Upload all models and assets for azb (20251001)
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- README.md +293 -133
- models/embeddings/monolingual/azb_128d.bin +2 -2
- models/embeddings/monolingual/azb_128d_metadata.json +5 -3
- models/embeddings/monolingual/azb_32d.bin +2 -2
- models/embeddings/monolingual/azb_32d_metadata.json +5 -3
- models/embeddings/monolingual/azb_64d.bin +2 -2
- models/embeddings/monolingual/azb_64d_metadata.json +5 -3
- models/subword_markov/azb_markov_ctx1_subword.parquet +2 -2
- models/subword_markov/azb_markov_ctx1_subword_metadata.json +2 -2
- models/subword_markov/azb_markov_ctx2_subword.parquet +2 -2
- models/subword_markov/azb_markov_ctx2_subword_metadata.json +2 -2
- models/subword_markov/azb_markov_ctx3_subword.parquet +2 -2
- models/subword_markov/azb_markov_ctx3_subword_metadata.json +2 -2
- models/subword_markov/azb_markov_ctx4_subword.parquet +2 -2
- models/subword_markov/azb_markov_ctx4_subword_metadata.json +2 -2
- models/subword_ngram/azb_2gram_subword.parquet +2 -2
- models/subword_ngram/azb_2gram_subword_metadata.json +2 -2
- models/subword_ngram/azb_3gram_subword.parquet +2 -2
- models/subword_ngram/azb_3gram_subword_metadata.json +2 -2
- models/subword_ngram/azb_4gram_subword.parquet +2 -2
- models/subword_ngram/azb_4gram_subword_metadata.json +2 -2
- models/tokenizer/azb_tokenizer_16k.model +2 -2
- models/tokenizer/azb_tokenizer_16k.vocab +0 -0
- models/tokenizer/azb_tokenizer_32k.model +2 -2
- models/tokenizer/azb_tokenizer_32k.vocab +0 -0
- models/tokenizer/azb_tokenizer_64k.model +2 -2
- models/tokenizer/azb_tokenizer_64k.vocab +0 -0
- models/tokenizer/azb_tokenizer_8k.model +2 -2
- models/tokenizer/azb_tokenizer_8k.vocab +0 -0
- models/vocabulary/azb_vocabulary.parquet +2 -2
- models/vocabulary/azb_vocabulary_metadata.json +10 -9
- models/word_markov/azb_markov_ctx1_word.parquet +2 -2
- models/word_markov/azb_markov_ctx1_word_metadata.json +2 -2
- models/word_markov/azb_markov_ctx2_word.parquet +2 -2
- models/word_markov/azb_markov_ctx2_word_metadata.json +2 -2
- models/word_markov/azb_markov_ctx3_word.parquet +2 -2
- models/word_markov/azb_markov_ctx3_word_metadata.json +2 -2
- models/word_markov/azb_markov_ctx4_word.parquet +2 -2
- models/word_markov/azb_markov_ctx4_word_metadata.json +2 -2
- models/word_ngram/azb_2gram_word.parquet +2 -2
- models/word_ngram/azb_2gram_word_metadata.json +2 -2
- models/word_ngram/azb_3gram_word.parquet +2 -2
- models/word_ngram/azb_3gram_word_metadata.json +2 -2
- models/word_ngram/azb_4gram_word.parquet +2 -2
- models/word_ngram/azb_4gram_word_metadata.json +2 -2
- visualizations/embedding_isotropy.png +0 -0
- visualizations/embedding_norms.png +0 -0
- visualizations/embedding_similarity.png +2 -2
- visualizations/markov_branching.png +0 -0
- visualizations/markov_contexts.png +0 -0
README.md
CHANGED
|
@@ -23,14 +23,14 @@ dataset_info:
|
|
| 23 |
metrics:
|
| 24 |
- name: best_compression_ratio
|
| 25 |
type: compression
|
| 26 |
-
value: 4.
|
| 27 |
- name: best_isotropy
|
| 28 |
type: isotropy
|
| 29 |
-
value: 0.
|
| 30 |
- name: vocabulary_size
|
| 31 |
type: vocab
|
| 32 |
-
value:
|
| 33 |
-
generated:
|
| 34 |
---
|
| 35 |
|
| 36 |
# AZB - Wikilangs Models
|
|
@@ -44,12 +44,13 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
|
|
| 44 |
### Models & Assets
|
| 45 |
|
| 46 |
- Tokenizers (8k, 16k, 32k, 64k)
|
| 47 |
-
- N-gram models (2, 3, 4-gram)
|
| 48 |
-
- Markov chains (context of 1, 2, 3 and
|
| 49 |
- Subword N-gram and Markov chains
|
| 50 |
-
- Embeddings in various sizes and dimensions
|
| 51 |
- Language Vocabulary
|
| 52 |
- Language Statistics
|
|
|
|
| 53 |

|
| 54 |
|
| 55 |
### Analysis and Evaluation
|
|
@@ -59,7 +60,8 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
|
|
| 59 |
- [3. Markov Chain Evaluation](#3-markov-chain-evaluation)
|
| 60 |
- [4. Vocabulary Analysis](#4-vocabulary-analysis)
|
| 61 |
- [5. Word Embeddings Evaluation](#5-word-embeddings-evaluation)
|
| 62 |
-
- [6.
|
|
|
|
| 63 |
- [Metrics Glossary](#appendix-metrics-glossary--interpretation-guide)
|
| 64 |
- [Visualizations Index](#visualizations-index)
|
| 65 |
|
|
@@ -68,51 +70,57 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
|
|
| 68 |
|
| 69 |

|
| 70 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 71 |
### Results
|
| 72 |
|
| 73 |
| Vocab Size | Compression | Avg Token Len | UNK Rate | Total Tokens |
|
| 74 |
|------------|-------------|---------------|----------|--------------|
|
| 75 |
-
| **8k** | 3.
|
| 76 |
-
| **16k** | 3.
|
| 77 |
-
| **32k** | 3.
|
| 78 |
-
| **64k** | 4.
|
| 79 |
|
| 80 |
### Tokenization Examples
|
| 81 |
|
| 82 |
Below are sample sentences tokenized with each vocabulary size:
|
| 83 |
|
| 84 |
-
**Sample 1:**
|
| 85 |
|
| 86 |
| Vocab | Tokens | Count |
|
| 87 |
|-------|--------|-------|
|
| 88 |
-
| 8k |
|
| 89 |
-
| 16k |
|
| 90 |
-
| 32k |
|
| 91 |
-
| 64k |
|
| 92 |
|
| 93 |
-
**Sample 2:**
|
| 94 |
|
| 95 |
| Vocab | Tokens | Count |
|
| 96 |
|-------|--------|-------|
|
| 97 |
-
| 8k |
|
| 98 |
-
| 16k |
|
| 99 |
-
| 32k |
|
| 100 |
-
| 64k |
|
| 101 |
|
| 102 |
-
**Sample 3:**
|
| 103 |
|
| 104 |
| Vocab | Tokens | Count |
|
| 105 |
|-------|--------|-------|
|
| 106 |
-
| 8k |
|
| 107 |
-
| 16k |
|
| 108 |
-
| 32k |
|
| 109 |
-
| 64k |
|
| 110 |
|
| 111 |
|
| 112 |
### Key Findings
|
| 113 |
|
| 114 |
-
- **Best Compression:** 64k achieves 4.
|
| 115 |
-
- **Lowest UNK Rate:** 8k with 0.
|
| 116 |
- **Trade-off:** Larger vocabularies improve compression but increase model size
|
| 117 |
- **Recommendation:** 32k vocabulary provides optimal balance for production use
|
| 118 |
|
|
@@ -121,57 +129,89 @@ Below are sample sentences tokenized with each vocabulary size:
|
|
| 121 |
|
| 122 |

