pub const KERNEL_KV_LORA_DIM: usize = 512;Expand description
2026-09-25: #define GLM_KV_LORA_DIM in
kernels/gb10/glm-5.3-flash/nvfp4/glm5next_dsa_mla_decode.cu, the latent width the
decode kernel is compiled for. Glm5NextDsaConfig::validate and attend::decode_attention
refuse a kv_lora_rank that differs.