fam/nanochat-omni
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix buggy midtrain and update all kwargs to be idiomatic. that is, argparse uses dashes variables use underscores. the underscores are just a remnant of the previous Configurator object. This is the right way

Browse Source
This commit is contained in:
Andrej Karpathy
2026-01-13 22:45:27 +00:00
parent 3b50b77ed3
commit 7312ec9898
11 changed files with 144 additions and 139 deletions
Download Patch File Download Diff File Expand all files Collapse all files
speedrun.sh
+2 -2
View File
@@ -59,7 +59,7 @@ python -m nanochat.dataset -n 8
python -m nanochat.dataset -n 370 &
DATASET_DOWNLOAD_PID=$!
# train the tokenizer with vocab size 2**16 = 65536 on ~2B characters of data
python -m scripts.tok_train --max_chars=2000000000 --vocab_size=65536
python -m scripts.tok_train --max-chars=2000000000 --vocab-size=65536
# evaluate the tokenizer (report compression ratio etc.)
python -m scripts.tok_eval
@@ -81,7 +81,7 @@ wait $DATASET_DOWNLOAD_PID
NPROC_PER_NODE=8
# pretrain the d20 model
torchrun --standalone --nproc_per_node=$NPROC_PER_NODE -m scripts.base_train -- --depth=20 --target_param_data_ratio=20 --run=$WANDB_RUN
torchrun --standalone --nproc_per_node=$NPROC_PER_NODE -m scripts.base_train -- --depth=20 --target-param-data-ratio=20 --run=$WANDB_RUN
# evaluate the model on a larger chunk of train/val data and draw some samples
torchrun --standalone --nproc_per_node=$NPROC_PER_NODE -m scripts.base_loss
# evaluate the model on CORE tasks