings/monolingual/bg_64d.meta.json

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

models/embeddings/monolingual/bg_64d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "bg",
+  "dimension": 64,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 64,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 784943
+}

models/subword_markov/bg_markov_ctx1_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd9e037e79ad13bab48d8eeb19edaefaa7ce9c28ff82a5a07912b24ad1dbaae5
+size 553134

models/subword_markov/bg_markov_ctx1_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "subword",
+  "language": "bg",
+  "unique_contexts": 7585,
+  "total_transitions": 748247572
+}

models/subword_markov/bg_markov_ctx2_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da9040e708654bc57ac14b886e7470e3d89de042483e7f8f4538568b3ebe116c
+size 4217824

models/subword_markov/bg_markov_ctx2_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "subword",
+  "language": "bg",
+  "unique_contexts": 81858,
+  "total_transitions": 747940816
+}

models/subword_markov/bg_markov_ctx3_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:747f1607cf523956d879650cfe449c357884a3cdb39c4513d39a8f6b5d5b567f
+size 20926937

models/subword_markov/bg_markov_ctx3_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "subword",
+  "language": "bg",
+  "unique_contexts": 542599,
+  "total_transitions": 747634060
+}

models/subword_markov/bg_markov_ctx4_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:580594f4c1b5219691b6c56a2d73434635ee09c032b5149e7db74064a6acb65d
+size 84747354

models/subword_markov/bg_markov_ctx4_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "subword",
+  "language": "bg",
+  "unique_contexts": 2836397,
+  "total_transitions": 747327304
+}

models/subword_ngram/bg_2gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67c8881953c951402dd6fdba26bfc9a9c946443e0e1ed1881dc14bec6d626101
+size 355833

models/subword_ngram/bg_2gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "subword",
+  "language": "bg",
+  "unique_ngrams": 25460,
+  "total_ngrams": 748247572
+}

models/subword_ngram/bg_3gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32fc5c6fa7c26961c9a441b21558109617f745d911b211285e49e74ef54da5f0
+size 3306412

models/subword_ngram/bg_3gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "subword",
+  "language": "bg",
+  "unique_ngrams": 263503,
+  "total_ngrams": 747940816
+}

models/subword_ngram/bg_4gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c7e026096ab3694db6e24aef689cbcdc5f2d4039e87e3716741d3e37f776c44
+size 21218611

models/subword_ngram/bg_4gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "subword",
+  "language": "bg",
+  "unique_ngrams": 1642365,
+  "total_ngrams": 747634060
+}

models/tokenizer/bg_tokenizer_16k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33c6944293a101991ea008701e30e9031604db2e52cd7e6ecdbe45d45ab069c8
+size 586044

models/tokenizer/bg_tokenizer_32k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ccfb2867a7a268411643ba09befd8b1a8af862bd8d825a39e257ee8375a64e1
+size 955868

models/tokenizer/bg_tokenizer_64k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb7a345ba1ca1260c37faed404185ea71b0a9e5a3a0f07e7c61a1c0ad0629672
+size 1713039

models/tokenizer/bg_tokenizer_8k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b413db4fd21958b13201f36fb037f0b71a092a43deac8595b08548d310f5741
+size 406965

models/vocabulary/bg_vocabulary.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c185c72f951b2a0b535c14e0db0c089b83b539e8e448f1c25ef74024e17cc9b5
+size 15858834

models/vocabulary/bg_vocabulary_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "bg",
+  "vocabulary_size": 960471,
+  "statistics": {
+    "type_token_ratio": 0.019838880375008985,
+    "coverage": {
+      "top_100": 0.3353316626354778,
+      "top_1000": 0.5264381386495268,
+      "top_5000": 0.6920902735329456,
+      "top_10000": 0.7614777009866214
+    },
+    "hapax_count": 1312970,
+    "hapax_ratio": 0.577525433912734,
+    "total_documents": 306756
+  }
+}

models/word_markov/bg_markov_ctx1_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86a3382cdb32fda378af8795d72ce4a5795e3868e31ac776f3e98413d191d916
+size 229240553

models/word_markov/bg_markov_ctx1_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "word",
+  "language": "bg",
+  "unique_contexts": 2274140,
+  "total_transitions": 144455456
+}

models/word_markov/bg_markov_ctx2_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bdb579252dcf1dfd5fefb2c16e59e61f8751fd9617e88957a50be897d5e7879a
+size 956955138

models/word_markov/bg_markov_ctx2_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "word",
+  "language": "bg",
+  "unique_contexts": 19851824,
+  "total_transitions": 144148700
+}

models/word_markov/bg_markov_ctx3_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c183bc9fb35d56a1d36a35c676087357430dbb1cbcd68cb95f60cff2617319bc
+size 2001387377

models/word_markov/bg_markov_ctx3_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "word",
+  "language": "bg",
+  "unique_contexts": 58038366,
+  "total_transitions": 143841946
+}

models/word_markov/bg_markov_ctx4_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6564dac2fd521efc9d195f037c4404f1b2abcc7d2611680d6efa069f8016a52d
+size 2808416888

models/word_markov/bg_markov_ctx4_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "word",
+  "language": "bg",
+  "unique_contexts": 90112776,
+  "total_transitions": 143535554
+}

models/word_ngram/bg_2gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:272feaf50f9917bf57090f67146ae6656c60415bd274fca46f527c0e6530255f
+size 52579739

models/word_ngram/bg_2gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "word",
+  "language": "bg",
+  "unique_ngrams": 2295348,
+  "total_ngrams": 144455456
+}

models/word_ngram/bg_3gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8e57efc5bab54762e166db24201676eefd1be3a4e403d456ed1b2c14bff5f13
+size 156219488

models/word_ngram/bg_3gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "word",
+  "language": "bg",
+  "unique_ngrams": 5989128,
+  "total_ngrams": 144148700
+}

models/word_ngram/bg_4gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0464427fba8035e05c53ccf7726df3c35aaed23d22e1af103e9f36d28c44eda
+size 316212735

models/word_ngram/bg_4gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "word",
+  "language": "bg",
+  "unique_ngrams": 11403312,
+  "total_ngrams": 143841946
+}