Upload all models and assets for as (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 +1,563 @@
+---
+language: as
+language_name: AS
+language_family: indoaryan_eastern
+tags:
+  - wikilangs
+  - nlp
+  - tokenizer
+  - embeddings
+  - n-gram
+  - markov
+  - wikipedia
+  - monolingual
+  - family-indoaryan_eastern
+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: 4.519
+  - name: best_isotropy
+    type: isotropy
+    value: 0.8484
+  - name: vocabulary_size
+    type: vocab
+    value: 69383
+generated: 2025-12-27
+---
+
+# AS - Wikilangs Models
+## Comprehensive Research Report & Full Ablation Study
+
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **AS** 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** | 3.390x | 3.36 | 0.0640% | 1,495,826 |
+| **16k** | 3.830x | 3.80 | 0.0723% | 1,324,012 |
+| **32k** | 4.212x | 4.18 | 0.0795% | 1,203,836 |
+| **64k** | 4.519x 🏆 | 4.48 | 0.0853% | 1,122,074 |
+
+### Tokenization Examples
+
+Below are sample sentences tokenized with each vocabulary size:
+
+**Sample 1:** `ইউৰেনাচ নামে তলৰ প্ৰৱন্ধসমূহ বুজাব পাৰে:
+ ইউৰেনাচ: সৌৰজগতৰ সপ্তম গ্ৰহ
+ ইউৰেনাচ (...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁ইউ ৰে না চ ▁নামে ▁তলৰ ▁প্ৰৱ ন্ধ সমূহ ▁বুজাব ... (+28 more)` | 38 |
+| 16k | `▁ইউৰে না চ ▁নামে ▁তলৰ ▁প্ৰৱ ন্ধ সমূহ ▁বুজাব ▁পাৰে ... (+23 more)` | 33 |
+| 32k | `▁ইউৰে নাচ ▁নামে ▁তলৰ ▁প্ৰৱন্ধ সমূহ ▁বুজাব ▁পাৰে : ▁ইউৰে ... (+18 more)` | 28 |
+| 64k | `▁ইউৰে নাচ ▁নামে ▁তলৰ ▁প্ৰৱন্ধ সমূহ ▁বুজাব ▁পাৰে : ▁ইউৰে ... (+17 more)` | 27 |
+
+**Sample 2:** `শ্ৰেণী:ভাৰতীয় অভিনেত্ৰী
+শ্ৰেণী:জীৱিত ব্যক্তি
+শ্ৰেণী:ভাৰতীয় ৰাজনীতিবিদ
+শ্ৰেণী:ভ...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁শ্ৰেণী : ভাৰতীয় ▁অভিনেত্ৰী ▁শ্ৰেণী : জীৱিত ▁ব্যক্তি ▁শ্ৰেণী : ... (+9 more)` | 19 |
+| 16k | `▁শ্ৰেণী : ভাৰতীয় ▁অভিনেত্ৰী ▁শ্ৰেণী : জীৱিত ▁ব্যক্তি ▁শ্ৰেণী : ... (+8 more)` | 18 |
+| 32k | `▁শ্ৰেণী : ভাৰতীয় ▁অভিনেত্ৰী ▁শ্ৰেণী : জীৱিত ▁ব্যক্তি ▁শ্ৰেণী : ... (+8 more)` | 18 |
+| 64k | `▁শ্ৰেণী : ভাৰতীয় ▁অভিনেত্ৰী ▁শ্ৰেণী : জীৱিত ▁ব্যক্তি ▁শ্ৰেণী : ... (+7 more)` | 17 |
+
+**Sample 3:** `তাল বুলিলে তলৰ পৃষ্ঠাসমূহ বুজাব পাৰে:
+
+তাল (বাদ্যযন্ত্ৰ)
+তাল (সংগীত)
+তাল (ফল)`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁তাল ▁বুল িলে ▁তলৰ ▁পৃষ্ঠ াসমূহ ▁বুজাব ▁পাৰে : ▁তাল ... (+13 more)` | 23 |
+| 16k | `▁তাল ▁বুলিলে ▁তলৰ ▁পৃষ্ঠ াসমূহ ▁বুজাব ▁পাৰে : ▁তাল ▁( ... (+11 more)` | 21 |
+| 32k | `▁তাল ▁বুলিলে ▁তলৰ ▁পৃষ্ঠ াসমূহ ▁বুজাব ▁পাৰে : ▁তাল ▁( ... (+11 more)` | 21 |
+| 64k | `▁তাল ▁বুলিলে ▁তলৰ ▁পৃষ্ঠাসমূহ ▁বুজাব ▁পাৰে : ▁তাল ▁( বাদ্যযন্ত্ৰ ... (+9 more)` | 19 |
+
+
+### Key Findings
+
+- **Best Compression:** 64k achieves 4.519x compression
+- **Lowest UNK Rate:** 8k with 0.0640% 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** | 2,255 🏆 | 11.14 | 144,083 | 37.8% | 74.6% |
+| **2-gram** | 712 🏆 | 9.48 | 14,933 | 47.2% | 91.5% |
+| **3-gram** | 20,803 | 14.34 | 558,435 | 14.3% | 40.2% |
+| **3-gram** | 6,279 | 12.62 | 135,792 | 17.8% | 52.9% |
+| **4-gram** | 109,133 | 16.74 | 1,749,119 | 7.8% | 23.0% |
+| **4-gram** | 34,064 | 15.06 | 727,754 | 9.4% | 29.5% |
+
+### Top 5 N-grams by Size
+
+**2-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `া ৰ` | 608,846 |
+| 2 | `য ়` | 511,382 |
+| 3 | `্ ৰ` | 409,497 |
+| 4 | `প ্` | 302,441 |
+| 5 | `ি ল` | 296,895 |
+
+**3-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `য ় া` | 192,871 |
+| 2 | `ি য ়` | 190,336 |
+| 3 | `ী য ়` | 164,267 |
+| 4 | `ত ে ও` | 138,798 |
+| 5 | `ে ও ঁ` | 138,171 |
+
+**4-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `ত ে ও ঁ` | 137,993 |
+| 2 | `ি য ় া` | 118,996 |
+| 3 | `ছ ি ল ।` | 81,488 |
+| 4 | `ি ছ ি ল` | 71,229 |
+| 5 | `শ ্ ৰ ে` | 58,374 |
+
+
+### Key Findings
+
+- **Best Perplexity:** 2-gram with 712
