Comment
eval/dclm/loss v. trainer.train_batch_size
eval/dclm/loss v. trainer.train_batch_size
eval/dclm/loss v. trainer.num_train_steps
eval/dclm/loss v. trainer.num_train_steps
eval/dclm/loss v. optimizer.weight_decay
eval/dclm/loss v. optimizer.weight_decay
eval/dclm/loss
eval/dclm/loss
4
8
8
weight decay
24
chinchilla single runs
4
Add a comment
Created with ❤️ on Weights & Biases.
https://wandb.ai/stanford-mercury/suhas-data-efficiency/reports/Standard-scaling-runs---VmlldzoxMzg2NDQ3OQ