Upload all models and assets for ay (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 +1,565 @@
+---
+language: ay
+language_name: AY
+language_family: american_aymara
+tags:
+  - wikilangs
+  - nlp
+  - tokenizer
+  - embeddings
+  - n-gram
+  - markov
+  - wikipedia
+  - monolingual
+  - family-american_aymara
+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: 4.018
+  - name: best_isotropy
+    type: isotropy
+    value: 0.7731
+  - name: vocabulary_size
+    type: vocab
+    value: 25329
+generated: 2025-12-27
+---
+
+# AY - Wikilangs Models
+## Comprehensive Research Report & Full Ablation Study
+
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **AY** 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** | 3.240x | 3.13 | 0.2414% | 169,821 |
+| **16k** | 3.519x | 3.40 | 0.2623% | 156,332 |
+| **32k** | 3.774x | 3.65 | 0.2812% | 145,778 |
+| **64k** | 4.018x 🏆 | 3.88 | 0.2994% | 136,918 |
+
+### Tokenization Examples
+
+Below are sample sentences tokenized with each vocabulary size:
+
+**Sample 1:** `Phuyu (), jamach'inakana janchipa imxatiriquña yänaka. Lulina phuyupaxa k'acha q...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁phuyu ▁(), ▁jamach ' i nakana ▁janchipa ▁im x atiri ... (+22 more)` | 32 |
+| 16k | `▁phuyu ▁(), ▁jamach ' i nakana ▁janchipa ▁im x atiri ... (+22 more)` | 32 |
+| 32k | `▁phuyu ▁(), ▁jamach ' inakana ▁janchipa ▁im x atiri quña ... (+18 more)` | 28 |
+| 64k | `▁phuyu ▁(), ▁jamach ' inakana ▁janchipa ▁im x atiri quña ... (+16 more)` | 26 |
+
+**Sample 2:** `1920 - mara.
+
+ Yuriña 
+Toshiro Mifune.
+
+ Jiwaña 
+
+ Uruyaña 
+Categoría:Maranaka`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁ 1 9 2 0 ▁- ▁mara . ▁yuriña ▁tos ... (+11 more)` | 21 |
+| 16k | `▁ 1 9 2 0 ▁- ▁mara . ▁yuriña ▁tos ... (+8 more)` | 18 |
+| 32k | `▁ 1 9 2 0 ▁- ▁mara . ▁yuriña ▁toshiro ... (+7 more)` | 17 |
+| 64k | `▁ 1 9 2 0 ▁- ▁mara . ▁yuriña ▁toshiro ... (+7 more)` | 17 |
+
+**Sample 3:** `Oaxaca (), marka istadu Mïxiku.
+ Oaxaca nayriri marka: Oaxaca de Juárez.
+
+Catego...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁o axa ca ▁(), ▁marka ▁istadu ▁mïxiku . ▁o axa ... (+13 more)` | 23 |
+| 16k | `▁oaxaca ▁(), ▁marka ▁istadu ▁mïxiku . ▁oaxaca ▁nayriri ▁marka : ... (+7 more)` | 17 |
+| 32k | `▁oaxaca ▁(), ▁marka ▁istadu ▁mïxiku . ▁oaxaca ▁nayriri ▁marka : ... (+7 more)` | 17 |
+| 64k | `▁oaxaca ▁(), ▁marka ▁istadu ▁mïxiku . ▁oaxaca ▁nayriri ▁marka : ... (+7 more)` | 17 |
+
+
+### Key Findings
+
+- **Best Compression:** 64k achieves 4.018x compression
+- **Lowest UNK Rate:** 8k with 0.2414% 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** | 1,629 🏆 | 10.67 | 14,390 | 42.1% | 70.8% |
+| **2-gram** | 335 🏆 | 8.39 | 3,058 | 62.7% | 98.6% |
+| **3-gram** | 2,436 | 11.25 | 23,975 | 39.1% | 65.6% |
+| **3-gram** | 2,346 | 11.20 | 22,295 | 28.7% | 70.7% |
+| **4-gram** | 4,646 | 12.18 | 47,728 | 34.2% | 57.4% |
+| **4-gram** | 9,050 | 13.14 | 97,814 | 19.6% | 48.1% |
+
+### Top 5 N-grams by Size
+
+**2-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `' a` | 22,172 |
+| 2 | `jisk '` | 16,424 |
+| 3 | `categoría :` | 12,913 |
+| 4 | `t '` | 11,645 |
+| 5 | `' aqa` | 10,721 |
+
+**3-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `jisk ' a` | 16,399 |
+| 2 | `t ' aqa` | 10,718 |
+| 3 | `a t '` | 10,632 |
+| 4 | `' a t` | 10,632 |
+| 5 | `' aqa suyu` | 8,509 |
+
+**4-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `' a t '` | 10,632 |
+| 2 | `a t ' aqa` | 10,628 |
+| 3 | `jisk ' a t` | 10,616 |
+| 4 | `t ' aqa suyu` | 8,508 |
+| 5 | `: jisk ' a` | 4,073 |
+
+
+### Key Findings
+
+- **Best Perplexity:** 2-gram with 335
+- **Entropy Trend:** Decreases with larger n-grams (more predictable)
+- **Coverage:** Top-1000 patterns cover ~48% 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** | 0.6043 | 1.520 | 3.69 | 65,172 | 39.6% |
+| **1** | 0.9793 | 1.971 | 7.56 | 953 | 2.1% |
