Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,31 @@
+---
+license: odc-by
+tags:
+  - gsa
+  - gist-sparse-attention
+  - long-context
+  - pretraining
+source_datasets:
+  - HuggingFaceFW/fineweb
+---
+
+# GSA-PT-Llama-3.2-1B-chunk16-data
+
+This is the **continue pretraining dataset** used for training GSA (Gist Sparse Attention) models with chunk size `chunk16`.
+
+## Dataset Description
+
+- **Source**: Filtered from [FineWeb (sample-10BT)](https://huggingface.co/datasets/HuggingFaceFW/fineweb/viewer/sample-10BT)
+- **Filtering**: Samples with sequence lengths ranging from **2K to 4K** tokens
+- **Chunk size**: `chunk16`
+- **Splits**: `train`, `eval`
+
+Each sample is tokenized and formatted with GSA gist tokens for continued pretraining.
+
+## Paper
+
+**GSA: Gist Sparse Attention via Learnable Compression and Selective Unfolding**
+
+## Related Models
+
+- [yuzhenm/GSA-PT-Llama-3.2-1B-chunk16](https://huggingface.co/yuzhenm/GSA-PT-Llama-3.2-1B-chunk16) — model trained on this dataset