Create README.md

README.md

@@ -0,0 +1,57 @@
+# Qwen3-Inspired Pre-training Dataset
+
+## Overview
+This is a demo version with sampled data from each source.
+
+This dataset is a curated mixture of high-quality text data designed for large language model pre-training, inspired by the Qwen3 methodology.
+
+## Dataset Statistics
+
+**Total Size:** 6.78 billion tokens
+
+### Data Sources
+
+- **dclm_baseline**: 3.71B tokens (54.78%) - 2,981,871 documents
+- **mini_pile**: 1.43B tokens (21.09%) - 998,100 documents
+- **the_stack**: 0.82B tokens (12.09%) - 148,195 documents
+- **common_corpus**: 0.82B tokens (12.04%) - 194,877 documents
+
+
+## Data Processing Pipeline
+
+1. **Data Collection**: Sourced from multiple high-quality datasets
+2. **Standardization**: All data transformed to consistent format with `text`, `info`, and `source_data` fields  
+3. **Exact Deduplication**: Removed identical documents
+4. **Near Deduplication**: Used MinHashLSH with Jaccard similarity threshold of 0.85
+5. **Quality Filtering**: Applied content-based filtering during processing
+
+## Data Format
+
+Each example contains:
+- `text`: The main text content
+- `info`: Metadata from the original dataset (as string)
+- `source_data`: Source dataset identifier
+
+## Tokenization
+
+Token counts were computed using the Llama3 tokenizer (`meta-llama/Meta-Llama-3-8B`).
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+dataset = load_dataset("bluelightai-dev/qwen_clt_pretrain_data")
+```
+
+## Dataset Sources
+
+The dataset combines data from the following sources:
+- **DCLM Baseline**: High-quality web text from DataComp-LM
+- **Common Corpus**: Curated common crawl data  
+- **The Stack**: Deduplicated source code
+- **Mini Pile**: Academic and reference texts
+
+## License
+
+Please refer to the individual source dataset licenses. This mixture is provided for research purposes.