+| **2** | 0.2085 | 1.155 | 1.54 | 239,858 | 79.2% |
+| **2** | 1.0055 | 2.008 | 6.09 | 7,201 | 0.0% |
+| **3** | 0.0898 | 1.064 | 1.20 | 369,469 | 91.0% |
+| **3** | 0.8336 | 1.782 | 3.92 | 43,833 | 16.6% |
+| **4** | 0.0469 🏆 | 1.033 | 1.11 | 443,066 | 95.3% |
+| **4** | 0.6060 🏆 | 1.522 | 2.52 | 171,955 | 39.4% |
+
+### Generated Text Samples
+
+Below are text samples generated from each Markov chain model:
+
+**Context Size 1:**
+
+1. `' a t ' aqa suyu kantarawi kantarawi tarata , punu jach ' a t '`
+2. `. wali alwat tukuñkamaw yuntampi wayt asitapjaniw . caritatis studium , bibliaxa mä jisk ' aq`
+3. `: freistaat thüringen suyu ( mendoza , ukax kunas manq ' a suyu asu 1889 mara`
+
+**Context Size 2:**
+
+1. `' a t ' aqa suyu ( piruw ) categoría : warmi categoría : warmi : consuelo`
+2. `jisk ' a t ' aqa suyu asu jaqinaka kurakanaka 2011 - 2014 : bernabé martí .`
+3. `categoría : warmi categoría : jiwäwi 1890 categoría : jisk ' a suyu nayriri marka ) ,`
+
+**Context Size 3:**
+
+1. `jisk ' a t ' aqa suyu asu jaqinaka kurakanaka 2011 - 2014 : wilfredo felimón vargas cotrado`
+2. `t ' aqa suyuxa ( kastilla aru : distrito de carabayllo ) nisqaqa huk jisk ' a t`
+3. `a t ' aqa suyu ( piruw ) categoría : jisk ' a t ' aqa suyu ;`
+
+**Context Size 4:**
+
+1. `' a t ' aqa suyuwa , amarujawira jach ' a suyu . nayra sarnaqawi kamasqa chichu phaxsi 29`
+2. `a t ' aqa suyuwa , 22px | la libertad la libertad jach ' a suyu , wankayu jisk`
+3. `jisk ' a t ' aqa suyu ( piruw ) categoría : jisk ' a suyu munisipyu web (`
+
+
+### Key Findings
+
+- **Best Predictability:** Context-4 with 95.3% predictability
+- **Branching Factor:** Decreases with context size (more deterministic)
+- **Memory Trade-off:** Larger contexts require more storage (171,955 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 | 25,329 |
+| Total Tokens | 596,604 |
+| Mean Frequency | 23.55 |
+| Median Frequency | 4 |
+| Frequency Std Dev | 287.74 |
+
+### Most Common Words
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | a | 23,874 |
+| 2 | jisk | 16,567 |
+| 3 | suyu | 14,640 |
+| 4 | categoría | 12,921 |
+| 5 | t | 11,916 |
+| 6 | de | 11,655 |
+| 7 | aqa | 10,725 |
+| 8 | jach | 7,168 |
+| 9 | piruw | 5,369 |
+| 10 | jaqinaka | 5,120 |
+
+### Least Common Words (from vocabulary)
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | wirnisa | 2 |
+| 2 | sawaru | 2 |
+| 3 | tuminku | 2 |
+| 4 | urupawa | 2 |
+| 5 | capitalapawa | 2 |
+| 6 | halloween | 2 |
+| 7 | uttar | 2 |
+| 8 | pradesh | 2 |
+| 9 | quqanakampi | 2 |
+| 10 | jawiranakat | 2 |
+
+### Zipf's Law Analysis
+
+| Metric | Value |
+|--------|-------|
+| Zipf Coefficient | 1.1050 |
+| R² (Goodness of Fit) | 0.997739 |
+| Adherence Quality | **excellent** |
+
+### Coverage Analysis
+
+| Top N Words | Coverage |
+|-------------|----------|
+| Top 100 | 46.7% |
+| Top 1,000 | 73.4% |
+| Top 5,000 | 87.8% |
+| Top 10,000 | 93.3% |
+
+### Key Findings
+
+- **Zipf Compliance:** R²=0.9977 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 46.7% of corpus
+- **Long Tail:** 15,329 words needed for remaining 6.7% 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** | 10,362 | 32 | 4.255 | 0.785 | 0.7731 🏆 |
+| **mono_64d** | 10,362 | 64 | 4.432 | 0.729 | 0.4752 |
+| **mono_128d** | 10,362 | 128 | 4.521 | 0.731 | 0.1287 |
+| **embeddings_enhanced** | 0 | 0 | 0.000 | 0.000 | 0.0000 |
+
+### Key Findings
+
+- **Best Isotropy:** mono_32d with 0.7731 (more uniform distribution)
+- **Dimension Trade-off:** Higher dimensions capture more semantics but reduce isotropy
+- **Vocabulary Coverage:** All models cover 10,362 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 (4.02x) with low UNK rate |
+| N-gram | **5-gram** | Lowest perplexity (335) |
+| Markov | **Context-4** | Highest predictability (95.3%) |
+| 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 20:48:58*