|
| 123 |
|
|
|
|
|
|
|
| 124 |

|
| 125 |
|
| 126 |
### Results
|
| 127 |
|
| 128 |
-
| N-gram | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
|
| 129 |
-
|
| 130 |
-
| **2-gram** |
|
| 131 |
-
| **2-gram** |
|
| 132 |
-
| **3-gram** |
|
| 133 |
-
| **3-gram** |
|
| 134 |
-
| **4-gram** |
|
| 135 |
-
| **4-gram** |
|
| 136 |
|
| 137 |
### Top 5 N-grams by Size
|
| 138 |
|
| 139 |
-
**2-grams:**
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 140 |
|
| 141 |
| Rank | N-gram | Count |
|
| 142 |
|------|--------|-------|
|
| 143 |
-
| 1 |
|
| 144 |
-
| 2 |
|
| 145 |
-
| 3 |
|
| 146 |
-
| 4 |
|
| 147 |
-
| 5 |
|
| 148 |
|
| 149 |
-
**3-grams:**
|
| 150 |
|
| 151 |
| Rank | N-gram | Count |
|
| 152 |
|------|--------|-------|
|
| 153 |
-
| 1 |
|
| 154 |
-
| 2 |
|
| 155 |
-
| 3 |
|
| 156 |
-
| 4 |
|
| 157 |
-
| 5 |
|
| 158 |
|
| 159 |
-
**4-grams:**
|
| 160 |
|
| 161 |
| Rank | N-gram | Count |
|
| 162 |
|------|--------|-------|
|
| 163 |
-
| 1 |
|
| 164 |
-
| 2 |
|
| 165 |
-
| 3 |
|
| 166 |
-
| 4 |
|
| 167 |
-
| 5 |
|
| 168 |
|
| 169 |
|
| 170 |
### Key Findings
|
| 171 |
|
| 172 |
-
- **Best Perplexity:** 2-gram with
|
| 173 |
- **Entropy Trend:** Decreases with larger n-grams (more predictable)
|
| 174 |
-
- **Coverage:** Top-1000 patterns cover ~
|
| 175 |
- **Recommendation:** 4-gram or 5-gram for best predictive performance
|
| 176 |
|
| 177 |
---
|
|
@@ -179,55 +219,86 @@ Below are sample sentences tokenized with each vocabulary size:
|
|
| 179 |
|
| 180 |

|
| 181 |
|
|
|
|
|
|
|
| 182 |

|
| 183 |
|
| 184 |
### Results
|
| 185 |
|
| 186 |
-
| Context | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
|
| 187 |
-
|
| 188 |
-
| **1** | 0.
|
| 189 |
-
| **1** | 1.
|
| 190 |
-
| **2** | 0.
|
| 191 |
-
| **2** |
|
| 192 |
-
| **3** | 0.
|
| 193 |
-
| **3** | 0.
|
| 194 |
-
| **4** | 0.
|
| 195 |
-
| **4** | 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 196 |
|
| 197 |
-
### Generated Text Samples
|
| 198 |
|
| 199 |
-
|
|
|
|
|
|
|
| 200 |
|
| 201 |
**Context Size 1:**
|
| 202 |
|
| 203 |
-
1.
|
| 204 |
-
2.
|
| 205 |
-
3.
|
| 206 |
|
| 207 |
**Context Size 2:**
|
| 208 |
|
| 209 |
-
1.
|
| 210 |
-
2.
|
| 211 |
-
3.
|
| 212 |
|
| 213 |
**Context Size 3:**
|
| 214 |
|
| 215 |
-
1.
|
| 216 |
-
2.
|
| 217 |
-
3.
|
| 218 |
|
| 219 |
**Context Size 4:**
|
| 220 |
|
| 221 |
-
1.
|
| 222 |
-
2.
|
| 223 |
-
3.
|
| 224 |
|
| 225 |
|
| 226 |
### Key Findings
|
| 227 |
|
| 228 |
-
- **Best Predictability:** Context-4 with
|
| 229 |
- **Branching Factor:** Decreases with context size (more deterministic)
|
| 230 |
-
- **Memory Trade-off:** Larger contexts require more storage (
|
| 231 |
- **Recommendation:** Context-3 or Context-4 for text generation
|
| 232 |
|
| 233 |
---
|
|
@@ -243,64 +314,64 @@ Below are text samples generated from each Markov chain model:
|
|
| 243 |
|
| 244 |
| Metric | Value |
|
| 245 |
|--------|-------|
|
| 246 |
-
| Vocabulary Size |
|
| 247 |
-
| Total Tokens |
|
| 248 |
-
| Mean Frequency |
|
| 249 |
| Median Frequency | 3 |
|
| 250 |
-
| Frequency Std Dev |
|
| 251 |
|
| 252 |
### Most Common Words
|
| 253 |
|
| 254 |
| Rank | Word | Frequency |
|
| 255 |
|------|------|-----------|
|
| 256 |
-
| 1 |
|
| 257 |
-
| 2 |
|
| 258 |
-
| 3 | اینگیلیسجه |
|
| 259 |
-
| 4 |
|
| 260 |
-
| 5 |
|
| 261 |
-
| 6 |
|
| 262 |
-
| 7 |
|
| 263 |
-
| 8 |
|
| 264 |
-
| 9 |
|
| 265 |
-
| 10 |
|
| 266 |
|
| 267 |
### Least Common Words (from vocabulary)
|
| 268 |
|
| 269 |
| Rank | Word | Frequency |
|
| 270 |
|------|------|-----------|
|
| 271 |
-
| 1 |
|
| 272 |
-
| 2 |
|
| 273 |
-
| 3 |
|
| 274 |
-
| 4 |
|
| 275 |
-
| 5 |
|
| 276 |
-
| 6 |
|
| 277 |
-
| 7 |
|
| 278 |
-
| 8 |
|
| 279 |
-
| 9 |
|
| 280 |
-
| 10 |
|
| 281 |
|
| 282 |
### Zipf's Law Analysis
|
| 283 |
|
| 284 |
| Metric | Value |
|
| 285 |
|--------|-------|
|
| 286 |
-
| Zipf Coefficient | 1.
|
| 287 |
-
| R² (Goodness of Fit) | 0.
|
| 288 |
| Adherence Quality | **excellent** |
|
| 289 |
|
| 290 |
### Coverage Analysis
|
| 291 |
|
| 292 |
| Top N Words | Coverage |
|
| 293 |
|-------------|----------|
|
| 294 |
-
| Top 100 |
|
| 295 |
-
| Top 1,000 |
|
| 296 |
-
| Top 5,000 |
|
| 297 |
-
| Top 10,000 |
|
| 298 |
|
| 299 |
### Key Findings
|
| 300 |
|
| 301 |
-
- **Zipf Compliance:** R²=0.
|
| 302 |
-
- **High Frequency Dominance:** Top 100 words cover
|
| 303 |
-
- **Long Tail:**
|
| 304 |
|
| 305 |
---
|
| 306 |
## 5. Word Embeddings Evaluation
|
|
@@ -313,24 +384,110 @@ Below are text samples generated from each Markov chain model:
|
|
| 313 |
|
| 314 |

|
| 315 |
|
| 316 |
-
### Model Comparison
|
| 317 |
|
| 318 |
-
|
| 319 |
-
|
| 320 |
-
|
| 321 |
-
|
| 322 |
-
|
| 323 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 324 |
|
| 325 |
### Key Findings
|
| 326 |
|
| 327 |
-
- **Best Isotropy:** mono_32d with 0.
|
| 328 |
-
- **
|
| 329 |
-
- **
|
| 330 |
-
- **Recommendation:**
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 331 |
|
| 332 |
---
|
| 333 |
-
##
|
| 334 |
|
| 335 |

