tacit/nanochat
1
0
Fork 0
mirror of https://github.com/karpathy/nanochat.git synced 2026-04-21 08:18:40 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

revert miniseries.sh to original, this should run fine with G4 instance in colab

Browse Source
This commit is contained in:
Your Name 2026-03-04 13:26:21 -08:00
parent c48aa05531
commit 676ddfdb46
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
runs/miniseries.sh
View File

@ -28,7 +28,7 @@ fi
# Series name: from arg, env var, or default to today's date (e.g., jan11)
SERIES_NAME="${1:-${SERIES_NAME:-$(date +%b%d | tr '[:upper:]' '[:lower:]')}}"
# Depths to train (the "miniseries")
DEPTHS=(12 14)
DEPTHS=(12 14 16 18 20 22 24 26)
# Hardware
NPROC_PER_NODE="${NPROC_PER_NODE:-8}"
# Logging
@ -63,7 +63,7 @@ for d in "${DEPTHS[@]}"; do
elif [ $d -ge 20 ]; then
DEVICE_BATCH_SIZE_ARG="--device-batch-size=16"
else
DEVICE_BATCH_SIZE_ARG="--device-batch-size=16"
DEVICE_BATCH_SIZE_ARG="--device-batch-size=32"
fi
torchrun --standalone --nproc_per_node=$NPROC_PER_NODE -m scripts.base_train -- \