merge_lora_and_push: job=82bce628b40a68e9a4a0c7f5eab74138feb67805dab5c5ccc7b32c931cc066ca checkpoint=checkpoint_9999.pt r=8 alpha=32 31a68f9 verified gdiamos commited on Jun 26
fix: restore 60 missing attention weights (k_proj/v_proj/k_norm layers 15-34) from base model 475472e verified gdiamos commited on Jun 25
merge_lora_and_push: job=82bce628b40a68e9a4a0c7f5eab74138feb67805dab5c5ccc7b32c931cc066ca checkpoint=checkpoint_9999.pt r=8 alpha=32 384de7c verified gdiamos commited on Jun 24