Fix AMD Triton re-installation issue in speedrun.sh

On AMD ROCm environments, `uv run` was detecting that the manually uninstalled `triton` package was missing (since it's a transitive dependency of `torch`) and reinstalling it during the tokenizer build step. This caused `ImportError: cannot import name 'Config' from 'triton'` due to conflict with `pytorch-triton-rocm`.

This change adds `--no-sync` to the `uv run` command for building the tokenizer, preventing `uv` from undoing the manual uninstallation of `triton`.
This commit is contained in:
google-labs-jules[bot] 2025-11-23 04:26:32 +00:00
parent 054394c708
commit d291a62ad8

View File

@ -74,7 +74,8 @@ curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y
source "$HOME/.cargo/env"
# Build the rustbpe Tokenizer
uv run --extra $EXTRAS maturin develop --release --manifest-path rustbpe/Cargo.toml
# use --no-sync to avoid re-installing triton on AMD, which we just uninstalled
uv run --no-sync --extra $EXTRAS maturin develop --release --manifest-path rustbpe/Cargo.toml
# Download the first ~2B characters of pretraining dataset
# look at dev/repackage_data_reference.py for details on how this data was prepared