Upload all models and assets for bn (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 +1,578 @@
+---
+language: bn
+language_name: Bengali
+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: 4.983
+  - name: best_isotropy
+    type: isotropy
+    value: 0.8133
+  - name: vocabulary_size
+    type: vocab
+    value: 250463
+generated: 2025-12-28
+---
+
+# Bengali - Wikilangs Models
+## Comprehensive Research Report & Full Ablation Study
+
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **Bengali** 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** | 3.731x | 3.70 | 0.0813% | 2,819,192 |
+| **16k** | 4.228x | 4.19 | 0.0921% | 2,487,647 |
+| **32k** | 4.650x | 4.61 | 0.1013% | 2,262,081 |
+| **64k** | 4.983x 🏆 | 4.94 | 0.1085% | 2,111,066 |
+
+### Tokenization Examples
+
+Below are sample sentences tokenized with each vocabulary size:
+
+**Sample 1:** `ইতিহাস 
+
+ ঘটনাবলী 
+
+ জন্ম 
+
+ মৃত্যু 
+মিশেল আফলাক
+
+ ছুটি এবং অন্যান্য 
+
+ বহিঃসংযো...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁ইতিহাস ▁ঘটনাবলী ▁জন্ম ▁মৃত্যু ▁মিশ েল ▁আ ফল াক ▁ছ ... (+16 more)` | 26 |
+| 16k | `▁ইতিহাস ▁ঘটনাবলী ▁জন্ম ▁মৃত্যু ▁মিশেল ▁আ ফল াক ▁ছুটি ▁এবং ... (+12 more)` | 22 |
+| 32k | `▁ইতিহাস ▁ঘটনাবলী ▁জন্ম ▁মৃত্যু ▁মিশেল ▁আ ফল াক ▁ছুটি ▁এবং ... (+11 more)` | 21 |
+| 64k | `▁ইতিহাস ▁ঘটনাবলী ▁জন্ম ▁মৃত্যু ▁মিশেল ▁আ ফল াক ▁ছুটি ▁এবং ... (+11 more)` | 21 |
+
+**Sample 2:** `চিলমারী ইউনিয়ন নামে বাংলাদেশে মোট ২টি ইউনিয়ন রয়েছে। যথা:
+ চিলমারী ইউনিয়ন, চি...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁চিল ম ারী ▁ইউনিয়ন ▁নামে ▁বাংলাদেশে ▁মোট ▁২টি ▁ইউনিয়ন ▁রয়েছে ... (+34 more)` | 44 |
+| 16k | `▁চিল মারী ▁ইউনিয়ন ▁নামে ▁বাংলাদেশে ▁মোট ▁২টি ▁ইউনিয়ন ▁রয়েছে । ... (+27 more)` | 37 |
+| 32k | `▁চিল মারী ▁ইউনিয়ন ▁নামে ▁বাংলাদেশে ▁মোট ▁২টি ▁ইউনিয়ন ▁রয়েছে । ... (+24 more)` | 34 |
+| 64k | `▁চিলমারী ▁ইউনিয়ন ▁নামে ▁বাংলাদেশে ▁মোট ▁২টি ▁ইউনিয়ন ▁রয়েছে । ▁যথা ... (+19 more)` | 29 |
+
+**Sample 3:** `ইতিহাস 
+
+ ঘটনাবলী 
+
+ জন্ম 
+
+ মৃত্যু 
+
+ ছুটি এবং অন্যান্য 
+
+ বহিঃসংযোগ 
+
+বিষয়শ্র...`
+
+| Vocab | Tokens | Count |
+|-------|--------|-------|
+| 8k | `▁ইতিহাস ▁ঘটনাবলী ▁জন্ম ▁মৃত্যু ▁ছ ুটি ▁এবং ▁অন্যান্য ▁বহিঃসংযোগ ▁বিষয়শ্রেণী ... (+12 more)` | 22 |
+| 16k | `▁ইতিহাস ▁ঘটনাবলী ▁জন্ম ▁মৃত্যু ▁ছুটি ▁এবং ▁অন্যান্য ▁বহিঃসংযোগ ▁বিষয়শ্রেণী : ... (+9 more)` | 19 |
+| 32k | `▁ইতিহাস ▁ঘটনাবলী ▁জন্ম ▁মৃত্যু ▁ছুটি ▁এবং ▁অন্যান্য ▁বহিঃসংযোগ ▁বিষয়শ্রেণী : ... (+9 more)` | 19 |
+| 64k | `▁ইতিহাস ▁ঘটনাবলী ▁জন্ম ▁মৃত্যু ▁ছুটি ▁এবং ▁অন্যান্য ▁বহিঃসংযোগ ▁বিষয়শ্রেণী : ... (+7 more)` | 17 |
+
+
+### Key Findings
+
+- **Best Compression:** 64k achieves 4.983x compression