|
| 336 |
|
|
@@ -338,11 +495,12 @@ Below are text samples generated from each Markov chain model:
|
|
| 338 |
|
| 339 |
| Component | Recommended | Rationale |
|
| 340 |
|-----------|-------------|-----------|
|
| 341 |
-
| Tokenizer | **
|
| 342 |
-
| N-gram | **
|
| 343 |
-
| Markov | **Context-4** | Highest predictability (
|
| 344 |
| Embeddings | **100d** | Balanced semantic capture and isotropy |
|
| 345 |
|
|
|
|
| 346 |
---
|
| 347 |
## Appendix: Metrics Glossary & Interpretation Guide
|
| 348 |
|
|
@@ -532,7 +690,8 @@ If you use these models in your research, please cite:
|
|
| 532 |
author = {Kamali, Omar},
|
| 533 |
title = {Wikilangs: Open NLP Models for Wikipedia Languages},
|
| 534 |
year = {2025},
|
| 535 |
-
|
|
|
|
| 536 |
url = {https://huggingface.co/wikilangs}
|
| 537 |
institution = {Omneity Labs}
|
| 538 |
}
|
|
@@ -548,7 +707,8 @@ MIT License - Free for academic and commercial use.
|
|
| 548 |
- 🤗 Models: [huggingface.co/wikilangs](https://huggingface.co/wikilangs)
|
| 549 |
- 📊 Data: [wikipedia-monthly](https://huggingface.co/datasets/omarkamali/wikipedia-monthly)
|
| 550 |
- 👤 Author: [Omar Kamali](https://huggingface.co/omarkamali)
|
|
|
|
| 551 |
---
|
| 552 |
*Generated by Wikilangs Models Pipeline*
|
| 553 |
|
| 554 |
-
*Report Date:
|
|
|
|
| 23 |
metrics:
|
| 24 |
- name: best_compression_ratio
|
| 25 |
type: compression
|
| 26 |
+
value: 4.148
|
| 27 |
- name: best_isotropy
|
| 28 |
type: isotropy
|
| 29 |
+
value: 0.8282
|
| 30 |
- name: vocabulary_size
|
| 31 |
type: vocab
|
| 32 |
+
value: 0
|
| 33 |
+
generated: 2026-01-03
|
| 34 |
---
|
| 35 |
|
| 36 |
# AZB - Wikilangs Models
|
|
|
|
| 44 |
### Models & Assets
|
| 45 |
|
| 46 |
- Tokenizers (8k, 16k, 32k, 64k)
|
| 47 |
+
- N-gram models (2, 3, 4, 5-gram)
|
| 48 |
+
- Markov chains (context of 1, 2, 3, 4 and 5)
|
| 49 |
- Subword N-gram and Markov chains
|
| 50 |
+
- Embeddings in various sizes and dimensions (aligned and unaligned)
|
| 51 |
- Language Vocabulary
|
| 52 |
- Language Statistics
|
| 53 |
+
|
| 54 |

|
| 55 |
|
| 56 |
### Analysis and Evaluation
|
|
|
|
| 60 |
- [3. Markov Chain Evaluation](#3-markov-chain-evaluation)
|
| 61 |
- [4. Vocabulary Analysis](#4-vocabulary-analysis)
|
| 62 |
- [5. Word Embeddings Evaluation](#5-word-embeddings-evaluation)
|
| 63 |
+
- [6. Morphological Analysis (Experimental)](#6-morphological-analysis)
|
| 64 |
+
- [7. Summary & Recommendations](#7-summary--recommendations)
|
| 65 |
- [Metrics Glossary](#appendix-metrics-glossary--interpretation-guide)
|
| 66 |
- [Visualizations Index](#visualizations-index)
|
| 67 |
|
|
|
|
| 70 |
|
| 71 |

|
| 72 |
|
| 73 |
+

|
| 74 |
+
|
| 75 |
+

|
| 76 |
+
|
| 77 |
+

|
| 78 |
+
|
| 79 |
### Results
|
| 80 |
|
| 81 |
| Vocab Size | Compression | Avg Token Len | UNK Rate | Total Tokens |
|
| 82 |
|------------|-------------|---------------|----------|--------------|
|
| 83 |
+
| **8k** | 3.135x | 3.14 | 0.5011% | 364,218 |
|
| 84 |
+
| **16k** | 3.510x | 3.51 | 0.5610% | 325,334 |
|
| 85 |
+
| **32k** | 3.852x | 3.86 | 0.6157% | 296,427 |
|
| 86 |
+
| **64k** | 4.148x 🏆 | 4.15 | 0.6629% | 275,291 |
|
| 87 |
|
| 88 |
### Tokenization Examples
|
| 89 |
|
| 90 |
Below are sample sentences tokenized with each vocabulary size:
|
| 91 |
|
| 92 |
+
**Sample 1:** `قالیکتیس (، ، ، ) ییٛرتیجیلار دستهسینه عایید حئیوان نؤعو. قایناقلار سیراسینا گؤ...`
|
| 93 |
|
| 94 |
| Vocab | Tokens | Count |
|
| 95 |
|-------|--------|-------|
|
| 96 |
+
| 8k | `▁قا لیک تیس ▁(، ▁، ▁، ▁) ▁ییٛرتیجیلار ▁دسته ▁سینه ... (+8 more)` | 18 |
|
| 97 |
+
| 16k | `▁قا لیک تیس ▁(، ▁، ▁، ▁) ▁ییٛرتیجیلار ▁دسته ▁سینه ... (+8 more)` | 18 |
|
| 98 |
+
| 32k | `▁قا لیک تیس ▁(، ▁، ▁، ▁) ▁ییٛرتیجیلار ▁دسته ▁سینه ... (+8 more)` | 18 |
|
| 99 |
+
| 64k | `▁قا لیک تیس ▁(، ▁، ▁، ▁) ▁ییٛرتیجیلار ▁دسته ▁سینه ... (+8 more)` | 18 |
|
| 100 |
|
| 101 |
+
**Sample 2:** `هیندوستان اؤلکهسینین کرالا ایالتینده بیر شهر دیر. بۇ شهرده مالایالم دیلی و اینگ...`
|
| 102 |
|
| 103 |
| Vocab | Tokens | Count |
|
| 104 |
|-------|--------|-------|
|
| 105 |
+
| 8k | `▁هیندوستان ▁اؤلکه ▁سینین ▁کرالا ▁ایالتینده ▁بیر ▁شهر ▁دیر . ▁بۇ ... (+10 more)` | 20 |
|
| 106 |
+
| 16k | `▁هیندوستان ▁اؤلکه ▁سینین ▁کرالا ▁ایالتینده ▁بیر ▁شهر ▁دیر . ▁بۇ ... (+10 more)` | 20 |
|
| 107 |
+
| 32k | `▁هیندوستان ▁اؤلکه ▁سینین ▁کرالا ▁ایالتینده ▁بیر ▁شهر ▁دیر . ▁بۇ ... (+10 more)` | 20 |
|
| 108 |
+
| 64k | `▁هیندوستان ▁اؤلکه ▁سینین ▁کرالا ▁ایالتینده ▁بیر ▁شهر ▁دیر . ▁بۇ ... (+10 more)` | 20 |
|
| 109 |
|
| 110 |
+
**Sample 3:** `آرقا, کارناتاکا Karnataka) هیندوستان اؤلکهسینین کارناتاکا ایالتینده بیر کند دیر...`
|
| 111 |
|
| 112 |
| Vocab | Tokens | Count |
|
| 113 |
|-------|--------|-------|
|
| 114 |
+
| 8k | `▁آر قا , ▁کارناتاکا ▁kar n at aka ) ▁هیندوستان ... (+16 more)` | 26 |
|
| 115 |
+
| 16k | `▁آر قا , ▁کارناتاکا ▁kar nat aka ) ▁هیندوستان ▁اؤلکه ... (+15 more)` | 25 |
|
| 116 |
+
| 32k | `▁آر قا , ▁کارناتاکا ▁karnataka ) ▁هیندوستان ▁اؤلکه ▁سینین ▁کارناتاکا ... (+13 more)` | 23 |
|
| 117 |
+
| 64k | `▁آر قا , ▁کارناتاکا ▁karnataka ) ▁هیندوستان ▁اؤلکه ▁سینین ▁کارناتاکا ... (+13 more)` | 23 |
|
| 118 |
|
| 119 |
|
| 120 |
### Key Findings
|
| 121 |
|
| 122 |
+
- **Best Compression:** 64k achieves 4.148x compression
|
| 123 |
+
- **Lowest UNK Rate:** 8k with 0.5011% unknown tokens
|
| 124 |
- **Trade-off:** Larger vocabularies improve compression but increase model size
|
| 125 |
- **Recommendation:** 32k vocabulary provides optimal balance for production use
|
| 126 |
|
|
|
|
| 129 |
|
| 130 |