models/embeddings/monolingual/ay_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:593d231b5c862c91d99a298b35d7a8a92af5889f7195a1096550a6cf493db26c
+size 1034781490

models/embeddings/monolingual/ay_128d.meta.json

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

models/embeddings/monolingual/ay_128d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "ay",
+  "dimension": 128,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 128,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 10362
+}

models/embeddings/monolingual/ay_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1148550baa373f84d4bcb57f3b890769155df89eb33f5ca94e031220a6178d22
+size 258823474

models/embeddings/monolingual/ay_32d.meta.json

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

models/embeddings/monolingual/ay_32d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "ay",
+  "dimension": 32,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 32,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 10362
+}

models/embeddings/monolingual/ay_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11f64944a6949f805320a9a0366bed7ddb5434a527e1796dc32cbf647cb04617
+size 517476146

models/embeddings/monolingual/ay_64d.meta.json

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

models/embeddings/monolingual/ay_64d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "ay",
+  "dimension": 64,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 64,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 10362
+}

models/subword_markov/ay_markov_ctx1_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a501564aa7bab749452432f4296ba393aa4888e0600291301c8acb9b46087910
+size 60606

models/subword_markov/ay_markov_ctx1_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "subword",
+  "language": "ay",
+  "unique_contexts": 953,
+  "total_transitions": 4432265
+}

models/subword_markov/ay_markov_ctx2_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:66f2745ff633512b340bdf8201dd98192e37615cc9042243528ef27b8cf231bb
+size 339502

models/subword_markov/ay_markov_ctx2_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "subword",
+  "language": "ay",
+  "unique_contexts": 7201,
+  "total_transitions": 4426837
+}

models/subword_markov/ay_markov_ctx3_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa6ef1a990648f3d87315c9fbfaba1abc08319ae83afd78f5259b68a6e9622fd
+size 1261852

models/subword_markov/ay_markov_ctx3_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "subword",
+  "language": "ay",
+  "unique_contexts": 43833,
+  "total_transitions": 4421409
+}

models/subword_markov/ay_markov_ctx4_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a65fd0941e7426fe960b1fee56ea748a94605c4412ad68c9094e3ae9643915cc
+size 3425718

models/subword_markov/ay_markov_ctx4_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "subword",
+  "language": "ay",
+  "unique_contexts": 171955,
+  "total_transitions": 4415981
+}

models/subword_ngram/ay_2gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b4e4597fceea0d7c8119438b61b0b4477dd0e8b8fde6f7f68febdb690bf5333
+size 41037