+- **Entropy Trend:** Decreases with larger n-grams (more predictable)
+- **Coverage:** Top-1000 patterns cover ~29% 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** | 0.5873 | 1.502 | 6.01 | 178,570 | 41.3% |
+| **1** | 1.0627 | 2.089 | 7.96 | 5,555 | 0.0% |
+| **2** | 0.3625 | 1.286 | 2.70 | 1,073,346 | 63.7% |
+| **2** | 0.9491 | 1.931 | 6.70 | 44,224 | 5.1% |
+| **3** | 0.2925 | 1.225 | 2.07 | 2,893,182 | 70.7% |
+| **3** | 0.8311 | 1.779 | 4.51 | 296,427 | 16.9% |
+| **4** | 0.2472 🏆 | 1.187 | 1.70 | 5,984,484 | 75.3% |
+| **4** | 0.6476 🏆 | 1.567 | 2.96 | 1,337,967 | 35.2% |
+
+### Generated Text Samples
+
+Below are text samples generated from each Markov chain model:
+
+**Context Size 1:**
+
+1. `া লয ় যদ ি ল ৰ া ন ে । পৰ ি ছ া উৰ`
+2. `্ চত ভ া শ ্ ট ো ক ভ ু ১ট া ’ ল ্`
+3. `ি য ় , ক া ল । ইয ় ন কৰ ম ৃ ষকসকলক প`
+
+**Context Size 2:**
+
+1. `া ৰ ব ি ব া গ া ঁ ৱত কচ ্ ‌ ২০০৫ – ২০১০ চনত`
+2. `য ় া ৰ ব ি প ি ছত anthropological survey of hinduism today , march 2009`
+3. `্ ৰ া থম ি ক ব ি ন ্ দ ্ ব া ৰ ব ি`
+
+**Context Size 3:**
+
+1. `য ় া ষ ্ ট ে চনত আশ ্ ৰফ ে চৰ১৯৬৮ফ ি ল ্ ল ী`
+2. `ি য ় া ধ ী ন হয ় আৰ ু মহ ি ল া ৰ দ ৌ`
+3. `ী য ় া ল ো কগ া থ া ক ে ব ি ব া হৰ ব`
+
+**Context Size 4:**
+
+1. `ত ে ও ঁ ৰ প ু ৰণ ি ড ি মৰ ু গছৰ পৰ া প ্ ৰ`
+2. `ি য ় া চ ি ন ্ ত ু স ে ই বছৰত ে ত ে ও ঁ`
+3. `ছ ি ল । ইয ় া ৰ অন ু ম ো দ ী ৰ ্ ঘদ ি ন`
+
+
+### Key Findings
+
+- **Best Predictability:** Context-4 with 75.3% predictability
+- **Branching Factor:** Decreases with context size (more deterministic)
+- **Memory Trade-off:** Larger contexts require more storage (1,337,967 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 | 69,383 |
+| Total Tokens | 22,699,994 |
+| Mean Frequency | 327.17 |
+| Median Frequency | 4 |
+| Frequency Std Dev | 13137.34 |
+
+### Most Common Words
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | ৰ | 1,751,445 |
+| 2 | ত | 1,226,468 |
+| 3 | ব | 988,164 |
+| 4 | য | 983,360 |
+| 5 | ক | 951,103 |
+| 6 | ন | 878,240 |
+| 7 | ল | 815,336 |
+| 8 | প | 750,050 |
+| 9 | ম | 610,982 |
+| 10 | স | 521,058 |
+
+### Least Common Words (from vocabulary)
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | zns | 2 |
+| 2 | tuklas | 2 |
+| 3 | bilqees | 2 |
+| 4 | surbhi | 2 |
+| 5 | ullasamga | 2 |
+| 6 | utsahamga | 2 |
+| 7 | manchস | 2 |
+| 8 | swfl | 2 |
+| 9 | manhunt | 2 |
+| 10 | megamodel | 2 |
+
+### Zipf's Law Analysis
+
+| Metric | Value |
+|--------|-------|
+| Zipf Coefficient | 1.5303 |
+| R² (Goodness of Fit) | 0.997399 |
+| Adherence Quality | **excellent** |
+
+### Coverage Analysis
+
+| Top N Words | Coverage |
+|-------------|----------|
+| Top 100 | 78.6% |
+| Top 1,000 | 93.9% |
+| Top 5,000 | 97.8% |
+| Top 10,000 | 98.6% |
+
+### Key Findings
+
+- **Zipf Compliance:** R²=0.9974 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 78.6% of corpus
+- **Long Tail:** 59,383 words needed for remaining 1.4% 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** | 113,429 | 32 | 3.011 | 0.609 | 0.8484 🏆 |
+| **mono_64d** | 113,429 | 64 | 3.498 | 0.644 | 0.8482 |
+| **mono_128d** | 113,429 | 128 | 4.089 | 0.704 | 0.8254 |
+| **embeddings_enhanced** | 0 | 0 | 0.000 | 0.000 | 0.0000 |
+
+### Key Findings
+
+- **Best Isotropy:** mono_32d with 0.8484 (more uniform distribution)
+- **Dimension Trade-off:** Higher dimensions capture more semantics but reduce isotropy
+- **Vocabulary Coverage:** All models cover 113,429 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 (4.52x) with low UNK rate |
+| N-gram | **5-gram** | Lowest perplexity (712) |
+| Markov | **Context-4** | Highest predictability (75.3%) |
+| 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 18:45:07*

