索引 A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | Z A activation (torchrec.modules.embedding_configs.QuantConfig 屬性) ADAGRAD (torchrec.distributed.embedding_types.OptimType 屬性) ADAM (torchrec.distributed.embedding_types.OptimType 屬性) ADAMW (torchrec.distributed.embedding_types.OptimType 屬性) add_param_group() (torchrec.optim.keyed.KeyedOptimizer 方法) (torchrec.optim.keyed.OptimizerWrapper 方法) add_params_from_parameter_sharding() (在 torchrec.distributed.utils 模組中) add_prefix_to_state_dict() (在 torchrec.distributed.utils 模組中) algorithm (torchrec.distributed.types.CacheParams 屬性), [1] All2All_Pooled_Req (torchrec.distributed.comm_ops 中的類別) all2all_pooled_sync() (在 torchrec.distributed.comm_ops 模組中) All2All_Pooled_Wait (torchrec.distributed.comm_ops 中的類別) All2All_Seq_Req (torchrec.distributed.comm_ops 中的類別) All2All_Seq_Req_Wait (torchrec.distributed.comm_ops 中的類別) all2all_sequence_sync() (在 torchrec.distributed.comm_ops 模組中) All2AllDenseInfo (torchrec.distributed.comm_ops 中的類別) All2AllPooledInfo (torchrec.distributed.comm_ops 中的類別) All2AllSequenceInfo (torchrec.distributed.comm_ops 中的類別) All2Allv_Req (torchrec.distributed.comm_ops 中的類別) all2allv_sync() (在 torchrec.distributed.comm_ops 模組中) All2Allv_Wait (torchrec.distributed.comm_ops 中的類別) All2AllVInfo (torchrec.distributed.comm_ops 中的類別) all_gather_base_pooled() (在 torchrec.distributed.comm_ops 模組中) all_gather_base_sync() (在 torchrec.distributed.comm_ops 模組中) all_gather_into_tensor_backward() (在 torchrec.distributed.comm_ops 模組中) all_gather_into_tensor_fake() (在 torchrec.distributed.comm_ops 模組中) all_gather_into_tensor_setup_context() (在 torchrec.distributed.comm_ops 模組中) all_to_all_single_backward() (在 torchrec.distributed.comm_ops 模組中) all_to_all_single_fake() (在 torchrec.distributed.comm_ops 模組中) all_to_all_single_setup_context() (在 torchrec.distributed.comm_ops 模組中) AllGatherBase_Req (torchrec.distributed.comm_ops 中的類別) AllGatherBase_Wait (torchrec.distributed.comm_ops 中的類別) AllGatherBaseInfo (torchrec.distributed.comm_ops 中的類別) allocate_budget() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 靜態方法) alltoall_pooled() (在 torchrec.distributed.comm_ops 模組中) alltoall_sequence() (在 torchrec.distributed.comm_ops 模組中) alltoallv() (在 torchrec.distributed.comm_ops 模組中) append_prefix() (在 torchrec.distributed.utils 模組中) apply() (torchrec.modules.lazy_extension.LazyModuleExtensionMixin 方法) apply_feature_processor() (torchrec.distributed.quant_embeddingbag.ShardedQuantFeatureProcessedEmbeddingBagCollection 方法) apply_mc_method_to_jt_dict() (在 torchrec.modules.mc_modules 模組中) average_threshold_filter() (在 torchrec.modules.mc_modules 模組中) Awaitable (torchrec.distributed.types 中的類別) B B_global (torchrec.distributed.comm_ops.All2AllVInfo 屬性), [1] B_local (torchrec.distributed.comm_ops.All2AllVInfo 屬性), [1] B_local_list (torchrec.distributed.comm_ops.All2AllVInfo 屬性), [1] backward (torchrec.distributed.types.QuantizedCommCodecs 屬性) backward() (torchrec.distributed.comm_ops.All2All_Pooled_Req 靜態方法) (torchrec.distributed.comm_ops.All2All_Pooled_Wait 靜態方法) (torchrec.distributed.comm_ops.All2All_Seq_Req 靜態方法) (torchrec.distributed.comm_ops.All2All_Seq_Req_Wait 靜態方法) (torchrec.distributed.comm_ops.All2Allv_Req 靜態方法) (torchrec.distributed.comm_ops.All2Allv_Wait 靜態方法) (torchrec.distributed.comm_ops.AllGatherBase_Req 靜態方法) (torchrec.distributed.comm_ops.AllGatherBase_Wait 