flashinfer.mla.supported_sparse_mla_sm120_configs¶
- flashinfer.mla.supported_sparse_mla_sm120_configs(*, kv_cache_format: str = 'fp8') dict[str, SparseMLASm120DecodeConfig]¶
Enumerate the instantiated SM120 sparse-MLA decode kernel configurations.
Lets callers validate a serving configuration at initialization time instead of discovering an uninstantiated
(num_heads, topk)pair on the first decode-form request.- Parameters:
kv_cache_format ({"fp8", "nvfp4"}) – Storage format whose independently calibrated kernel envelope is requested. Defaults to
"fp8"for backward compatibility.- Returns:
Mapping from kernel family to its instantiated decode set. FP8 returns DSv4, DSv3.2, GLM-NSA, GLM53_NOPE, and DOTS3_SWA entries. NVFP4 currently returns the independently calibrated DSv4 entry.
- Return type:
dict[str, SparseMLASm120DecodeConfig]
Examples
>>> import flashinfer >>> configs = flashinfer.mla.supported_sparse_mla_sm120_configs() >>> configs["dsv4"].supports_decode(num_heads=64, topk=256) True >>> nvfp4 = flashinfer.mla.supported_sparse_mla_sm120_configs( ... kv_cache_format="nvfp4" ... ) >>> nvfp4["dsv4"].bytes_per_token 384