Best model so far - eval_loss: 0.3902 (Trained with Unsloth)

Best model so far - eval_loss: 0.3946 (Trained with Unsloth)

Best model so far - eval_loss: 0.4001 (Trained with Unsloth)

Best model so far - eval_loss: 0.4075 (Trained with Unsloth)

Best model so far - eval_loss: 0.4195 (Trained with Unsloth)

Best model so far - eval_loss: 0.4401 (Trained with Unsloth)

Initial model state (Trained with Unsloth)

initial commit