Upload n6ja1e77q4yn1viilgzbv32a.zip

Upload folder using huggingface_hub

Upload distill_sharded.py with huggingface_hub

Upload eval_kl.py with huggingface_hub

add phase-2 ultra-conservative sweep (J,K,L,M) + waiter that auto-launches after phase 1 from the best ckpt

add 9-config hparam sweep + new_layer_lr_mul param-groups support

fix scheduler bug: don't prepare scheduler with accelerate (was over-stepping cosine by num_processes); add grow40_winning_v2 config

grow40_winning: switch student to bf16 to fit in B200 memory + 40-layer Adam state

add retry loop around load_dataset for transient HF Hub 5xx

add micro_batch_size config key + per-micro inner loop in train step (fixes OOM for fp32+seq2048)

fix OOM: chunked KL with checkpointing + PYTORCH_CUDA_ALLOC_CONF expandable_segments; add kl_chunk_size config key

add grow_layers, sweep configs (replicate_zero4, grow40_winning, grow40_simple), sweep runner

initial scaffold: distill.py + base/zero_14_17 configs + accelerate yaml

initial commit