+- **Lowest UNK Rate:** 8k with 0.0813% 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** | 1,729 🏆 | 10.76 | 514,602 | 43.5% | 78.4% |
+| **2-gram** | 617 🏆 | 9.27 | 33,180 | 50.6% | 92.9% |
+| **3-gram** | 17,027 | 14.06 | 2,087,238 | 16.5% | 45.3% |
+| **3-gram** | 5,420 | 12.40 | 328,172 | 19.8% | 55.7% |
+| **4-gram** | 101,821 | 16.64 | 6,958,889 | 8.9% | 25.5% |
+| **4-gram** | 30,323 | 14.89 | 1,934,479 | 10.4% | 31.7% |
+
+### Top 5 N-grams by Size
+
+**2-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `া র` | 6,292,455 |
+| 2 | `ে র` | 5,308,140 |
+| 3 | `্ র` | 4,972,400 |
+| 4 | `য ়` | 4,738,290 |
+| 5 | `্ য` | 4,282,578 |
+
+**3-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `্ য া` | 2,328,201 |
+| 2 | `্ র ে` | 1,699,814 |
+| 3 | `ি য ়` | 1,654,809 |
+| 4 | `া য ়` | 1,550,712 |
+| 5 | `য ় া` | 1,358,961 |
+
+**4-grams:**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `শ ্ র ে` | 1,127,020 |
+| 2 | `ব ি ষয ়` | 1,117,741 |
+| 3 | `্ র ে ণ` | 1,092,505 |
+| 4 | `র ে ণ ী` | 1,077,879 |
+| 5 | `় শ ্ র` | 1,049,641 |
+
+
+### Key Findings
+
+- **Best Perplexity:** 2-gram with 617
+- **Entropy Trend:** Decreases with larger n-grams (more predictable)
+- **Coverage:** Top-1000 patterns cover ~32% of corpus
+- **Recommendation:** 4-gram or 5-gram for best predictive performance
+
+---
+## 3. Markov Chain Evaluation
+
+![Markov Entropy](visualizations/markov_entropy.png)
+
+![Markov Branching](visualizations/markov_branching.png)
+
+### Results
+
+| Context | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
+|---------|-------------|------------|------------------|-----------------|----------------|
+| **1** | 0.4721 | 1.387 | 5.55 | 741,598 | 52.8% |
+| **1** | 1.3088 | 2.477 | 9.81 | 11,996 | 0.0% |
+| **2** | 0.3508 | 1.275 | 2.80 | 4,117,166 | 64.9% |
+| **2** | 0.8170 | 1.762 | 6.18 | 117,615 | 18.3% |
+| **3** | 0.2622 | 1.199 | 2.09 | 11,539,749 | 73.8% |
+| **3** | 0.8527 | 1.806 | 4.88 | 727,384 | 14.7% |
+| **4** | 0.2332 🏆 | 1.175 | 1.80 | 24,106,813 | 76.7% |
+| **4** | 0.6914 🏆 | 1.615 | 3.38 | 3,549,554 | 30.9% |
+
+### Generated Text Samples
+
+Below are text samples generated from each Markov chain model:
+
+**Context Size 1:**
+
+1. `া র া মর ্ র ু ড ় ঐত ি ল ো নও ধরন ে`
+2. `্ ড০০ তথ ্ ভরশ ী শ া স ৃ হহ ী দ ি ক া`
+3. `ে ণ ী বন ্ দ ্ র ্ টগ ্ ন , দক ্ দ`
+
+**Context Size 2:**
+
+1. `া র হ া ত ি প ি ত া র া ল ী ব ি প`
+2. `ে র স ্ তর আয ় শ ্ ব া ইর া ক ি ন ব`
+3. `্ র ে ণ ী : য ু ৎ ক ে আমর া ত া ড া`
+
+**Context Size 3:**
+
+1. `্ য া ঙ ্ ক া য ় য ে শব ্ দ ে গ ু ল`
