type eval | step 0 | loss 4.8501 | checkpoint False type eval | step 0 | loss 4.8489 | checkpoint False type eval | step 250 | loss 3.1139 | checkpoint True type eval | step 500 | loss 2.5083 | checkpoint True type eval | step 750 | loss 2.3826 | checkpoint True type eval | step 1000 | loss 2.2400 | checkpoint True type eval | step 1250 | loss 2.0680 | checkpoint True type eval | step 1500 | loss 1.9792 | checkpoint True type eval | step 1750 | loss 1.9354 | checkpoint True type eval | step 2000 | loss 1.8793 | checkpoint True type eval | step 2250 | loss 1.8414 | checkpoint True type eval | step 2500 | loss 1.8133 | checkpoint True type eval | step 2750 | loss 1.7787 | checkpoint True type eval | step 3000 | loss 1.7515 | checkpoint True type eval | step 3250 | loss 1.7513 | checkpoint True type eval | step 3500 | loss 1.7423 | checkpoint True type eval | step 3750 | loss 1.7245 | checkpoint True type eval | step 4000 | loss 1.7110 | checkpoint True type eval | step 4250 | loss 1.7084 | checkpoint True type eval | step 4500 | loss 1.6869 | checkpoint True type eval | step 4750 | loss 1.6772 | checkpoint True type eval | step 5000 | loss 1.6680 | checkpoint True type eval | step 5250 | loss 1.6416 | checkpoint True type eval | step 5500 | loss 1.6447 | checkpoint False type eval | step 5750 | loss 1.6367 | checkpoint True type eval | step 6000 | loss 1.6424 | checkpoint False type eval | step 6250 | loss 1.6370 | checkpoint False type eval | step 6500 | loss 1.6343 | checkpoint True type eval | step 6750 | loss 1.6303 | checkpoint True type eval | step 7000 | loss 1.6229 | checkpoint True type eval | step 7250 | loss 1.6121 | checkpoint True type eval | step 7500 | loss 1.6108 | checkpoint True type eval | step 7750 | loss 1.6004 | checkpoint True type eval | step 8000 | loss 1.5971 | checkpoint True type eval | step 8250 | loss 1.5931 | checkpoint True type eval | step 8500 | loss 1.5917 | checkpoint True type eval | step 8750 | loss 1.5924 | checkpoint False type eval | step 9000 | loss 1.5881 | checkpoint True type eval | step 9250 | loss 1.5856 | checkpoint True type eval | step 9500 | loss 1.5845 | checkpoint True type eval | step 9750 | loss 1.5839 | checkpoint True type eval | step 10000 | loss 1.5833 | checkpoint True type eval | step 0 | loss 4.8544 | checkpoint False type eval | step 250 | loss 3.0956 | checkpoint True type eval | step 500 | loss 2.4943 | checkpoint True type eval | step 750 | loss 2.3488 | checkpoint True type eval | step 1000 | loss 2.1962 | checkpoint True type eval | step 1250 | loss 2.0515 | checkpoint True type eval | step 1500 | loss 1.9533 | checkpoint True type eval | step 1750 | loss 1.9070 | checkpoint True type eval | step 2000 | loss 1.8539 | checkpoint True type eval | step 2250 | loss 1.8142 | checkpoint True type eval | step 2500 | loss 1.7810 | checkpoint True type eval | step 2750 | loss 1.7395 | checkpoint True type eval | step 3000 | loss 1.7162 | checkpoint True type eval | step 3250 | loss 1.7088 | checkpoint True type eval | step 3500 | loss 1.6995 | checkpoint True type eval | step 3750 | loss 1.6880 | checkpoint True type eval | step 4000 | loss 1.6766 | checkpoint True type eval | step 4250 | loss 1.6753 | checkpoint True type eval | step 4500 | loss 1.6584 | checkpoint True type eval | step 4750 | loss 1.6419 | checkpoint True type eval | step 5000 | loss 1.6300 | checkpoint True type eval | step 5250 | loss 1.6117 | checkpoint True type eval | step 5500 | loss 1.6086 | checkpoint True type eval | step 5750 | loss 1.6152 | checkpoint False type eval | step 6000 | loss 1.6096 | checkpoint False type eval | step 6250 | loss 1.6190 | checkpoint False type eval | step 6500 | loss 1.6118 | checkpoint False type eval | step 6750 | loss 1.6055 | checkpoint True type eval | step 7000 | loss 1.6066 | checkpoint False type eval | step 7250 | loss 1.6080 | checkpoint False type eval | step 7500 | loss 1.6004 | checkpoint True type eval | step 7750 | loss 1.5821 | checkpoint True type eval | step 8000 | loss 1.5792 | checkpoint True type eval | step 8250 | loss 1.5776 | checkpoint True type eval | step 8500 | loss 1.5804 | checkpoint False type eval | step 8750 | loss 1.5805 | checkpoint False type eval | step 9000 | loss 1.5763 | checkpoint True type eval | step 9250 | loss 1.5718 | checkpoint True type eval | step 9500 | loss 1.5695 | checkpoint True type eval | step 9750 | loss 1.5686 | checkpoint True type eval | step 10000 | loss 1.5683 | checkpoint True