Upload scripts/dl_t32.sh with huggingface_hub
Browse files- scripts/dl_t32.sh +13 -0
scripts/dl_t32.sh
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/usr/bin/env bash
|
| 2 |
+
set -uo pipefail; cd "$(dirname "$0")"
|
| 3 |
+
export PATH=/root/Adhi/BluTrain/envv/bin:$PATH
|
| 4 |
+
R=unparallelled/BluScriptCP-data
|
| 5 |
+
get(){ for i in $(seq -f '%05g' "$2" "$3"); do
|
| 6 |
+
hf download "$R" --repo-type dataset --quiet \
|
| 7 |
+
--include "cp_data_by_length/$1/train_shard_$i.bin" --local-dir dl32 || return 1; done; }
|
| 8 |
+
echo "b4 32-64k (docs >= the 32k window: contiguous single-doc context)"; get b4_32768_65536 0 6
|
| 9 |
+
echo "b3 16-32k"; get b3_16384_32768 0 3
|
| 10 |
+
echo "b1 4-8k"; get b1_4096_8192 0 2
|
| 11 |
+
for d in dl32/cp_data_by_length/*; do n=$(basename "$d"); mkdir -p "src32/$n"
|
| 12 |
+
for f in "$d"/*.bin; do [ -e "$f" ] && mv -n "$f" "src32/$n/"; done; done
|
| 13 |
+
rm -rf dl32; du -sh src32/*; df -h . | tail -1; echo DL_DONE
|