{ "cls_name": "RBLNQwen3ForCausalLMConfig", "dtype": "float16", "optimum_rbln_version": "0.10.0.post2", "_compile_cfgs": [ { "compiled_model_name": "prefill", "input_info": [ [ "input_ids", [ 1, 128 ], "int64" ], [ "cache_position", [ 1, 128 ], "int32" ], [ "block_tables", [ 2 ], "int16" ], [ "query_position", [], "int16" ], [ "past_key_values_0", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_1", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_2", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_3", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_4", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_5", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_6", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_7", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_8", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_9", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_10", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_11", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_12", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_13", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_14", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_15", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_16", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_17", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_18", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_19", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_20", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_21", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_22", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_23", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_24", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_25", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_26", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_27", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_28", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_29", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_30", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_31", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_32", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_33", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_34", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_35", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_36", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_37", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_38", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_39", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_40", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_41", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_42", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_43", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_44", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_45", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_46", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_47", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_48", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_49", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_50", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_51", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_52", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_53", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_54", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_55", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_56", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_57", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_58", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_59", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_60", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_61", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_62", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_63", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_64", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_65", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_66", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_67", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_68", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_69", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_70", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_71", [ 1, 8, 16384, 128 ], "float16" ] ], "npu": "RBLN-CA25", "tensor_parallel_size": 4 }, { "compiled_model_name": "decoder_batch_12", "input_info": [ [ "input_ids", [ 12, 1 ], "int64" ], [ "cache_position", [ 12, 1 ], "int32" ], [ "block_tables", [ 12, 2 ], "int16" ], [ "past_key_values_0", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_1", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_2", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_3", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_4", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_5", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_6", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_7", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_8", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_9", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_10", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_11", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_12", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_13", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_14", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_15", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_16", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_17", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_18", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_19", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_20", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_21", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_22", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_23", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_24", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_25", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_26", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_27", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_28", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_29", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_30", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_31", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_32", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_33", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_34", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_35", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_36", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_37", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_38", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_39", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_40", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_41", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_42", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_43", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_44", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_45", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_46", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_47", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_48", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_49", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_50", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_51", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_52", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_53", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_54", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_55", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_56", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_57", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_58", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_59", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_60", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_61", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_62", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_63", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_64", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_65", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_66", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_67", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_68", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_69", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_70", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_71", [ 1, 8, 16384, 128 ], "float16" ] ], "npu": "RBLN-CA25", "tensor_parallel_size": 4 }, { "compiled_model_name": "decoder_batch_8", "input_info": [ [ "input_ids", [ 8, 1 ], "int64" ], [ "cache_position", [ 8, 1 ], "int32" ], [ "block_tables", [ 8, 2 ], "int16" ], [ "past_key_values_0", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_1", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_2", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_3", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_4", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_5", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_6", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_7", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_8", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_9", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_10", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_11", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_12", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_13", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_14", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_15", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_16", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_17", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_18", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_19", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_20", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_21", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_22", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_23", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_24", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_25", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_26", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_27", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_28", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_29", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_30", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_31", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_32", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_33", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_34", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_35", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_36", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_37", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_38", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_39", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_40", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_41", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_42", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_43", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_44", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_45", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_46", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_47", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_48", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_49", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_50", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_51", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_52", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_53", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_54", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_55", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_56", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_57", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_58", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_59", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_60", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_61", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_62", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_63", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_64", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_65", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_66", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_67", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_68", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_69", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_70", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_71", [ 1, 8, 16384, 128 ], "float16" ] ], "npu": "RBLN-CA25", "tensor_parallel_size": 4 }, { "compiled_model_name": "decoder_batch_4", "input_info": [ [ "input_ids", [ 4, 1 ], "int64" ], [ "cache_position", [ 4, 1 ], "int32" ], [ "block_tables", [ 4, 2 ], "int16" ], [ "past_key_values_0", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_1", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_2", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_3", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_4", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_5", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_6", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_7", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_8", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_9", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_10", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_11", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_12", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_13", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_14", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_15", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_16", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_17", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_18", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_19", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_20", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_21", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_22", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_23", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_24", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_25", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_26", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_27", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_28", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_29", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_30", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_31", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_32", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_33", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_34", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_35", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_36", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_37", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_38", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_39", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_40", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_41", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_42", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_43", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_44", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_45", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_46", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_47", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_48", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_49", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_50", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_51", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_52", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_53", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_54", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_55", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_56", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_57", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_58", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_59", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_60", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_61", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_62", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_63", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_64", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_65", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_66", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_67", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_68", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_69", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_70", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_71", [ 1, 8, 16384, 128 ], "float16" ] ], "npu": "RBLN-CA25", "tensor_parallel_size": 4 }, { "compiled_model_name": "decoder_batch_1", "input_info": [ [ "input_ids", [ 1, 1 ], "int64" ], [ "cache_position", [ 1, 1 ], "int32" ], [ "block_tables", [ 1, 2 ], "int16" ], [ "past_key_values_0", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_1", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_2", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_3", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_4", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_5", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_6", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_7", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_8", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_9", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_10", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_11", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_12", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_13", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_14", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_15", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_16", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_17", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_18", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_19", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_20", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_21", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_22", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_23", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_24", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_25", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_26", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_27", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_28", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_29", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_30", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_31", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_32", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_33", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_34", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_35", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_36", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_37", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_38", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_39", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_40", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_41", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_42", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_43", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_44", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_45", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_46", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_47", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_48", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_49", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_50", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_51", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_52", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_53", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_54", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_55", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_56", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_57", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_58", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_59", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_60", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_61", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_62", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_63", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_64", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_65", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_66", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_67", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_68", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_69", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_70", [ 1, 8, 16384, 128 ], "float16" ], [ "past_key_values_71", [ 1, 8, 16384, 128 ], "float16" ] ], "npu": "RBLN-CA25", "tensor_parallel_size": 4 } ], "batch_size": 12, "max_seq_len": 32768, "use_inputs_embeds": false, "use_position_ids": false, "use_attention_mask": false, "quantization": { "format": "rbln", "weights": "int8", "activations": "fp16", "kv_caches": "fp16" }, "lora_config": null, "attn_impl": "flash_attn", "kvcache_partition_len": 16384, "kvcache_block_size": 16384, "prefill_chunk_size": 128, "kvcache_num_blocks": 24, "cache_impl": "static", "sliding_window": null, "sliding_window_layers": [], "phases": [ "prefill", "decode" ], "logits_to_keep": 1, "output_hidden_states": false, "decoder_batch_sizes": [ 12, 8, 4, 1 ], "kvcache_metas": [ { "name": "past_key_values_0", "layer_index": 0, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_1", "layer_index": 0, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_2", "layer_index": 1, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_3", "layer_index": 1, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_4", "layer_index": 2, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_5", "layer_index": 2, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_6", "layer_index": 3, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_7", "layer_index": 3, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_8", "layer_index": 4, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_9", "layer_index": 4, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_10", "layer_index": 5, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_11", "layer_index": 5, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_12", "layer_index": 6, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_13", "layer_index": 6, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_14", "layer_index": 7, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_15", "layer_index": 7, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_16", "layer_index": 8, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_17", "layer_index": 8, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_18", "layer_index": 9, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_19", "layer_index": 9, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_20", "layer_index": 10, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_21", "layer_index": 10, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_22", "layer_index": 11, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_23", "layer_index": 11, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_24", "layer_index": 12, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_25", "layer_index": 12, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_26", "layer_index": 13, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_27", "layer_index": 13, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_28", "layer_index": 14, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_29", "layer_index": 14, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_30", "layer_index": 15, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_31", "layer_index": 15, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_32", "layer_index": 16, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_33", "layer_index": 16, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_34", "layer_index": 17, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_35", "layer_index": 17, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_36", "layer_index": 18, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_37", "layer_index": 18, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_38", "layer_index": 19, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_39", "layer_index": 19, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_40", "layer_index": 20, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_41", "layer_index": 20, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_42", "layer_index": 21, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_43", "layer_index": 21, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_44", "layer_index": 22, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_45", "layer_index": 22, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_46", "layer_index": 23, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_47", "layer_index": 23, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_48", "layer_index": 24, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_49", "layer_index": 24, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_50", "layer_index": 25, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_51", "layer_index": 25, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_52", "layer_index": 26, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_53", "layer_index": 26, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_54", "layer_index": 27, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_55", "layer_index": 27, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_56", "layer_index": 28, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_57", "layer_index": 28, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_58", "layer_index": 29, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_59", "layer_index": 29, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_60", "layer_index": 30, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_61", "layer_index": 30, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_62", "layer_index": 31, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_63", "layer_index": 31, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_64", "layer_index": 32, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_65", "layer_index": 32, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_66", "layer_index": 33, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_67", "layer_index": 33, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_68", "layer_index": 34, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_69", "layer_index": 34, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_70", "layer_index": 35, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" }, { "name": "past_key_values_71", "layer_index": 35, "shape": [ 24, 8, 16384, 128 ], "layer_type": "full_attention", "is_auto": true, "dtype": "float16" } ] }