|
| 131 |
|
| 132 |
+

|
| 133 |
+
|
| 134 |

|
| 135 |
|
| 136 |
### Results
|
| 137 |
|
| 138 |
+
| N-gram | Variant | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
|
| 139 |
+
|--------|---------|------------|---------|----------------|------------------|-------------------|
|
| 140 |
+
| **2-gram** | Word | 8,048 | 12.97 | 158,908 | 25.7% | 56.1% |
|
| 141 |
+
| **2-gram** | Subword | 528 🏆 | 9.04 | 12,648 | 51.6% | 95.7% |
|
| 142 |
+
| **3-gram** | Word | 10,249 | 13.32 | 236,749 | 22.6% | 53.6% |
|
| 143 |
+
| **3-gram** | Subword | 3,765 | 11.88 | 106,644 | 23.1% | 62.4% |
|
| 144 |
+
| **4-gram** | Word | 17,175 | 14.07 | 426,395 | 19.0% | 47.9% |
|
| 145 |
+
| **4-gram** | Subword | 15,100 | 13.88 | 581,225 | 14.6% | 44.8% |
|
| 146 |
|
| 147 |
### Top 5 N-grams by Size
|
| 148 |
|
| 149 |
+
**2-grams (Word):**
|
| 150 |
+
|
| 151 |
+
| Rank | N-gram | Count |
|
| 152 |
+
|------|--------|-------|
|
| 153 |
+
| 1 | `ایشلدنلری طرفیندن` | 75,584 |
|
| 154 |
+
| 2 | `مقالهسیندن گؤتورولوبدور` | 75,503 |
|
| 155 |
+
| 3 | `ویکیپدیاسینین ایشلدنلری` | 73,734 |
|
| 156 |
+
| 4 | `اینگیلیسجه ویکیپدیاسینین` | 71,132 |
|
| 157 |
+
| 5 | `قایناقلار اینگیلیسجه` | 70,880 |
|
| 158 |
+
|
| 159 |
+
**3-grams (Word):**
|
| 160 |
+
|
| 161 |
+
| Rank | N-gram | Count |
|
| 162 |
+
|------|--------|-------|
|
| 163 |
+
| 1 | `ویکیپدیاسینین ایشلدنلری طرفیندن` | 73,734 |
|
| 164 |
+
| 2 | `اینگیلیسجه ویکیپدیاسینین ایشلدنلری` | 71,132 |
|
| 165 |
+
| 3 | `قایناقلار اینگیلیسجه ویکیپدیاسینین` | 70,806 |
|
| 166 |
+
| 4 | `بیر یاشاییش منطقهسیدیر` | 40,399 |
|
| 167 |
+
| 5 | `بیر کند دیر` | 30,448 |
|
| 168 |
+
|
| 169 |
+
**4-grams (Word):**
|
| 170 |
+
|
| 171 |
+
| Rank | N-gram | Count |
|
| 172 |
+
|------|--------|-------|
|
| 173 |
+
| 1 | `اینگیلیسجه ویکیپدیاسینین ایشلدنلری طرفیندن` | 71,132 |
|
| 174 |
+
| 2 | `قایناقلار اینگیلیسجه ویکیپدیاسینین ایشلدنلری` | 70,806 |
|
| 175 |
+
| 3 | `سوْن نۆفوس ساییمی اساسيندا` | 24,568 |
|
| 176 |
+
| 4 | `شهرلرین لیستی قایناقلار اینگیلیسجه` | 22,937 |
|
| 177 |
+
| 5 | `لیستی قایناقلار اینگیلیسجه ویکیپدیاسینین` | 22,937 |
|
| 178 |
+
|
| 179 |
+
**2-grams (Subword):**
|
| 180 |
|
| 181 |
| Rank | N-gram | Count |
|
| 182 |
|------|--------|-------|
|
| 183 |
+
| 1 | `ی ن` | 1,868,991 |
|
| 184 |
+
| 2 | `_ ا` | 1,658,104 |
|
| 185 |
+
| 3 | `ی _` | 1,437,263 |
|
| 186 |
+
| 4 | `ا ی` | 1,393,221 |
|
| 187 |
+
| 5 | `ن _` | 1,215,806 |
|
| 188 |
|
| 189 |
+
**3-grams (Subword):**
|
| 190 |
|
| 191 |
| Rank | N-gram | Count |
|
| 192 |
|------|--------|-------|
|
| 193 |
+
| 1 | `_ ا ی` | 717,380 |
|
| 194 |
+
| 2 | `ی ن د` | 658,977 |
|
| 195 |
+
| 3 | `د ه _` | 585,522 |
|
| 196 |
+
| 4 | `ل ا ر` | 580,226 |
|
| 197 |
+
| 5 | `ا ی ن` | 470,621 |
|
| 198 |
|
| 199 |
+
**4-grams (Subword):**
|
| 200 |
|
| 201 |
| Rank | N-gram | Count |
|
| 202 |
|------|--------|-------|
|
| 203 |
+
| 1 | `ن د ه _` | 347,347 |
|
| 204 |
+
| 2 | `ل ا ر _` | 329,379 |
|
| 205 |
+
| 3 | `ی ن د ه` | 320,994 |
|
| 206 |
+
| 4 | `_ ب ی ر` | 258,707 |
|
| 207 |
+
| 5 | `ن ی ن _` | 257,628 |
|
| 208 |
|
| 209 |
|
| 210 |
### Key Findings
|
| 211 |
|
| 212 |
+
- **Best Perplexity:** 2-gram (subword) with 528
|
| 213 |
- **Entropy Trend:** Decreases with larger n-grams (more predictable)
|
| 214 |
+
- **Coverage:** Top-1000 patterns cover ~45% of corpus
|
| 215 |
- **Recommendation:** 4-gram or 5-gram for best predictive performance
|
| 216 |
|
| 217 |
---
|
|
|
|
| 219 |
|
| 220 |

