Training update: run 243, 500 iters, cache_key=ci

Training update: run 242, 500 iters, cache_key=ci

Training update: run 241, 500 iters, cache_key=ci

Training update: run 240, 500 iters, cache_key=ci

Training update: run 239, 500 iters, cache_key=ci

Training update: run 238, 500 iters, cache_key=ci

Training update: run 237, 500 iters, cache_key=ci

Training update: run 236, 500 iters, cache_key=ci

Training update: run 235, 500 iters, cache_key=ci

Training update: run 234, 500 iters, cache_key=ci

Training update: run 233, 500 iters, cache_key=ci

Training update: run 232, 500 iters, cache_key=ci

Training update: run 231, 500 iters, cache_key=ci

Training update: run 230, 500 iters, cache_key=ci

Training update: run 229, 500 iters, cache_key=ci

Training update: run 228, 500 iters, cache_key=ci

Training update: run 227, 500 iters, cache_key=ci

Training update: run 226, 500 iters, cache_key=ci

Training update: run 225, 500 iters, cache_key=ci

Training update: run 224, 500 iters, cache_key=ci

Training update: run 223, 500 iters, cache_key=ci

Training update: run 222, 500 iters, cache_key=ci

Training update: run 221, 500 iters, cache_key=ci

Deploy EchoSelf NanEcho model (workflow run 173)

Deploy EchoSelf NanEcho model (workflow run 163)

Fix Conv1D weight transposition for HF GPT-2 compatibility

Deploy NanEcho CI checkpoint (4L/4H/256E, 200 iters, val_loss=1.9258)

initial commit