Duplicate from OptimalScale/ClimbMix

Co-authored-by: LMFlow <lmflow-optimalscale@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,159 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mds filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz 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
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text
+part_79.jsonl filter=lfs diff=lfs merge=lfs -text
+part_80.jsonl filter=lfs diff=lfs merge=lfs -text
+part_74.jsonl filter=lfs diff=lfs merge=lfs -text
+part_78.jsonl filter=lfs diff=lfs merge=lfs -text
+part_73.jsonl filter=lfs diff=lfs merge=lfs -text
+part_67.jsonl filter=lfs diff=lfs merge=lfs -text
+part_68.jsonl filter=lfs diff=lfs merge=lfs -text
+part_76.jsonl filter=lfs diff=lfs merge=lfs -text
+part_75.jsonl filter=lfs diff=lfs merge=lfs -text
+part_63.jsonl filter=lfs diff=lfs merge=lfs -text
+part_72.jsonl filter=lfs diff=lfs merge=lfs -text
+part_54.jsonl filter=lfs diff=lfs merge=lfs -text
+part_71.jsonl filter=lfs diff=lfs merge=lfs -text
+part_44.jsonl filter=lfs diff=lfs merge=lfs -text
+part_69.jsonl filter=lfs diff=lfs merge=lfs -text
+part_66.jsonl filter=lfs diff=lfs merge=lfs -text
+part_81.jsonl filter=lfs diff=lfs merge=lfs -text
+part_59.jsonl filter=lfs diff=lfs merge=lfs -text
+part_65.jsonl filter=lfs diff=lfs merge=lfs -text
+part_55.jsonl filter=lfs diff=lfs merge=lfs -text
+part_62.jsonl filter=lfs diff=lfs merge=lfs -text
+part_4.jsonl filter=lfs diff=lfs merge=lfs -text
+part_53.jsonl filter=lfs diff=lfs merge=lfs -text
+part_19.jsonl filter=lfs diff=lfs merge=lfs -text
+part_43.jsonl filter=lfs diff=lfs merge=lfs -text
+part_98.jsonl filter=lfs diff=lfs merge=lfs -text
+part_57.jsonl filter=lfs diff=lfs merge=lfs -text
+part_70.jsonl filter=lfs diff=lfs merge=lfs -text
+part_5.jsonl filter=lfs diff=lfs merge=lfs -text
+part_99.jsonl filter=lfs diff=lfs merge=lfs -text
+part_23.jsonl filter=lfs diff=lfs merge=lfs -text
+part_56.jsonl filter=lfs diff=lfs merge=lfs -text
+part_3.jsonl filter=lfs diff=lfs merge=lfs -text
+part_0.jsonl filter=lfs diff=lfs merge=lfs -text
+part_34.jsonl filter=lfs diff=lfs merge=lfs -text
+part_61.jsonl filter=lfs diff=lfs merge=lfs -text
+part_22.jsonl filter=lfs diff=lfs merge=lfs -text
+part_96.jsonl filter=lfs diff=lfs merge=lfs -text
+part_39.jsonl filter=lfs diff=lfs merge=lfs -text
+part_9.jsonl filter=lfs diff=lfs merge=lfs -text
+part_20.jsonl filter=lfs diff=lfs merge=lfs -text
+part_36.jsonl filter=lfs diff=lfs merge=lfs -text
+part_6.jsonl filter=lfs diff=lfs merge=lfs -text
+part_37.jsonl filter=lfs diff=lfs merge=lfs -text
+part_58.jsonl filter=lfs diff=lfs merge=lfs -text
+part_29.jsonl filter=lfs diff=lfs merge=lfs -text
+part_21.jsonl filter=lfs diff=lfs merge=lfs -text
+part_40.jsonl filter=lfs diff=lfs merge=lfs -text
+part_48.jsonl filter=lfs diff=lfs merge=lfs -text
+part_50.jsonl filter=lfs diff=lfs merge=lfs -text
+part_1.jsonl filter=lfs diff=lfs merge=lfs -text
+part_45.jsonl filter=lfs diff=lfs merge=lfs -text
+part_28.jsonl filter=lfs diff=lfs merge=lfs -text
+part_95.jsonl filter=lfs diff=lfs merge=lfs -text
+part_49.jsonl filter=lfs diff=lfs merge=lfs -text
+part_11.jsonl filter=lfs diff=lfs merge=lfs -text
+part_46.jsonl filter=lfs diff=lfs merge=lfs -text
+part_85.jsonl filter=lfs diff=lfs merge=lfs -text
+part_18.jsonl filter=lfs diff=lfs merge=lfs -text
+part_93.jsonl filter=lfs diff=lfs merge=lfs -text
+part_8.jsonl filter=lfs diff=lfs merge=lfs -text
+part_38.jsonl filter=lfs diff=lfs merge=lfs -text
+part_94.jsonl filter=lfs diff=lfs merge=lfs -text
+part_52.jsonl filter=lfs diff=lfs merge=lfs -text
+part_51.jsonl filter=lfs diff=lfs merge=lfs -text