|
| 221 |
|
| 222 |
+

|
| 223 |
+
|
| 224 |

|
| 225 |
|
| 226 |
### Results
|
| 227 |
|
| 228 |
+
| Context | Variant | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
|
| 229 |
+
|---------|---------|-------------|------------|------------------|-----------------|----------------|
|
| 230 |
+
| **1** | Word | 0.6640 | 1.584 | 5.09 | 726,930 | 33.6% |
|
| 231 |
+
| **1** | Subword | 1.0592 | 2.084 | 9.07 | 3,409 | 0.0% |
|
| 232 |
+
| **2** | Word | 0.1970 | 1.146 | 1.48 | 3,693,091 | 80.3% |
|
| 233 |
+
| **2** | Subword | 0.9308 | 1.906 | 6.57 | 30,905 | 6.9% |
|
| 234 |
+
| **3** | Word | 0.0689 | 1.049 | 1.14 | 5,447,170 | 93.1% |
|
| 235 |
+
| **3** | Subword | 0.8408 | 1.791 | 4.70 | 203,056 | 15.9% |
|
| 236 |
+
| **4** | Word | 0.0340 🏆 | 1.024 | 1.07 | 6,178,524 | 96.6% |
|
| 237 |
+
| **4** | Subword | 0.7000 | 1.625 | 3.22 | 953,883 | 30.0% |
|
| 238 |
+
|
| 239 |
+
### Generated Text Samples (Word-based)
|
| 240 |
+
|
| 241 |
+
Below are text samples generated from each word-based Markov chain model:
|
| 242 |
+
|
| 243 |
+
**Context Size 1:**
|
| 244 |
+
|
| 245 |
+
1. `و شمشددیل سلطانلیغی ایله برابر ایدی قایناقلار شهرلری آمریکا بیرلشمیش ایالتلرینین ایداری بؤلوملری ائش...`
|
| 246 |
+
2. `بیر کند دیر و یا هچ اینگیلیسجه ویکیپدیاسینین ایشلدنلری طرفیندن mountain a few years until 5`
|
| 247 |
+
3. `اینگیلیسجه ویکیپدیاسینین ایشلدنلری طرفیندن mountain wilbert minnesota مقالهسیندن گؤتورولوبدور ۸ آ...`
|
| 248 |
+
|
| 249 |
+
**Context Size 2:**
|
| 250 |
+
|
| 251 |
+
1. `ایشلدنلری طرفیندن مقالهسیندن گؤتورولوبدور ۸ آقوست تاریخینده یوْخلانیلیبدیر شهرلری en güləh`
|
| 252 |
+
2. `مقالهسیندن گؤتورولوبدور ۳۰ نوْوامبر تاریخینده یوْخلانیلیبدیر کولوبلاری en araks ararat fc مقالهسین...`
|
| 253 |
+
3. `ویکیپدیاسینین ایشلدنلری طرفیندن indiana مقالهسیندن گؤتورولوبدور ۲۱ دسامبر تاریخینده یوْخلانیلیبدی...`
|
| 254 |
+
|
| 255 |
+
**Context Size 3:**
|
| 256 |
+
|
| 257 |
+
1. `ویکیپدیاسینین ایشلدنلری طرفیندن مقالهسیندن گؤتورولوبدور ۳۰ نوْوامبر تاریخینده یوْخلانیلیبدیر گؤرو...`
|
| 258 |
+
2. `اینگیلیسجه ویکیپدیاسینین ایشلدنلری طرفیندن rutherfurd مقالهسیندن گؤتورولوبدور ۲۲ ژانویه تاریخینده...`
|
| 259 |
+
3. `قایناقلار اینگیلیسجه ویکیپدیاسینین ایشلدنلری طرفیندن مقالهسیندن گؤتورولوبدور ۱۹ جولای یوْخلانیلی...`
|
| 260 |
+
|
| 261 |
+
**Context Size 4:**
|
| 262 |
+
|
| 263 |
+
1. `اینگیلیسجه ویکیپدیاسینین ایشلدنلری طرفیندن مقالهسیندن گؤتورولوبدور ۸ آقوست تاریخینده یوْخلانیلیبد...`
|
| 264 |
+
2. `قایناقلار اینگیلیسجه ویکیپدیاسینین ایشلدنلری طرفیندن georgia مقالهسیندن گؤتورولوبدور ۸ آقوست تار...`
|
| 265 |
+
3. `سوْن نۆفوس ساییمی اساسيندا ۱۳۷ نفر ایمیش و ویسوچینا اوستانیندا یئرلشیب بۆتون چکدا اوْلدوغو کیمی بۇ ...`
|
| 266 |
|
|
|
|
| 267 |
|
| 268 |
+
### Generated Text Samples (Subword-based)
|
| 269 |
+
|
| 270 |
+
Below are text samples generated from each subword-based Markov chain model:
|
| 271 |
|
| 272 |
**Context Size 1:**
|
| 273 |
|
| 274 |
+
1. `_متیشاورلهستبشلی`
|
| 275 |
+
2. `یتا_d_by_jottiep`
|
| 276 |
+
3. `الاقلر)_ژادالين_`
|
| 277 |
|
| 278 |
**Context Size 2:**
|
| 279 |
|
| 280 |
+
1. `ین_حیازیل_دیکان_ا`
|
| 281 |
+
2. `_ایلدا_۹_جی_giran`
|
| 282 |
+
3. `ی_حافع_سان_آماری_`
|
| 283 |
|
| 284 |
**Context Size 3:**
|
| 285 |
|
| 286 |
+
1. `_ایدان_عوضوو._۲۷۳_`
|
| 287 |
+
2. `ینده_روس_سال_منطقه`
|
| 288 |
+
3. `ده_یوْخلانی_آما_خوب`
|
| 289 |
|
| 290 |
**Context Size 4:**
|
| 291 |
|
| 292 |
+
1. `نده_یوْخلانیلیبدیر).`
|
| 293 |
+
2. `لار_یولو_۳۰_دسامبر_`
|
| 294 |
+
3. `ینده_هر_گونئی_کاروا`
|
| 295 |
|
| 296 |
|
| 297 |
### Key Findings
|
| 298 |
|
| 299 |
+
- **Best Predictability:** Context-4 (word) with 96.6% predictability
|
| 300 |
- **Branching Factor:** Decreases with context size (more deterministic)
|
| 301 |
+
- **Memory Trade-off:** Larger contexts require more storage (953,883 contexts)
|
| 302 |
- **Recommendation:** Context-3 or Context-4 for text generation
|
| 303 |
|
| 304 |
---
|
|
|
|
| 314 |
|
| 315 |
| Metric | Value |
|
| 316 |
|--------|-------|
|
| 317 |
+
| Vocabulary Size | 271,198 |
|
| 318 |
+
| Total Tokens | 12,478,531 |
|
| 319 |
+
| Mean Frequency | 46.01 |
|
| 320 |
| Median Frequency | 3 |
|
| 321 |
+
| Frequency Std Dev | 1145.11 |
|
| 322 |
|
| 323 |
### Most Common Words
|
| 324 |
|
| 325 |
| Rank | Word | Frequency |
|
| 326 |
|------|------|-----------|
|
| 327 |
+
| 1 | و | 284,031 |
|
| 328 |
+
| 2 | بیر | 169,280 |
|
| 329 |
+
| 3 | اینگیلیسجه | 149,737 |
|
| 330 |
+
| 4 | قایناقلار | 141,945 |
|
| 331 |
+
| 5 | the | 114,439 |
|
| 332 |
+
| 6 | تاریخینده | 92,079 |
|
| 333 |
+
| 7 | قایناقلار | 90,963 |
|
| 334 |
+
| 8 | ایلده | 83,679 |
|
| 335 |
+
| 9 | شهرلری | 81,894 |
|
| 336 |
+
| 10 | طرفیندن | 80,132 |
|
| 337 |
|
| 338 |
### Least Common Words (from vocabulary)
|
| 339 |
|
| 340 |
| Rank | Word | Frequency |
|
| 341 |
|------|------|-----------|
|
| 342 |
+
| 1 | کساسیاسی | 2 |
|
| 343 |
+
| 2 | کالابری�� | 2 |
|
| 344 |
+
| 3 | کونتینوا | 2 |
|
| 345 |
+
| 4 | تحقیقلری | 2 |
|
| 346 |
+
| 5 | romanzo | 2 |
|
| 347 |
+
| 6 | strage | 2 |
|
| 348 |
+
| 7 | سۆرهجینده | 2 |
|
| 349 |
+
| 8 | ایلکهلر | 2 |
|
| 350 |
+
| 9 | لائیکلیک | 2 |
|
| 351 |
+
| 10 | شاسکوه | 2 |
|
| 352 |
|
| 353 |
### Zipf's Law Analysis
|
| 354 |
|
| 355 |
| Metric | Value |
|
| 356 |
|--------|-------|
|
| 357 |
+
| Zipf Coefficient | 1.1609 |
|
| 358 |
+
| R² (Goodness of Fit) | 0.995521 |
|
| 359 |
| Adherence Quality | **excellent** |
|
| 360 |
|
| 361 |
### Coverage Analysis
|
| 362 |
|
| 363 |
| Top N Words | Coverage |
|
| 364 |
|-------------|----------|
|
| 365 |
+
| Top 100 | 34.4% |
|
| 366 |
+
| Top 1,000 | 64.8% |
|
| 367 |
+
| Top 5,000 | 79.6% |
|
| 368 |
+
| Top 10,000 | 84.6% |
|
| 369 |
|
| 370 |
### Key Findings
|
| 371 |
|
| 372 |
+
- **Zipf Compliance:** R²=0.9955 indicates excellent adherence to Zipf's law
|
| 373 |
+
- **High Frequency Dominance:** Top 100 words cover 34.4% of corpus
|
| 374 |
+
- **Long Tail:** 261,198 words needed for remaining 15.4% coverage
|
| 375 |
|
| 376 |
---
|
| 377 |
## 5. Word Embeddings Evaluation
|
|
|
|
| 384 |
|
| 385 |

