rlm-arithmetic-training / train_arithmetic.py

Commit History

Fix hang: remove use_cpu parameter, reduce generations to 2, batch to 2, steps to 20
95008ad
verified

mindchain commited on

Fix NoneType.shape error: device handling, CPU optimizer, safe tensor ops
0168a3e
verified

mindchain commited on

Upload train_arithmetic.py with huggingface_hub
61cc0c7
verified

mindchain commited on

Upload train_arithmetic.py with huggingface_hub
306c5f0
verified

mindchain commited on