models/embeddings/monolingual/as_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7294ee96f3b0c61ba7c46af8e348b551645a26d74537adc45861ba9171026d24
+size 1143382361

models/embeddings/monolingual/as_128d.meta.json

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

models/embeddings/monolingual/as_128d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "as",
+  "dimension": 128,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 128,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 113429
+}

models/embeddings/monolingual/as_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c0125149319853f09b65df93575e1bf1665cf0cf98a0a4e52df3248b94a776d
+size 288268889

models/embeddings/monolingual/as_32d.meta.json

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

models/embeddings/monolingual/as_32d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "as",
+  "dimension": 32,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 32,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 113429
+}

models/embeddings/monolingual/as_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b271a32a3a16aa6c5905bd4ef669972e7fa9d2fcad872330bcbfe8211a6cbf6
+size 573306713

models/embeddings/monolingual/as_64d.meta.json

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

models/embeddings/monolingual/as_64d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "as",
+  "dimension": 64,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 64,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 113429
+}

models/subword_markov/as_markov_ctx1_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:304b81823be6914953bac63df3fb3c2d3d5e01a1f9b4b8f4a8878717e903a2e3
+size 319610

models/subword_markov/as_markov_ctx1_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "subword",
+  "language": "as",
+  "unique_contexts": 5555,
+  "total_transitions": 65140856
+}

models/subword_markov/as_markov_ctx2_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:072a0a77d5f9b149b82195090a2f7ec01b569c51640f1c61e11c74d314891b20
+size 2244616

models/subword_markov/as_markov_ctx2_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "subword",
+  "language": "as",
+  "unique_contexts": 44224,
+  "total_transitions": 65120064
+}

models/subword_markov/as_markov_ctx3_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7c22336529a0ded174b0241383329d3154a9858bd06639bd78f20c1e786cf00
+size 10618800

models/subword_markov/as_markov_ctx3_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "subword",
+  "language": "as",
+  "unique_contexts": 296427,
+  "total_transitions": 65099272
+}

models/subword_markov/as_markov_ctx4_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fba617be292f432879d2b6f88a252644b452cbe07fc436c4d4b98898b0dd2f33
+size 34480970

