080-ai/

flintlock_3B_v0.1

A fine-tuned openllama 3B model, using primary sources from US History to provide a deeper understanding of the historical context.

Run history:

train/epoch ▁▁▂▂▂▃▃▄▄▄▅▅▅▆▆▇▇▇███
train/global_step ▁▁▂▂▂▃▃▄▄▄▅▅▅▆▆▇▇▇███
train/grad_norm ██▄▅▄▅▃▂▂▄▂▂▂▂▁▁▁▁▁▁
train/learning_rate ▂▂▃▄▅▅▆▇▇█▇▇▆▅▅▄▃▂▂▁
train/loss ▇█▇▆▅▅▄▄▃▃▂▂▂▁▂▁▁▁▁▁
train/total_flos ▁
train/train_loss ▁
train/train_runtime ▁
train/train_samples_per_second ▁
train/train_steps_per_second ▁

Run summary:

train/epoch 2.0
train/global_step 20
train/grad_norm 0.13779
train/learning_rate 0.0
train/loss 1.1365
train/total_flos 4.579249185376512e+16
train/train_loss 1.29891
train/train_runtime 1552.5749
train/train_samples_per_second 1.649
train/train_steps_per_second 0.013