+2. `্ র ে ণ ী : ১৯৭৯ - এ প ্ রত ি রক ্ ষ ি ত`
+3. `ি য ় ে ছ ে , ত া দ ে র য া ত ি ক ব`
+
+**Context Size 4:**
+
+1. `শ ্ র ে ণ ী : ১৮০৬ - এ জন ্ ম ব ি ষয ় শ ্`
+2. `ব ি ষয ় শ ্ র ে ণ ী : ভ া রত ী য ় ধ া`
+3. `্ র ে ণ ী : ভ া রত ী য ় স ্ থ ১৯৬৬ - স ৌ`
+
+
+### Key Findings
+
+- **Best Predictability:** Context-4 with 76.7% predictability
+- **Branching Factor:** Decreases with context size (more deterministic)
+- **Memory Trade-off:** Larger contexts require more storage (3,549,554 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 | 250,463 |
+| Total Tokens | 201,406,640 |
+| Mean Frequency | 804.14 |
+| Median Frequency | 4 |
+| Frequency Std Dev | 71598.48 |
+
+### Most Common Words
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | র | 21,510,159 |
+| 2 | ন | 11,295,997 |
+| 3 | য | 10,867,520 |
+| 4 | ত | 9,714,992 |
+| 5 | ব | 9,151,878 |
+| 6 | ক | 8,792,666 |
+| 7 | স | 7,892,623 |
+| 8 | ল | 7,373,804 |
+| 9 | প | 5,927,003 |
+| 10 | ম | 5,815,632 |
+
+### Least Common Words (from vocabulary)
+
+| Rank | Word | Frequency |
+|------|------|-----------|
+| 1 | kovidara | 2 |
+| 2 | ৩৩w | 2 |
+| 3 | vuer | 2 |
+| 4 | ২০১৯২১ | 2 |
+| 5 | dilg | 2 |
+| 6 | দ১70px | 2 |
+| 7 | ciacrimes | 2 |
+| 8 | এএমএইচএস | 2 |
+| 9 | yohanna | 2 |
+| 10 | katanacho | 2 |
+
+### Zipf's Law Analysis
+
+| Metric | Value |
+|--------|-------|
+| Zipf Coefficient | 1.5776 |
+| R² (Goodness of Fit) | 0.996866 |
+| Adherence Quality | **excellent** |
+
+### Coverage Analysis
+
+| Top N Words | Coverage |
+|-------------|----------|
+| Top 100 | 81.9% |
+| Top 1,000 | 95.2% |
+| Top 5,000 | 98.2% |
+| Top 10,000 | 98.8% |
+
+### Key Findings
+
+- **Zipf Compliance:** R²=0.9969 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 81.9% of corpus
+- **Long Tail:** 240,463 words needed for remaining 1.2% 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** | 502,529 | 32 | 3.036 | 0.774 | 0.8133 🏆 |
+| **mono_64d** | 502,529 | 64 | 3.445 | 0.763 | 0.7938 |
+| **mono_128d** | 502,529 | 128 | 3.881 | 0.802 | 0.7502 |
+| **embeddings_enhanced** | 0 | 0 | 0.000 | 0.000 | 0.0000 |
+
+### Key Findings
+
+- **Best Isotropy:** mono_32d with 0.8133 (more uniform distribution)
+- **Dimension Trade-off:** Higher dimensions capture more semantics but reduce isotropy
+- **Vocabulary Coverage:** All models cover 502,529 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 (4.98x) with low UNK rate |
+| N-gram | **5-gram** | Lowest perplexity (617) |
+| Markov | **Context-4** | Highest predictability (76.7%) |
+| 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 07:23:06*