靜態方法) (torchrec.distributed.comm_ops.ReduceScatter_Req 靜態方法) (torchrec.distributed.comm_ops.ReduceScatter_Wait 靜態方法) (torchrec.distributed.comm_ops.ReduceScatterBase_Req 靜態方法) (torchrec.distributed.comm_ops.ReduceScatterBase_Wait 靜態方法) (torchrec.distributed.comm_ops.ReduceScatterV_Req 靜態方法) (torchrec.distributed.comm_ops.ReduceScatterV_Wait 靜態方法) (torchrec.distributed.comm_ops.Variable_Batch_All2All_Pooled_Req 靜態方法) (torchrec.distributed.comm_ops.Variable_Batch_All2All_Pooled_Wait 靜態方法) (torchrec.distributed.embedding_lookup.CommOpGradientScaling 靜態方法) backward_recat_tensor (torchrec.distributed.comm_ops.All2AllSequenceInfo 屬性), [1] bare_named_parameters() (torchrec.distributed.model_parallel.DistributedModelParallel 方法) BaseCwEmbeddingSharding (torchrec.distributed.sharding.cw_sharding 中的類別) BaseDpEmbeddingSharding (torchrec.distributed.sharding.dp_sharding 中的類別) BaseEmbeddingConfig (torchrec.modules.embedding_configs 中的類別) BaseEmbeddingDist (torchrec.distributed.embedding_sharding 中的類別) BaseEmbeddingLookup (torchrec.distributed.embedding_types 中的類別) BaseEmbeddingSharder (torchrec.distributed.embedding_types 中的類別) BaseFeatureProcessor (torchrec.modules.feature_processor 中的類別) BaseGroupedFeatureProcessor (torchrec.distributed.embedding_types 中的類別) (torchrec.modules.feature_processor 中的類別) BaseManagedCollisionEmbeddingCollection (torchrec.modules.mc_embedding_modules 中的類別) BaseQuantEmbeddingSharder (torchrec.distributed.embedding_types 中的類別) BaseRwEmbeddingSharding (torchrec.distributed.sharding.rw_sharding 中的類別) BaseSparseFeaturesDist (torchrec.distributed.embedding_sharding 中的類別) BaseTwEmbeddingSharding (torchrec.distributed.sharding.tw_sharding 中的類別) BaseTwRwEmbeddingSharding (torchrec.distributed.sharding.twrw_sharding 中的類別) batch_size (torchrec.distributed.comm_ops.All2AllDenseInfo 屬性) (torchrec.distributed.planner.types.ShardingOption 屬性) batch_size_per_feature_pre_a2a (torchrec.distributed.comm_ops.VariableBatchAll2AllPooledInfo 屬性), [1] batch_size_per_rank (torchrec.distributed.comm_ops.All2AllPooledInfo 屬性), [1] batch_size_per_rank_per_feature (torchrec.distributed.comm_ops.VariableBatchAll2AllPooledInfo 屬性), [1] batch_sizes (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] batching_metadata() (torchrec.inference.modules.PredictFactory 方法) batching_metadata_json() (torchrec.inference.modules.PredictFactory 方法) BatchingMetadata (torchrec.inference.modules 中的類別) best() (torchrec.distributed.planner.utils.LuusJaakolaSearch 方法) BinarySearchPredicate (torchrec.distributed.planner.utils 中的類別) bounds_check_mode (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.types.ParameterSharding 屬性) bucket_mapping_tensor (torchrec.distributed.embedding_types.InputDistOutputs 屬性) bucketize_kjt_before_all2all() (在 torchrec.distributed.embedding_sharding 模組中) bucketize_kjt_inference() (在 torchrec.distributed.embedding_sharding 模組中) bucketized_length (torchrec.distributed.embedding_types.InputDistOutputs 屬性) build_affine_storage_model() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 靜態方法) bwd_comms (torchrec.distributed.planner.types.Perf 屬性) bwd_compute (torchrec.distributed.planner.types.Perf 屬性) bwd_compute_multiplier (torchrec.distributed.planner.types.Topology 屬性) bytes_to_gb() (在 torchrec.distributed.planner.utils 模組中) bytes_to_mb() (在 torchrec.distributed.planner.utils 