|
| 386 |
|
|
|
|
| 387 |
|
| 388 |
+
### 5.1 Cross-Lingual Alignment
|
| 389 |
+
|
| 390 |
+
> *Note: Multilingual alignment visualization not available for this language.*
|
| 391 |
+
|
| 392 |
+
|
| 393 |
+
### 5.2 Model Comparison
|
| 394 |
+
|
| 395 |
+
| Model | Dimension | Isotropy | Semantic Density | Alignment R@1 | Alignment R@10 |
|
| 396 |
+
|-------|-----------|----------|------------------|---------------|----------------|
|
| 397 |
+
| **mono_32d** | 32 | 0.8282 🏆 | 0.3614 | N/A | N/A |
|
| 398 |
+
| **mono_64d** | 64 | 0.7952 | 0.3099 | N/A | N/A |
|
| 399 |
+
| **mono_128d** | 128 | 0.7570 | 0.2493 | N/A | N/A |
|
| 400 |
|
| 401 |
### Key Findings
|
| 402 |
|
| 403 |
+
- **Best Isotropy:** mono_32d with 0.8282 (more uniform distribution)
|
| 404 |
+
- **Semantic Density:** Average pairwise similarity of 0.3069. Lower values indicate better semantic separation.
|
| 405 |
+
- **Alignment Quality:** No aligned models evaluated in this run.
|
| 406 |
+
- **Recommendation:** 128d aligned for best cross-lingual performance
|
| 407 |
+
|
| 408 |
+
---
|
| 409 |
+
## 6. Morphological Analysis (Experimental)
|
| 410 |
+
|
| 411 |
+
> ⚠️ **Warning:** This language shows low morphological productivity. The statistical signals used for this analysis may be noisy or less reliable than for morphologically rich languages.
|
| 412 |
+
|
| 413 |
+
This section presents an automated morphological analysis derived from the statistical divergence between word-level and subword-level models. By analyzing where subword predictability spikes and where word-level coverage fails, we can infer linguistic structures without supervised data.
|
| 414 |
+
|
| 415 |
+
### 6.1 Productivity & Complexity
|
| 416 |
+
|
| 417 |
+
| Metric | Value | Interpretation | Recommendation |
|
| 418 |
+
|--------|-------|----------------|----------------|
|
| 419 |
+
| Productivity Index | **0.000** | Low morphological productivity | ⚠️ Likely unreliable |
|
| 420 |
+
| Idiomaticity Gap | **-1.000** | Low formulaic content | - |
|
| 421 |
+
|
| 422 |
+
### 6.2 Affix Inventory (Productive Units)
|
| 423 |
+
|
| 424 |
+
These are the most productive prefixes and suffixes identified by sampling the vocabulary for global substitutability patterns. A unit is considered an affix if stripping it leaves a valid stem that appears in other contexts.
|
| 425 |
+
|
| 426 |
+
#### Productive Prefixes
|
| 427 |
+
| Prefix | Examples |
|
| 428 |
+
|--------|----------|
|
| 429 |
+
|
| 430 |
+
#### Productive Suffixes
|
| 431 |
+
| Suffix | Examples |
|
| 432 |
+
|--------|----------|
|
| 433 |
+
| `-ین` | ائتدیگینین, دالینین, لشکرینین |
|
| 434 |
+
| `-ان` | سيران, کاپیتان, تاپیلمایان |
|
| 435 |
+
|
| 436 |
+
### 6.3 Bound Stems (Lexical Roots)
|
| 437 |
+
|
| 438 |
+
Bound stems are high-frequency subword units that are semantically cohesive but rarely appear as standalone words. These often correspond to the 'core' of a word that requires inflection or derivation to be valid.
|
| 439 |
+
|
| 440 |
+
| Stem | Cohesion | Substitutability | Examples |
|
| 441 |
+
|------|----------|------------------|----------|
|
| 442 |
+
| `رلری` | 1.93x | 205 contexts | ارلری, یرلری, دیرلری |
|
| 443 |
+
| `اقلا` | 1.95x | 131 contexts | ناقلا, آیاقلا, آياقلا |
|
| 444 |
+
| `قلار` | 2.11x | 54 contexts | لیقلار, حقلاری, ماقلار |
|
| 445 |
+
| `تیند` | 1.93x | 72 contexts | تیندل, تینده, اتیندن |
|
| 446 |
+
| `یبدی` | 2.32x | 31 contexts | آلیبدی, ییبدیر, گلیبدی |
|
| 447 |
+
| `اریخ` | 1.93x | 41 contexts | تا��یخ, تاریخه, تاریخ |
|
| 448 |
+
| `ولوب` | 1.70x | 60 contexts | کولوب, گولوب, سولوب |
|
| 449 |
+
| `ئرلش` | 2.00x | 24 contexts | يئرلشن, یئرلشن, یئرلشه |
|
| 450 |
+
| `ریخی` | 2.03x | 22 contexts | مریخی, ریخین, تاریخی |
|
| 451 |
+
| `یناق` | 1.87x | 27 contexts | سیناق, قیناق, ایناق |
|
| 452 |
+
| `هرلر` | 2.14x | 17 contexts | شهرلر, شهرلره, شهرلري |
|
| 453 |
+
| `یلیس` | 1.56x | 43 contexts | هیلیس, یلیسی, تیلیس |
|
| 454 |
+
|
| 455 |
+
### 6.4 Affix Compatibility (Co-occurrence)
|
| 456 |
+
|
| 457 |
+
This table shows which prefixes and suffixes most frequently co-occur on the same stems, revealing the 'stacking' rules of the language's morphology.
|
| 458 |
+
|
| 459 |
+
*No significant affix co-occurrences detected.*
|
| 460 |
+
|
| 461 |
+
|
| 462 |
+
### 6.5 Recursive Morpheme Segmentation
|
| 463 |
+
|
| 464 |
+
Using **Recursive Hierarchical Substitutability**, we decompose complex words into their constituent morphemes. This approach handles nested affixes (e.g., `prefix-prefix-root-suffix`).
|
| 465 |
+
|
| 466 |
+
| Word | Suggested Split | Confidence | Stem |
|
| 467 |
+
|------|-----------------|------------|------|
|
| 468 |
+
| وطنداشلارینین | **`وطنداشلار-ین-ین`** | 6.0 | `وطنداشلار` |
|
| 469 |
+
| تورپاقلارینین | **`تورپاقلار-ین-ین`** | 6.0 | `تورپاقلار` |
|
| 470 |
+
| دئموکراتلارینین | **`دئموکراتلار-ین-ین`** | 6.0 | `دئموکراتلار` |
|
| 471 |
+
| اوستانلارینان | **`اوستانلار-ین-ان`** | 6.0 | `اوستانلار` |
|
| 472 |
+
| تولیدینین | **`تولید-ین-ین`** | 6.0 | `تولید` |
|
| 473 |
+
| المنتلرینین | **`المنتلر-ین-ین`** | 6.0 | `المنتلر` |
|
| 474 |
+
| جومهوریتلرین | **`جومهوریتلر-ین`** | 4.5 | `جومهوریتلر` |
|
| 475 |
+
| سیمالارین | **`سیمالار-ین`** | 4.5 | `سیمالار` |
|
| 476 |
+
| ژوآنلارین | **`ژوآنلار-ین`** | 4.5 | `ژوآنلار` |
|
| 477 |
+
| ماشینلارین | **`ماشینلار-ین`** | 4.5 | `ماشینلار` |
|
| 478 |
+
| بدویلرین | **`بدویلر-ین`** | 4.5 | `بدویلر` |
|
| 479 |
+
| تبریزلیلرین | **`تبریزلیلر-ین`** | 4.5 | `تبریزلیلر` |
|
| 480 |
+
| ناخوشلوقلارین | **`ناخوشلوقلار-ین`** | 4.5 | `ناخوشلوقلار` |
|
| 481 |
+
| اویکونیمین | **`اویکونیم-ین`** | 4.5 | `اویکونیم` |
|
| 482 |
+
| ائرککلرین | **`ائرککلر-ین`** | 4.5 | `ائرککلر` |
|
| 483 |
+
|
| 484 |
+
### 6.6 Linguistic Interpretation
|
| 485 |
+
|
| 486 |
+
> **Automated Insight:**
|
| 487 |
+
The language AZB appears to be more isolating or has a highly fixed vocabulary. Word-level models perform nearly as well as subword models, indicating fewer productive morphological processes.
|
| 488 |
|
| 489 |
---
|
| 490 |
+
## 7. Summary & Recommendations
|
| 491 |
|
| 492 |