models/embeddings/monolingual/bn_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26e2e2343255830a534f320fde4263aaa50541a543b5fbc99b4a836d1708f6ed
+size 1554303761

models/embeddings/monolingual/bn_128d.meta.json

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

models/embeddings/monolingual/bn_128d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "bn",
+  "dimension": 128,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 128,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 502529
+}

models/embeddings/monolingual/bn_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:911654690f1911e79d553aca3a334439dafa8c950bae483cf0c07d3ea4e1772b
+size 400361489

models/embeddings/monolingual/bn_32d.meta.json

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

models/embeddings/monolingual/bn_32d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "bn",
+  "dimension": 32,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 32,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 502529
+}

models/embeddings/monolingual/bn_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94bb8ad1b61a560bfdb17d79f222765c5f109aca6fedef51a4f335af7d665217
+size 785008913

models/embeddings/monolingual/bn_64d.meta.json

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

models/embeddings/monolingual/bn_64d_metadata.json

@@ -0,0 +1,13 @@
+{
+  "language": "bn",
+  "dimension": 64,
+  "version": "monolingual",
+  "training_params": {
+    "dim": 64,
+    "min_count": 5,
+    "window": 5,
+    "negative": 5,
+    "epochs": 5
+  },
+  "vocab_size": 502529
+}

models/subword_markov/bn_markov_ctx1_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13a0b6a9ac2105c1af4f886f06a71d9db23ecb5a32652e818e7c2ddf640bc671
+size 772286

models/subword_markov/bn_markov_ctx1_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "subword",
+  "language": "bn",
+  "unique_contexts": 11996,
+  "total_transitions": 553695182
+}

models/subword_markov/bn_markov_ctx2_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e1c81da270a2a82d5b8c625aa355045c0a0670e6228e198544f60acda4472ec5
+size 5606771

models/subword_markov/bn_markov_ctx2_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "subword",
+  "language": "bn",
+  "unique_contexts": 117615,
+  "total_transitions": 553516718
+}

models/subword_markov/bn_markov_ctx3_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4b88518401525add742d578b9efa13a5511741c545ce0dcc6ceeb6acb7704ab
+size 26159855

models/subword_markov/bn_markov_ctx3_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "subword",
+  "language": "bn",
+  "unique_contexts": 727384,
+  "total_transitions": 553338254
+}

models/subword_markov/bn_markov_ctx4_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c54eb916a5048ffb603a48ea1d6d64f6bfe3b00818a8d1bffa427255daa9c99
+size 100901318

models/subword_markov/bn_markov_ctx4_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "subword",
+  "language": "bn",
+  "unique_contexts": 3549554,
+  "total_transitions": 553159790
+}

models/subword_ngram/bn_2gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87c44bccc6719de458cb056280276826654f9ecfa18d99048424106a8f1b8f40
+size 469388

models/subword_ngram/bn_2gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "subword",
+  "language": "bn",
+  "unique_ngrams": 33180,
+  "total_ngrams": 553695182
+}

models/subword_ngram/bn_3gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78fb7840741ea86971d4925542c9e6ce6e6c79ac8ce399abba98433705fb070b
+size 4114566

models/subword_ngram/bn_3gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "subword",
+  "language": "bn",
+  "unique_ngrams": 328172,
+  "total_ngrams": 553516718
+}

models/subword_ngram/bn_4gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46b2845e143d5e9021ee5b87cb38819dcdfb00f50601547315a5117b4da42688
+size 25302225

models/subword_ngram/bn_4gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "subword",
+  "language": "bn",
+  "unique_ngrams": 1934479,
+  "total_ngrams": 553338254
+}

