Best model so far - eval_loss: 0.1945 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.1959 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.1980 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2007 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2083 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2109 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2201 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2210 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2292 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2420 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2421 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2461 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2512 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2550 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2607 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2743 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2829 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2879 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.3008 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.3145 -------------------------------------------------------- (Trained with Unsloth)

Initial model state ------------------------------------------------------------------------ (Trained with Unsloth)

Best model so far - eval_loss: 0.2111 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2111 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2111 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2112 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2115 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2146 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2158 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2218 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2259 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2270 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2320 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2333 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2344 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2349 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2379 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2454 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2567 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2587 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2597 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2611 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2614 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2832 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2952 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2971 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.2976 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.3186 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.3264 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.3385 -------------------------------------------------------- (Trained with Unsloth)

Best model so far - eval_loss: 0.3466 -------------------------------------------------------- (Trained with Unsloth)