|
| 493 |
|
|
|
|
| 495 |
|
| 496 |
| Component | Recommended | Rationale |
|
| 497 |
|-----------|-------------|-----------|
|
| 498 |
+
| Tokenizer | **64k BPE** | Best compression (4.15x) |
|
| 499 |
+
| N-gram | **2-gram** | Lowest perplexity (528) |
|
| 500 |
+
| Markov | **Context-4** | Highest predictability (96.6%) |
|
| 501 |
| Embeddings | **100d** | Balanced semantic capture and isotropy |
|
| 502 |
|
| 503 |
+
|
| 504 |
---
|
| 505 |
## Appendix: Metrics Glossary & Interpretation Guide
|
| 506 |
|
|
|
|
| 690 |
author = {Kamali, Omar},
|
| 691 |
title = {Wikilangs: Open NLP Models for Wikipedia Languages},
|
| 692 |
year = {2025},
|
| 693 |
+
doi = {10.5281/zenodo.18073153},
|
| 694 |
+
publisher = {Zenodo},
|
| 695 |
url = {https://huggingface.co/wikilangs}
|
| 696 |
institution = {Omneity Labs}
|
| 697 |
}
|
|
|
|
| 707 |
- 🤗 Models: [huggingface.co/wikilangs](https://huggingface.co/wikilangs)
|
| 708 |
- 📊 Data: [wikipedia-monthly](https://huggingface.co/datasets/omarkamali/wikipedia-monthly)
|
| 709 |
- 👤 Author: [Omar Kamali](https://huggingface.co/omarkamali)
|
| 710 |
+
- 🤝 Sponsor: [Featherless AI](https://featherless.ai)
|
| 711 |
---
|
| 712 |
*Generated by Wikilangs Models Pipeline*
|
| 713 |
|
| 714 |
+
*Report Date: 2026-01-03 06:14:53*
|
models/embeddings/monolingual/azb_128d.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0affe475b2c95d7dbb76d561e1c19235d6d09956c11cb77486d2ed42c5bc91f8
|
| 3 |
+
size 1145287401
|
models/embeddings/monolingual/azb_128d_metadata.json
CHANGED
|
@@ -3,11 +3,13 @@
|
|
| 3 |
"dimension": 128,
|
| 4 |
"version": "monolingual",
|
| 5 |
"training_params": {
|
| 6 |
-
"
|
| 7 |
"min_count": 5,
|
| 8 |
"window": 5,
|
| 9 |
"negative": 5,
|
| 10 |
-
"epochs": 5
|
|
|
|
|
|
|
| 11 |
},
|
| 12 |
-
"vocab_size":
|
| 13 |
}
|
|
|
|
| 3 |
"dimension": 128,
|
| 4 |
"version": "monolingual",
|
| 5 |
"training_params": {
|
| 6 |
+
"algorithm": "skipgram",
|
| 7 |
"min_count": 5,
|
| 8 |
"window": 5,
|
| 9 |
"negative": 5,
|
| 10 |
+
"epochs": 5,
|
| 11 |
+
"encoding_method": "rope",
|
| 12 |
+
"dim": 128
|
| 13 |
},
|
| 14 |
+
"vocab_size": 116118
|
| 15 |
}
|
models/embeddings/monolingual/azb_32d.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd0d8e4ac83c586e8f75145c0ce099bef44ff8d39e955040bc9454b60e32022c
|
| 3 |
+
size 288108777
|
models/embeddings/monolingual/azb_32d_metadata.json
CHANGED
|
@@ -3,11 +3,13 @@
|
|
| 3 |
"dimension": 32,
|
| 4 |
"version": "monolingual",
|
| 5 |
"training_params": {
|
| 6 |
-
"
|
| 7 |
"min_count": 5,
|
| 8 |
"window": 5,
|
| 9 |
"negative": 5,
|
| 10 |
-
"epochs": 5
|
|
|
|
|
|
|
| 11 |
},
|
| 12 |
-
"vocab_size":
|
| 13 |
}
|
|
|
|
| 3 |
"dimension": 32,
|
| 4 |
"version": "monolingual",
|
| 5 |
"training_params": {
|
| 6 |
+
"algorithm": "skipgram",
|
| 7 |
"min_count": 5,
|
| 8 |
"window": 5,
|
| 9 |
"negative": 5,
|
| 10 |
+
"epochs": 5,
|
| 11 |
+
"encoding_method": "rope",
|
| 12 |
+
"dim": 32
|
| 13 |
},
|
| 14 |
+
"vocab_size": 116118
|
| 15 |
}
|
models/embeddings/monolingual/azb_64d.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2aae88ec4bdf46e7654ecacada31d2b3be0690719379b657e44df7793b45fa65
|
| 3 |
+
size 573834985
|
models/embeddings/monolingual/azb_64d_metadata.json
CHANGED
|
@@ -3,11 +3,13 @@
|
|
| 3 |
"dimension": 64,
|
| 4 |
"version": "monolingual",
|
| 5 |
"training_params": {
|
| 6 |
-
"
|
| 7 |
"min_count": 5,
|
| 8 |
"window": 5,
|
| 9 |
"negative": 5,
|
| 10 |
-
"epochs": 5
|
|
|
|
|
|
|
| 11 |
},
|
| 12 |
-
"vocab_size":
|
| 13 |
}
|
|
|
|
| 3 |
"dimension": 64,
|
| 4 |
"version": "monolingual",
|
| 5 |
"training_params": {
|
| 6 |
+
"algorithm": "skipgram",
|
| 7 |
"min_count": 5,
|
| 8 |
"window": 5,
|
| 9 |
"negative": 5,
|
| 10 |
+
"epochs": 5,
|
| 11 |
+
"encoding_method": "rope",
|
| 12 |
+
"dim": 64
|
| 13 |
},
|
| 14 |
+
"vocab_size": 116118
|
| 15 |
}
|
models/subword_markov/azb_markov_ctx1_subword.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4ff3da37abf00dd52bdcb28478a692c4366754cc1a111aa42559670cd7ea3aee
|
| 3 |
+
size 247087
|
models/subword_markov/azb_markov_ctx1_subword_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"context_size": 1,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_contexts":
|
| 6 |
-
"total_transitions":
|
| 7 |
}
|
|
|
|
| 2 |
"context_size": 1,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_contexts": 3409,
|
| 6 |
+
"total_transitions": 91613484
|
| 7 |
}
|
models/subword_markov/azb_markov_ctx2_subword.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1090e7d0037ce29d3cddcfe27f87428fcd5c872805dcf5e3d4202b9f83929b66
|
| 3 |
+
size 1650829
|
models/subword_markov/azb_markov_ctx2_subword_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"context_size": 2,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_contexts":
|
| 6 |
-
"total_transitions":
|
| 7 |
}
|
|
|
|
| 2 |
"context_size": 2,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_contexts": 30905,
|
| 6 |
+
"total_transitions": 91370829
|
| 7 |
}
|
models/subword_markov/azb_markov_ctx3_subword.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06a5083f143c1f0688e95d617ee8d409f5c78a44f6ea6c79e014253dcf249ac5
|
| 3 |
+
size 7956042
|
models/subword_markov/azb_markov_ctx3_subword_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"context_size": 3,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_contexts":
|
| 6 |
-
"total_transitions":
|
| 7 |
}
|
|
|
|
| 2 |
"context_size": 3,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_contexts": 203056,
|
| 6 |
+
"total_transitions": 91128174
|
| 7 |
}
|
models/subword_markov/azb_markov_ctx4_subword.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d4bd1bdcfb24b934a84145355781bb3e8cf414b875f467fc28d048b5d64ec5d1
|
| 3 |
+
size 26328560
|
models/subword_markov/azb_markov_ctx4_subword_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"context_size": 4,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_contexts":
|
| 6 |
-
"total_transitions":
|
| 7 |
}
|
|
|
|
| 2 |
"context_size": 4,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_contexts": 953883,
|
| 6 |
+
"total_transitions": 90885519
|
| 7 |
}
|
models/subword_ngram/azb_2gram_subword.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:383d8448d14aee343af41336d2c34807c417414896fb0905958afee3cec1c929
|
| 3 |
+
size 182774
|
models/subword_ngram/azb_2gram_subword_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"n": 2,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_ngrams":
|
| 6 |
-
"total_ngrams":
|
| 7 |
}
|
|
|
|
| 2 |
"n": 2,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_ngrams": 12648,
|
| 6 |
+
"total_ngrams": 91613484
|
| 7 |
}
|
models/subword_ngram/azb_3gram_subword.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed1d402ce0eac47e37053808b7981cf3eb00b76c5e2e5e887890b036ff01d869
|
| 3 |
+
size 1377099
|
models/subword_ngram/azb_3gram_subword_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"n": 3,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_ngrams":
|
| 6 |
-
"total_ngrams":
|
| 7 |
}
|
|
|
|
| 2 |