模組中) C cache_load_factor (torchrec.distributed.planner.types.ShardingOption 屬性) cache_params (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.types.ParameterSharding 屬性) cacheability (torchrec.distributed.planner.shard_estimators.EmbeddingOffloadStats 屬性) (torchrec.distributed.types.CacheStatistics 屬性) CacheParams (torchrec.distributed.types 中的類別) CacheStatistics (torchrec.distributed.types 中的類別) calc_quantized_size() (torchrec.distributed.types.NoOpQuantizedCommCodec 方法) (torchrec.distributed.types.QuantizedCommCodec 方法) calculate_pipeline_io_cost() (在 torchrec.distributed.planner.shard_estimators 模組中) calculate_shard_storages() (在 torchrec.distributed.planner.shard_estimators 模組中) callbacks (torchrec.distributed.dist_data.PooledEmbeddingsAllToAll 屬性), [1] (torchrec.distributed.dist_data.PooledEmbeddingsAwaitable 屬性), [1] (torchrec.distributed.dist_data.VariableBatchPooledEmbeddingsAllToAll 屬性), [1] (torchrec.distributed.types.Awaitable 屬性) check_module_output_dimension() (在 torchrec.modules.utils 模組中) clamp() (torchrec.distributed.planner.utils.LuusJaakolaSearch 方法) clf_to_bytes() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 靜態方法) coalesce_history_metadata() (torchrec.modules.mc_modules.DistanceLFU_EvictionPolicy 方法) (torchrec.modules.mc_modules.LFU_EvictionPolicy 方法) （torchrec.modules.mc_modules.LRU_EvictionPolicy 方法） （torchrec.modules.mc_modules.MCHEvictionPolicy 方法） codecs（torchrec.distributed.comm_ops.All2AllPooledInfo 屬性），[1] （torchrec.distributed.comm_ops.All2AllSequenceInfo 屬性），[1] （torchrec.distributed.comm_ops.All2AllVInfo 屬性） （torchrec.distributed.comm_ops.AllGatherBaseInfo 屬性） （torchrec.distributed.comm_ops.ReduceScatterBaseInfo 屬性） （torchrec.distributed.comm_ops.ReduceScatterInfo 屬性） （torchrec.distributed.comm_ops.ReduceScatterVInfo 屬性），[1] （torchrec.distributed.comm_ops.VariableBatchAll2AllPooledInfo 屬性），[1] collective_plan()（torchrec.distributed.planner.planners.EmbeddingShardingPlanner 方法） （torchrec.distributed.planner.planners.HeteroEmbeddingShardingPlanner 方法） （torchrec.distributed.types.ShardingPlanner 方法） COLUMN_WISE（torchrec.distributed.types.ShardingType 屬性） （torchrec.modules.embedding_configs.ShardingType 屬性） CombinedOptimizer（torchrec.optim.keyed 中的類別） CommOp（torchrec.distributed.types 中的類別） CommOpGradientScaling（torchrec.distributed.embedding_lookup 中的類別） compute()（torchrec.distributed.embedding.ShardedEmbeddingCollection 方法） （torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法） （torchrec.distributed.embeddingbag.ShardedEmbeddingBagCollection 方法） （torchrec.distributed.mc_modules.ShardedManagedCollisionCollection 方法） （torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法） （torchrec.distributed.quant_embeddingbag.ShardedQuantFeatureProcessedEmbeddingBagCollection 方法） （torchrec.distributed.types.ShardedModule 方法） compute_and_output_dist()（torchrec.distributed.embedding.ShardedEmbeddingCollection 方法） （torchrec.distributed.embeddingbag.ShardedEmbeddingBagCollection 方法） （torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法） （torchrec.distributed.types.ShardedModule 方法） compute_device（torchrec.distributed.planner.types.Topology 屬性） compute_kernel（torchrec.distributed.embedding_types.EmbeddingAttributes 屬性） （torchrec.distributed.embedding_types.GroupedEmbeddingConfig 屬性） (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.types.ParameterSharding 