Delta-Vector/Rei-24B-Base
115
base model for Rei-24B-KTO
I would not use this - better to use the downstream KTO, Released because i <3 OSS
wandb is linked in the KTO repo.
trained on 8xA100s for 8 hours for 2 epochs.
base model for Rei-24B-KTO
I would not use this - better to use the downstream KTO, Released because i <3 OSS
wandb is linked in the KTO repo.
trained on 8xA100s for 8 hours for 2 epochs.