"n": 3,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_ngrams": 106644,
|
| 6 |
+
"total_ngrams": 91370829
|
| 7 |
}
|
models/subword_ngram/azb_4gram_subword.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cde924d7fc46a3c6a14d25e38135fbf37a14e774606b67187d527dd381b1d75f
|
| 3 |
+
size 7368829
|
models/subword_ngram/azb_4gram_subword_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"n": 4,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_ngrams":
|
| 6 |
-
"total_ngrams":
|
| 7 |
}
|
|
|
|
| 2 |
"n": 4,
|
| 3 |
"variant": "subword",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_ngrams": 581225,
|
| 6 |
+
"total_ngrams": 91128174
|
| 7 |
}
|
models/tokenizer/azb_tokenizer_16k.model
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c4f3786bae820ff2d6cc627dbdc96f1ede4586a6afa95391543c078559a7c125
|
| 3 |
+
size 527370
|
models/tokenizer/azb_tokenizer_16k.vocab
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
models/tokenizer/azb_tokenizer_32k.model
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d5380be73c3e18000f3f24bd0d0309c1b6d15259c73623f50c42cea3c6f8992d
|
| 3 |
+
size 820944
|
models/tokenizer/azb_tokenizer_32k.vocab
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
models/tokenizer/azb_tokenizer_64k.model
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:efb9e4835aafb1f0f10715a303c2ac20f1f169ed4ffc46f49d685eea42a8bfae
|
| 3 |
+
size 1430567
|
models/tokenizer/azb_tokenizer_64k.vocab
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
models/tokenizer/azb_tokenizer_8k.model
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e6b1776330fc0caaa7cc96df48505e83e46f3a00daa53e66e6a94354397d3dd8
|
| 3 |
+
size 384993
|
models/tokenizer/azb_tokenizer_8k.vocab
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
models/vocabulary/azb_vocabulary.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1346fa231de1649149828cb0cae257f644b1914797c84a5a8c4b56c62e716253
|
| 3 |
+
size 4348211
|
models/vocabulary/azb_vocabulary_metadata.json
CHANGED
|
@@ -1,16 +1,17 @@
|
|
| 1 |
{
|
| 2 |
"language": "azb",
|
| 3 |
-
"vocabulary_size":
|
|
|
|
| 4 |
"statistics": {
|
| 5 |
-
"type_token_ratio": 0.
|
| 6 |
"coverage": {
|
| 7 |
-
"top_100": 0.
|
| 8 |
-
"top_1000": 0.
|
| 9 |
-
"top_5000": 0.
|
| 10 |
-
"top_10000": 0.
|
| 11 |
},
|
| 12 |
-
"hapax_count":
|
| 13 |
-
"hapax_ratio": 0.
|
| 14 |
-
"total_documents":
|
| 15 |
}
|
| 16 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"language": "azb",
|
| 3 |
+
"vocabulary_size": 271198,
|
| 4 |
+
"variant": "full",
|
| 5 |
"statistics": {
|
| 6 |
+
"type_token_ratio": 0.05623983023140009,
|
| 7 |
"coverage": {
|
| 8 |
+
"top_100": 0.33209927062556827,
|
| 9 |
+
"top_1000": 0.6250560987377987,
|
| 10 |
+
"top_5000": 0.7676335196346162,
|
| 11 |
+
"top_10000": 0.8163255618590586
|
| 12 |
},
|
| 13 |
+
"hapax_count": 456252,
|
| 14 |
+
"hapax_ratio": 0.627193621554746,
|
| 15 |
+
"total_documents": 242655
|
| 16 |
}
|
| 17 |
}
|
models/word_markov/azb_markov_ctx1_word.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:816eaec35f734257259dcbb9d574a8c019635474f0de8b0daca5e593a77ac78b
|
| 3 |
+
size 42654567
|
models/word_markov/azb_markov_ctx1_word_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"context_size": 1,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_contexts":
|
| 6 |
-
"total_transitions":
|
| 7 |
}
|
|
|
|
| 2 |
"context_size": 1,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_contexts": 726930,
|
| 6 |
+
"total_transitions": 12692128
|
| 7 |
}
|
models/word_markov/azb_markov_ctx2_word.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:46d9babbdaaaad5f5dc9ac4a05883d1447bd376822b6b3dc01d9bfa81167154f
|
| 3 |
+
size 101955475
|
models/word_markov/azb_markov_ctx2_word_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"context_size": 2,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_contexts":
|
| 6 |
-
"total_transitions":
|
| 7 |
}
|
|
|
|
| 2 |
"context_size": 2,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_contexts": 3693091,
|
| 6 |
+
"total_transitions": 12449473
|
| 7 |
}
|
models/word_markov/azb_markov_ctx3_word.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:950723c3851972daff338e7c32d25b6b2711a549f145cfc4f36896cd3c7ac632
|
| 3 |
+
size 137860035
|
models/word_markov/azb_markov_ctx3_word_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"context_size": 3,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_contexts":
|
| 6 |
-
"total_transitions":
|
| 7 |
}
|
|
|
|
| 2 |
"context_size": 3,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_contexts": 5447170,
|
| 6 |
+
"total_transitions": 12206818
|
| 7 |
}
|
models/word_markov/azb_markov_ctx4_word.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8542b69e6498bf99f08ac81a2c69e6e10cfaec85f6041228b2d33e47374fa2f
|
| 3 |
+
size 167254340
|
models/word_markov/azb_markov_ctx4_word_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"context_size": 4,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_contexts":
|
| 6 |
-
"total_transitions":
|
| 7 |
}
|
|
|
|
| 2 |
"context_size": 4,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_contexts": 6178524,
|
| 6 |
+
"total_transitions": 11964164
|
| 7 |
}
|
models/word_ngram/azb_2gram_word.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:38d788975186225cbd44f6dc3f5933b5ef0372dd125cccbf762b7c146ac00818
|
| 3 |
+
size 2898203
|
models/word_ngram/azb_2gram_word_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"n": 2,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_ngrams":
|
| 6 |
-
"total_ngrams":
|
| 7 |
}
|
|
|
|
| 2 |
"n": 2,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_ngrams": 158908,
|
| 6 |
+
"total_ngrams": 12692128
|
| 7 |
}
|
models/word_ngram/azb_3gram_word.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e3b2c502cac90ffea868839ff83a17e6f50d9e2b94a2406bbe1479ad2b587f88
|
| 3 |
+
size 4982020
|
models/word_ngram/azb_3gram_word_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"n": 3,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_ngrams":
|
| 6 |
-
"total_ngrams":
|
| 7 |
}
|
|
|
|
| 2 |
"n": 3,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_ngrams": 236749,
|
| 6 |
+
"total_ngrams": 12449473
|
| 7 |
}
|
models/word_ngram/azb_4gram_word.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ca451d9d2e6c1c4cd584e9ce07d419f547d1fe4ce2dc69563f37268bba623e4e
|
| 3 |
+
size 10534721
|
models/word_ngram/azb_4gram_word_metadata.json
CHANGED
|
@@ -2,6 +2,6 @@
|
|
| 2 |
"n": 4,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
-
"unique_ngrams":
|
| 6 |
-
"total_ngrams":
|
| 7 |
}
|
|
|
|
| 2 |
"n": 4,
|
| 3 |
"variant": "word",
|
| 4 |
"language": "azb",
|
| 5 |
+
"unique_ngrams": 426395,
|
| 6 |
+
"total_ngrams": 12206818
|
| 7 |
}
|
visualizations/embedding_isotropy.png
CHANGED
|
|
visualizations/embedding_norms.png
CHANGED
|
|
visualizations/embedding_similarity.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|
visualizations/markov_branching.png
CHANGED
|
|
visualizations/markov_contexts.png
CHANGED
|
|