屬性) compute_kernel_to_embedding_location()（位於模組 torchrec.distributed.embedding_types 中） compute_kernels（torchrec.distributed.planner.types.ParameterConstraints 屬性），[1] compute_kernels()（torchrec.distributed.embedding_types.BaseEmbeddingSharder 方法） （torchrec.distributed.embedding_types.BaseQuantEmbeddingSharder 方法） （torchrec.distributed.quant_embeddingbag.QuantFeatureProcessedEmbeddingBagCollectionSharder 方法） （torchrec.distributed.types.ModuleSharder 方法） ComputeJTDictToKJT（torchrec.sparse.jagged_tensor 中的類別） ComputeKernel（torchrec.distributed.types 中的類別） ComputeKJTToJTDict（torchrec.sparse.jagged_tensor 中的類別） concat()（torchrec.sparse.jagged_tensor.KeyedJaggedTensor 靜態方法） CONSTANT（torchrec.optim.warmup.WarmupPolicy 屬性） construct_jagged_tensors()（位於模組 torchrec.modules.utils 中） construct_jagged_tensors_inference()（位於模組 torchrec.modules.utils 中） construct_modulelist_from_single_module()（位於模組 torchrec.modules.utils 中） construct_output_kt()（位於模組 torchrec.distributed.embeddingbag 中） convert_list_of_modules_to_modulelist()（位於模組 torchrec.modules.utils 中） convert_to_fbgemm_types()（位於模組 torchrec.distributed.utils 中） copy()（torchrec.distributed.model_parallel.DistributedModelParallel 方法） （torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法） （torchrec.distributed.utils.CopyableMixin 方法） copy_to_device()（位於模組 torchrec.distributed.utils 中） CopyableMixin（torchrec.distributed.utils 中的類別） COSINE_ANNEALING_WARM_RESTARTS（torchrec.optim.warmup.WarmupPolicy 屬性） create_arg()（torchrec.fx.tracer.Tracer 方法） create_context()（torchrec.distributed.embedding.ShardedEmbeddingCollection 方法） （torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法） （torchrec.distributed.embeddingbag.ShardedEmbeddingBagCollection 方法） （torchrec.distributed.mc_embedding.ShardedManagedCollisionEmbeddingCollection 方法） （torchrec.distributed.mc_embeddingbag.ShardedManagedCollisionEmbeddingBagCollection 方法） （torchrec.distributed.mc_modules.ShardedManagedCollisionCollection 方法） （torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法） （torchrec.distributed.types.NoOpQuantizedCommCodec 方法） (torchrec.distributed.types.QuantizedCommCodec 方法) （torchrec.distributed.types.ShardedModule 方法） create_embedding_bag_sharding()（位於模組 torchrec.distributed.embeddingbag 中） create_embedding_sharding()（位於模組 torchrec.distributed.embedding 中） create_infer_embedding_bag_sharding()（位於模組 torchrec.distributed.quant_embeddingbag 中） create_input_dist()（torchrec.distributed.embedding_sharding.EmbeddingSharding 方法） （torchrec.distributed.sharding.cw_sharding.CwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.cw_sharding.InferCwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.dp_sharding.DpPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.InferRwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.RwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.InferTwEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.TwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.twrw_sharding.TwRwPooledEmbeddingSharding 方法） create_lookup()（torchrec.distributed.embedding_sharding.EmbeddingSharding 