+part_47.jsonl filter=lfs diff=lfs merge=lfs -text
+part_7.jsonl filter=lfs diff=lfs merge=lfs -text
+part_35.jsonl filter=lfs diff=lfs merge=lfs -text
+part_24.jsonl filter=lfs diff=lfs merge=lfs -text
+part_17.jsonl filter=lfs diff=lfs merge=lfs -text
+part_33.jsonl filter=lfs diff=lfs merge=lfs -text
+part_88.jsonl filter=lfs diff=lfs merge=lfs -text
+part_60.jsonl filter=lfs diff=lfs merge=lfs -text
+part_87.jsonl filter=lfs diff=lfs merge=lfs -text
+part_27.jsonl filter=lfs diff=lfs merge=lfs -text
+part_2.jsonl filter=lfs diff=lfs merge=lfs -text
+part_41.jsonl filter=lfs diff=lfs merge=lfs -text
+part_32.jsonl filter=lfs diff=lfs merge=lfs -text
+part_91.jsonl filter=lfs diff=lfs merge=lfs -text
+part_42.jsonl filter=lfs diff=lfs merge=lfs -text
+part_13.jsonl filter=lfs diff=lfs merge=lfs -text
+part_83.jsonl filter=lfs diff=lfs merge=lfs -text
+part_92.jsonl filter=lfs diff=lfs merge=lfs -text
+part_86.jsonl filter=lfs diff=lfs merge=lfs -text
+part_31.jsonl filter=lfs diff=lfs merge=lfs -text
+part_16.jsonl filter=lfs diff=lfs merge=lfs -text
+part_30.jsonl filter=lfs diff=lfs merge=lfs -text
+part_89.jsonl filter=lfs diff=lfs merge=lfs -text
+part_14.jsonl filter=lfs diff=lfs merge=lfs -text
+part_84.jsonl filter=lfs diff=lfs merge=lfs -text
+part_10.jsonl filter=lfs diff=lfs merge=lfs -text
+part_12.jsonl filter=lfs diff=lfs merge=lfs -text
+part_26.jsonl filter=lfs diff=lfs merge=lfs -text
+part_77.jsonl filter=lfs diff=lfs merge=lfs -text
+part_97.jsonl filter=lfs diff=lfs merge=lfs -text
+part_25.jsonl filter=lfs diff=lfs merge=lfs -text
+part_64.jsonl filter=lfs diff=lfs merge=lfs -text
+part_90.jsonl filter=lfs diff=lfs merge=lfs -text
+part_82.jsonl filter=lfs diff=lfs merge=lfs -text
+part_15.jsonl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,31 @@
+---
+license: apache-2.0
+task_categories:
+- text-generation
+language:
+- en
+---
+[ClimbMix](https://huggingface.co/datasets/nvidia/ClimbMix) is a high-quality pre-training corpus released by NVIDIA. Here is the description:
+
+>ClimbMix is a compact yet powerful 400-billion-token dataset designed for efficient pre-training that delivers superior performance under an equal token budget.  It was introduced in [this paper](https://huggingface.co/papers/2504.13161).
+We proposed a new algorithm to filter and mix the dataset. First, we grouped the data into 1,000 groups based on topic information. Then we applied two classifiers: one to detect advertisements and another to assess the educational value of the text. Each group was scored accordingly, and low-quality data with low scores was removed. Finally, the remaining high-quality groups were mixed using certain weights to generate the final dataset.
+
+But it is released in gpt-2 tokens which is not easy-to-use. Therefore,we use gpt-2 tokenizer to detokenize them into raw texts.
+
+⚠️ Please note: This version is not officially released or maintained by NVIDIA. We are not responsible for the content, accuracy, or updates of this dataset.
+
+
+## Citation:
+If you find this dataset helpful, please cite the following [paper](https://arxiv.org/abs/2504.13161):
+
+```
+@article{diao2025climb,
+  author    = {Shizhe Diao and Yu Yang and Yonggan Fu and Xin Dong and Dan Su and Markus Kliegl and Zijia Chen and Peter Belcak and Yoshi Suhara and Hongxu Yin and Mostofa Patwary and Celine Lin and Jan Kautz and Pavlo Molchanov},
+  title={CLIMB: CLustering-based Iterative Data Mixture Bootstrapping for Language Model Pre-training}, 
+  journal   = {arXiv preprint},
+  year      = {2025},
+  archivePrefix = {arXiv},
+  primaryClass = {cs.CL},
+  url={https://arxiv.org/abs/2504.13161}, 
+}
+```