Update README.md

update sampling parameters to match evals

Add mlp_only_layers, decoder_sparse_step, norm_topk_prob; set sliding_window=0 for vLLM 0.21.0 native Laguna

Bump max_position_embeddings to 256K; double yarn rope factor to 64

Update README.md

Update overlay image with H100 fp8_w8a8 fused-MoE tuned config (E=256,N=512)

Upload chat_template.jinja

Upload image/laguna-vllm-v0.19.0-overlay.tar.gz with huggingface_hub

Upload image/LOAD.md with huggingface_hub

Add files using upload-large-folder tool

Push corrected vLLM image (rope_parameters guard + parser registrations)

Add files using upload-large-folder tool

Add files using upload-large-folder tool

initial commit