models/subword_ngram/ay_2gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "subword",
+  "language": "ay",
+  "unique_ngrams": 3058,
+  "total_ngrams": 4432265
+}

models/subword_ngram/ay_3gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00b41467becbb77b0b605997b7fefc5dfb74440266d63b49cbe7a413cc01e080
+size 270869

models/subword_ngram/ay_3gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "subword",
+  "language": "ay",
+  "unique_ngrams": 22295,
+  "total_ngrams": 4426837
+}

models/subword_ngram/ay_4gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47d09c14900c8b9ae9a7d8df02df50e5b5485695a6fc0b931bef0d14df9ce8b1
+size 1123557

models/subword_ngram/ay_4gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "subword",
+  "language": "ay",
+  "unique_ngrams": 97814,
+  "total_ngrams": 4421409
+}

models/tokenizer/ay_tokenizer_16k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bacd6b6b67eecb376d4d6587a319913a3beba867c2d3084857448cfac3757e48
+size 503226

models/tokenizer/ay_tokenizer_32k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff707a0d80083da5281395b809fd5fa3d8f96da2ef90f415d80f52d3eef2121a
+size 777476

models/tokenizer/ay_tokenizer_64k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e56a192de4ae4f2d2a0ff3ab901c71c57726cc607781da09d9ccbcc07365fc0
+size 1370860

models/tokenizer/ay_tokenizer_8k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06a75da91da908126d1fafd3569251faaa4952976ef9181f20ec4f333bf6005d
+size 371206

models/vocabulary/ay_vocabulary.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de3094a5549fa43e268c332bf5eb6ff19a145b1edf8b5ccfbb2e2d399a2272e4
+size 427486

models/vocabulary/ay_vocabulary_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "ay",
+  "vocabulary_size": 25329,
+  "statistics": {
+    "type_token_ratio": 0.10227701301150292,
+    "coverage": {
+      "top_100": 0.43783393047960273,
+      "top_1000": 0.6883022188698221,
+      "top_5000": 0.8231488465648376,
+      "top_10000": 0.8748287133069332
+    },
+    "hapax_count": 39756,
+    "hapax_ratio": 0.6108319889375432,
+    "total_documents": 5428
+  }
+}

models/word_markov/ay_markov_ctx1_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9235d44170713caeb8fc80b3dba5782d1f796ed803992b0531ceb65661a19aa4
+size 2137508

models/word_markov/ay_markov_ctx1_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "word",
+  "language": "ay",
+  "unique_contexts": 65172,
+  "total_transitions": 888867
+}

models/word_markov/ay_markov_ctx2_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56c569c3d809da2135b500827e40c8db0797fbcb18c0c4e781c6af0595edd8c7
+size 4305196

models/word_markov/ay_markov_ctx2_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "word",
+  "language": "ay",
+  "unique_contexts": 239858,
+  "total_transitions": 883439
+}

models/word_markov/ay_markov_ctx3_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9f338f5f12683bff8d312db02c7920bbeba0835ee748bb9152d82925a294286
+size 6062956

models/word_markov/ay_markov_ctx3_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "word",
+  "language": "ay",
+  "unique_contexts": 369469,
+  "total_transitions": 878012
+}

models/word_markov/ay_markov_ctx4_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49f18f764c5f4fb242282040dee518fce36fdc22459fbf93e81df9c68865595e
+size 7355175

models/word_markov/ay_markov_ctx4_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "word",
+  "language": "ay",
+  "unique_contexts": 443066,
+  "total_transitions": 872587
+}

models/word_ngram/ay_2gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa957dcf8bd6a2084464d7f441e000f2d48cc10f34f222fd5e62ba3ee4cc9c65
+size 198830

models/word_ngram/ay_2gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "word",
+  "language": "ay",
+  "unique_ngrams": 14390,
+  "total_ngrams": 888867
+}

models/word_ngram/ay_3gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2a940917149974038eb678645a67aaf32fec2cfe5ad07fcd47a47b766ae7e81
+size 367171

models/word_ngram/ay_3gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "word",
+  "language": "ay",
+  "unique_ngrams": 23975,
+  "total_ngrams": 883439
+}

models/word_ngram/ay_4gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7c85c1f3bf8ded5071febc58f660d755b21c15511a8d4b49dff4d6eb7ff1566
+size 742404

models/word_ngram/ay_4gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "word",
+  "language": "ay",
+  "unique_ngrams": 47728,
+  "total_ngrams": 878012
+}