models/subword_markov/as_markov_ctx4_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "subword",
+  "language": "as",
+  "unique_contexts": 1337967,
+  "total_transitions": 65078480
+}

models/subword_ngram/as_2gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13f2a80a67d004f512051da3fb0f08c9a906b943f4d8c6e170756448689a3095
+size 206387

models/subword_ngram/as_2gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "subword",
+  "language": "as",
+  "unique_ngrams": 14933,
+  "total_ngrams": 65140856
+}

models/subword_ngram/as_3gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:566ad2045c2c52180f92dc29b097fad25c84b062a65319808f9431fad7f538cb
+size 1714165

models/subword_ngram/as_3gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "subword",
+  "language": "as",
+  "unique_ngrams": 135792,
+  "total_ngrams": 65120064
+}

models/subword_ngram/as_4gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:645ec172809e1ba4475428602980c6d539af311c472d34ab195bc24111837333
+size 9556409

models/subword_ngram/as_4gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "subword",
+  "language": "as",
+  "unique_ngrams": 727754,
+  "total_ngrams": 65099272
+}

models/tokenizer/as_tokenizer_16k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92c4a0d6d3435730960b337cbe00d6561bb754c82a66840b10b3effdf114a9c0
+size 627322

models/tokenizer/as_tokenizer_32k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfd4440400274db9e0e200c32c56c9b2f8b5d1e4a80cd88b665224da4685840d
+size 1042352

models/tokenizer/as_tokenizer_64k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f1a070ac3d05309f6fc98145f97c0714ee5da0ecebf4b99231acd5001077e67f
+size 1892860

models/tokenizer/as_tokenizer_8k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3ac9ae44b7976a9a581b8c12e38d04180fcbc38165d660d6b96b537c7954700
+size 427062

models/vocabulary/as_vocabulary.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd7c512db097eb79bf2751927085a424d4d4be0b149ff56f3b19b0846f20e25
+size 1236046

models/vocabulary/as_vocabulary_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "as",
+  "vocabulary_size": 69383,
+  "statistics": {
+    "type_token_ratio": 0.007789323101723796,
+    "coverage": {
+      "top_100": 0.7824795758310844,
+      "top_1000": 0.9345050339631166,
+      "top_5000": 0.9733725115168742,
+      "top_10000": 0.9816343824731659
+    },
+    "hapax_count": 108278,
+    "hapax_ratio": 0.6094640917252521,
+    "total_documents": 20792
+  }
+}

models/word_markov/as_markov_ctx1_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d6086bc53947c43f34edcda8e0a034c9779c3d6c44de35c31cc2ba88f8aa774
+size 9625358

models/word_markov/as_markov_ctx1_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "word",
+  "language": "as",
+  "unique_contexts": 178570,
+  "total_transitions": 41886369
+}

models/word_markov/as_markov_ctx2_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1542cf7905c52aefaff0db643b5107e37e6e10fccc17a863ba95e7ee172740fd
+size 27241048

models/word_markov/as_markov_ctx2_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "word",
+  "language": "as",
+  "unique_contexts": 1073346,
+  "total_transitions": 41865578
+}

models/word_markov/as_markov_ctx3_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09d22e9d5414b02524a41246b9410c78ab3bf52bcade25ce445168e3445ee4d8
+size 63382094

models/word_markov/as_markov_ctx3_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "word",
+  "language": "as",
+  "unique_contexts": 2893182,
+  "total_transitions": 41844788
+}

models/word_markov/as_markov_ctx4_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05e3849b2bcb7ebe1f9ae63c37ed5ea9973336b012b7a2c60020150cd53b6a7d
+size 124869442

models/word_markov/as_markov_ctx4_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "word",
+  "language": "as",
+  "unique_contexts": 5984484,
+  "total_transitions": 41824000
+}

models/word_ngram/as_2gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46ca72b889b191be977809c46f9e5c601830c691249878aa061223811c6c065a
+size 1981156

models/word_ngram/as_2gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "word",
+  "language": "as",
+  "unique_ngrams": 144083,
+  "total_ngrams": 41886369
+}

models/word_ngram/as_3gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b16c54cf2957c0d5c96809a5a7c39da4fe7c1dd6c13b3345c3073e0cc4a9da1
+size 8106622

models/word_ngram/as_3gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "word",
+  "language": "as",
+  "unique_ngrams": 558435,
+  "total_ngrams": 41865578
+}

models/word_ngram/as_4gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c567cb651bd6273d82124f43b6ac4426b9ece8a16302648d995617ab25da8d18
+size 26163353

models/word_ngram/as_4gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "word",
+  "language": "as",
+  "unique_ngrams": 1749119,
+  "total_ngrams": 41844788
+}