方法） （torchrec.distributed.sharding.cw_sharding.CwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.cw_sharding.InferCwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.dp_sharding.DpPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.InferRwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.RwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.InferTwEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.TwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.twrw_sharding.TwRwPooledEmbeddingSharding 方法） create_mc_sharding()（位於模組 torchrec.distributed.mc_modules 中） create_output_dist()（torchrec.distributed.embedding_sharding.EmbeddingSharding 方法） （torchrec.distributed.sharding.cw_sharding.CwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.cw_sharding.InferCwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.dp_sharding.DpPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.InferRwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.RwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.InferTwEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.TwPooledEmbeddingSharding 方法） （torchrec.distributed.sharding.twrw_sharding.TwRwPooledEmbeddingSharding 方法） create_predict_module()（torchrec.inference.modules.PredictFactory 方法） create_sharding_infos_by_sharding()（位於模組 torchrec.distributed.embedding 中） （位於模組 torchrec.distributed.embeddingbag 中） create_sharding_infos_by_sharding_device_group()（位於模組 torchrec.distributed.embedding 中） （位於模組 torchrec.distributed.embeddingbag 中） CrossNet（torchrec.modules.crossnet 中的類別） cumsum_dim_sum_per_rank_tensor（torchrec.distributed.comm_ops.All2AllPooledInfo 屬性），[1] CustomTopologyData（torchrec.distributed.planner.types 中的類別） CwPooledEmbeddingSharding（torchrec.distributed.sharding.cw_sharding 中的類別） D D_local_list（torchrec.distributed.comm_ops.All2AllVInfo 屬性），[1] DATA_PARALLEL（torchrec.distributed.types.ShardingType 屬性） （torchrec.modules.embedding_configs.ShardingType 屬性） data_type（torchrec.distributed.embedding_types.GroupedEmbeddingConfig 屬性） （torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性） data_type_to_dtype()（位於模組 torchrec.modules.embedding_configs 中） data_type_to_sparse_type()（位於模組 torchrec.modules.embedding_configs 中） DataParallelWrapper（torchrec.distributed.model_parallel 中的類別） ddr（torchrec.distributed.planner.types.Storage 屬性） DDR（torchrec.distributed.types.ParameterStorage 屬性） ddr_mem_bw（torchrec.distributed.planner.types.Topology 屬性） decay_iters（torchrec.optim.warmup.WarmupStage 屬性） decode()（torchrec.distributed.types.NoOpQuantizedCommCodec 方法） (torchrec.distributed.types.QuantizedCommCodec 方法) DeepFM（torchrec.modules.deepfm 中的類別） DEFAULT（torchrec.distributed.types.ComputeKernel 屬性） DefaultDataParallelWrapper（torchrec.distributed.model_parallel 中的類別） DENSE（torchrec.distributed.embedding_types.EmbeddingComputeKernel 屬性） DenseArch（torchrec.models.deepfm 中的類別） dependency（torchrec.distributed.planner.types.ShardingOption 屬性） deterministic_dedup()（位於模組 torchrec.modules.utils 中） device（torchrec.distributed.embedding_sharding.KJTSplitsAllToAllMeta 屬性） （torchrec.distributed.planner.partitioners.OrderedDeviceHardware 屬性） DEVICE（torchrec.distributed.planner.types.PartitionByType 屬性） device（torchrec.inference.modules.BatchingMetadata 屬性） （torchrec.modules.embedding_modules.EmbeddingBagCollection 屬性） （torchrec.modules.embedding_modules.EmbeddingCollection 屬性） （torchrec.modules.mc_modules.ManagedCollisionModule 屬性） （torchrec.quant.embedding_modules.EmbeddingBagCollection 屬性） （torchrec.quant.embedding_modules.EmbeddingCollection 屬性） device()（torchrec.sparse.jagged_tensor.JaggedTensor 