Stage 1 checkpoint: [Step 50/20000] loss=7.7500

Stage 1 complete: 10K steps, CKA layer alignment final checkpoint

Stage 1 checkpoint: [Step 9100/10000] loss=0.0141 cka_mean=0.1489

Stage 1 checkpoint: [Step 8400/10000] loss=0.0131 cka_mean=0.1052

Stage 1 checkpoint: [Step 7100/10000] loss=0.0161 cka_mean=0.1370

Stage 1 checkpoint: [Step 6400/10000] loss=0.0199 cka_mean=0.3752

Stage 1 checkpoint: [Step 5050/10000] loss=0.0500 cka_mean=0.4527

Stage 1 checkpoint: [Step 4400/10000] loss=0.6573 cka_mean=0.3674

Stage 1 checkpoint: [Step 3100/10000] loss=0.0432 cka_mean=0.4536

Stage 1 checkpoint: [Step 2450/10000] loss=0.0617 cka_mean=0.4075

Stage 1 checkpoint: [Step 1800/10000] loss=0.1092 cka_mean=0.4199

Stage 1: step 1000+, loss dropping, cka=0.60

Stage 1 initial: step 1000, loss=0.29, cka=0.60