models/tokenizer/bn_tokenizer_16k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68e1ac6a6a41154f9552ef7791eb5aaad0d47e5f8c5657a570b191e0f483c20a
+size 641106

models/tokenizer/bn_tokenizer_32k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39be559e8432dcd4b0837f4c1116b1daaaeb8f85ead3a786b09b36cd27681a88
+size 1074049

models/tokenizer/bn_tokenizer_64k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d1cb3f78cd4392dd3e19ff9cde695d104459cebc3f75d35d45933142a9489e9
+size 1967297

models/tokenizer/bn_tokenizer_8k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe2dd93cbc1830854f019c57ff5e0efd8b11fcda823d0bb65cfbbef3ad4eca53
+size 433093

models/vocabulary/bn_vocabulary.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8da3f061defc453cbf1bb0c1e36b511077f05f0d399b060245d1a6a437f8c4a1
+size 3901209

models/vocabulary/bn_vocabulary_metadata.json

@@ -0,0 +1,16 @@
+{
+  "language": "bn",
+  "vocabulary_size": 250463,
+  "statistics": {
+    "type_token_ratio": 0.0036646333232905415,
+    "coverage": {
+      "top_100": 0.8173536449340773,
+      "top_1000": 0.9499165937132837,
+      "top_5000": 0.9795696203113472,
+      "top_10000": 0.9855947158392181
+    },
+    "hapax_count": 489412,
+    "hapax_ratio": 0.6614793039364758,
+    "total_documents": 178464
+  }
+}

models/word_markov/bn_markov_ctx1_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e4832f62be8b99ea4eeb5d4269e7bbe06d6de2b4621946fad64f9ec22957eea
+size 38908723

models/word_markov/bn_markov_ctx1_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 1,
+  "variant": "word",
+  "language": "bn",
+  "unique_contexts": 741598,
+  "total_transitions": 376897683
+}

models/word_markov/bn_markov_ctx2_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4ba7c5295c6739716c57443fe46ae2ac3abef527a4acbe3169b854b8d2cda2f
+size 117081865

models/word_markov/bn_markov_ctx2_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 2,
+  "variant": "word",
+  "language": "bn",
+  "unique_contexts": 4117166,
+  "total_transitions": 376719219
+}

models/word_markov/bn_markov_ctx3_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a549e1e8faa30b5a0df58aae56496c8fa1e4c6ee4a74731a88aa60c91e577e74
+size 267027059

models/word_markov/bn_markov_ctx3_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 3,
+  "variant": "word",
+  "language": "bn",
+  "unique_contexts": 11539749,
+  "total_transitions": 376540755
+}

models/word_markov/bn_markov_ctx4_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89cc548abb01820903720d01a51c51890b88692832d86bcb209344e00ea779dd
+size 528654949

models/word_markov/bn_markov_ctx4_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "context_size": 4,
+  "variant": "word",
+  "language": "bn",
+  "unique_contexts": 24106813,
+  "total_transitions": 376362293
+}

models/word_ngram/bn_2gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cae303864fc970c7cc1f5114d7e1fc497c15d5ffe73ff023a5a75b1f487e2440
+size 7351010

models/word_ngram/bn_2gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 2,
+  "variant": "word",
+  "language": "bn",
+  "unique_ngrams": 514602,
+  "total_ngrams": 376897683
+}

models/word_ngram/bn_3gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a30750e0ec3ee8a1c69954761f3031b82f408d71010731f6775934668e7641e4
+size 31502716

models/word_ngram/bn_3gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 3,
+  "variant": "word",
+  "language": "bn",
+  "unique_ngrams": 2087238,
+  "total_ngrams": 376719219
+}

models/word_ngram/bn_4gram_word.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4ac57a189323e9d54a2b3aef96a23026160aa10063c1d4c12814e420a050d98
+size 107643827

models/word_ngram/bn_4gram_word_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 4,
+  "variant": "word",
+  "language": "bn",
+  "unique_ngrams": 6958889,
+  "total_ngrams": 376540755
+}