Update README.md

README.md

@@ -25,3 +25,23 @@ configs:
   - split: train
     path: data/train-*
 ---
+
+Generated using https://github.com/FranxYao/Long-Context-Data-Engineering with the below command:
+
+```bash
+mkdir logs
+mkdir data
+mkdir data/slimpajama
+mkdir data/slimpajama/per_source_downsample
+cd data_engineering
+
+PATH_TO_SLIMPAJAMA=rokset3/slim_pajama_chunk_1
+nohup python -u slimpajama_packing.py\
+    --dataset_size=100m\
+    --print_interval=100 --num_process=200\
+    --chunk_size=1000001 \
+    --dataset_path=$PATH_TO_SLIMPAJAMA\
+    --output_path=../data/slimpajama/per_source_downsample/ --down_sample_ratio=0.1 --down_sample_mode=per_source\
+    > ../logs/slimpajama_packing_dist_per_source_downsample_0.1.log 2>&1 &
+tail -f ../logs/slimpajama_packing_dist_per_source_downsample_0.1.log
+```