方法） （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） （torchrec.sparse.jagged_tensor.KeyedTensor 方法） device_group（torchrec.distributed.planner.types.ParameterConstraints 屬性），[1] DeviceHardware（torchrec.distributed.planner.types 中的類別） devices（torchrec.distributed.planner.types.Topology 屬性） dim_sum()（torchrec.distributed.embedding_types.GroupedEmbeddingConfig 方法） dim_sum_per_rank（torchrec.distributed.comm_ops.All2AllPooledInfo 屬性），[1] （torchrec.distributed.comm_ops.All2AllVInfo 屬性） dim_sum_per_rank_tensor（torchrec.distributed.comm_ops.All2AllPooledInfo 屬性），[1] dims_sum_per_rank（torchrec.distributed.comm_ops.All2AllVInfo 屬性） dist_init()（torchrec.sparse.jagged_tensor.KeyedJaggedTensor 靜態方法） dist_labels()（torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） dist_splits()（torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） dist_tensors()（torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） DistanceLFU_EvictionPolicy（torchrec.modules.mc_modules 中的類別） DistributedModelParallel（torchrec.distributed.model_parallel 中的類別） divisor（torchrec.distributed.embeddingbag.EmbeddingBagCollectionContext 屬性） DpPooledEmbeddingDist（torchrec.distributed.sharding.dp_sharding 中的類別） DpPooledEmbeddingSharding（torchrec.distributed.sharding.dp_sharding 中的類別） DpSparseFeaturesDist（torchrec.distributed.sharding.dp_sharding 中的類別） dtensor_metadata（torchrec.distributed.embedding_types.ShardedMetaConfig 屬性） DTensorMetadata（torchrec.distributed.embedding_types 中的類別） dtype_to_data_type()（位於模組 torchrec.modules.embedding_configs 中） dummy_tensor()（位於模組 torchrec.distributed.embedding_lookup 中） dynamic_threshold_filter()（位於模組 torchrec.modules.mc_modules 中） E emb_dim_per_rank_per_feature（torchrec.distributed.comm_ops.VariableBatchAll2AllPooledInfo 屬性），[1] embedding_bag_configs()（torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法） （torchrec.modules.embedding_modules.EmbeddingBagCollection 方法） （torchrec.modules.embedding_modules.EmbeddingBagCollectionInterface 方法） （torchrec.quant.embedding_modules.EmbeddingBagCollection 方法） embedding_bags（torchrec.distributed.quant_embeddingbag.ShardedQuantFeatureProcessedEmbeddingBagCollection 屬性） （torchrec.quant.embedding_modules.FeatureProcessedEmbeddingBagCollection 屬性） embedding_config（torchrec.distributed.embedding_sharding.EmbeddingShardingInfo 屬性） embedding_configs()（torchrec.modules.embedding_modules.EmbeddingCollection 方法） （torchrec.modules.embedding_modules.EmbeddingCollectionInterface 方法） （torchrec.modules.mc_modules.ManagedCollisionCollection 方法） （torchrec.quant.embedding_modules.EmbeddingCollection 方法） embedding_dim（torchrec.distributed.comm_ops.All2AllSequenceInfo 屬性），[1] （torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性） （torchrec.modules.embedding_configs.EmbeddingConfig 屬性） embedding_dim()（torchrec.modules.embedding_modules.EmbeddingCollection 方法） （torchrec.modules.embedding_modules.EmbeddingCollectionInterface 方法） （torchrec.quant.embedding_modules.EmbeddingCollection 方法） embedding_dims()（torchrec.distributed.embedding_sharding.EmbeddingSharding 方法） （torchrec.distributed.embedding_types.GroupedEmbeddingConfig 方法） （torchrec.distributed.sharding.cw_sharding.BaseCwEmbeddingSharding 方法） （torchrec.distributed.sharding.dp_sharding.BaseDpEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.BaseRwEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.BaseTwEmbeddingSharding 