Edit model card

Time train more 12 H.

image/png

Metrics.

global_step=3783,
training_loss=0.2294789322652169,
metrics={'train_loss': 0.2294789322652169})

image/png

Colab

https://colab.research.google.com/drive/1iZRjkTm7Sv3_JZJx2MAk384hQuk-fq1W?usp=sharing
Downloads last month
6
Safetensors
Model size
2.51B params
Tensor type
F32
·
BF16
·