方法） （torchrec.distributed.sharding.twrw_sharding.BaseTwRwEmbeddingSharding 方法） embedding_names（torchrec.modules.embedding_configs.EmbeddingTableConfig 屬性） embedding_names()（torchrec.distributed.embedding_sharding.EmbeddingSharding 方法） （torchrec.distributed.embedding_types.GroupedEmbeddingConfig 方法） （torchrec.distributed.sharding.cw_sharding.BaseCwEmbeddingSharding 方法） （torchrec.distributed.sharding.dp_sharding.BaseDpEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.BaseRwEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.BaseTwEmbeddingSharding 方法） （torchrec.distributed.sharding.twrw_sharding.BaseTwRwEmbeddingSharding 方法） embedding_names_by_table()（torchrec.modules.embedding_modules.EmbeddingCollection 方法） （torchrec.modules.embedding_modules.EmbeddingCollectionInterface 方法） （torchrec.quant.embedding_modules.EmbeddingCollection 方法） embedding_names_per_rank()（torchrec.distributed.embedding_sharding.EmbeddingSharding 方法） （torchrec.distributed.sharding.dp_sharding.BaseDpEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.BaseRwEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.BaseTwEmbeddingSharding 方法） （torchrec.distributed.sharding.twrw_sharding.BaseTwRwEmbeddingSharding 方法） embedding_shard_metadata()（torchrec.distributed.embedding_sharding.EmbeddingSharding 方法） （torchrec.distributed.embedding_types.GroupedEmbeddingConfig 方法） （torchrec.distributed.sharding.dp_sharding.BaseDpEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.BaseRwEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.BaseTwEmbeddingSharding 方法） （torchrec.distributed.sharding.twrw_sharding.BaseTwRwEmbeddingSharding 方法） embedding_tables（torchrec.distributed.embedding_types.GroupedEmbeddingConfig 屬性） embedding_tables()（torchrec.distributed.embedding_sharding.EmbeddingSharding 方法） （torchrec.distributed.sharding.dp_sharding.BaseDpEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.BaseRwEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.BaseTwEmbeddingSharding 方法） EmbeddingAttributes（torchrec.distributed.embedding_types 中的類別） EmbeddingAwaitable（torchrec.distributed.embeddingbag 中的類別） EmbeddingBagCollection（torchrec.modules.embedding_modules 中的類別） （torchrec.quant.embedding_modules 中的類別） EmbeddingBagCollectionAwaitable（torchrec.distributed.embeddingbag 中的類別） EmbeddingBagCollectionContext（torchrec.distributed.embeddingbag 中的類別） EmbeddingBagCollectionInterface（torchrec.modules.embedding_modules 中的類別） EmbeddingBagCollectionSharder（torchrec.distributed.embeddingbag 中的類別） EmbeddingBagConfig（torchrec.modules.embedding_configs 中的類別） EmbeddingBagSharder（torchrec.distributed.embeddingbag 中的類別） EmbeddingCollection（torchrec.modules.embedding_modules 中的類別） （torchrec.quant.embedding_modules 中的類別） EmbeddingCollectionAwaitable（torchrec.distributed.embedding 中的類別） EmbeddingCollectionContext（torchrec.distributed.embedding 中的類別） EmbeddingCollectionInterface（torchrec.modules.embedding_modules 中的類別） EmbeddingCollectionSharder（torchrec.distributed.embedding 中的類別） EmbeddingComputeKernel（torchrec.distributed.embedding_types 中的類別） EmbeddingConfig（torchrec.modules.embedding_configs 中的類別） EmbeddingEnumerator（torchrec.distributed.planner.enumerators 中的類別） EmbeddingModuleShardingPlan（torchrec.distributed.types 中的類別） EmbeddingOffloadScaleupProposer（torchrec.distributed.planner.proposers 中的類別） EmbeddingOffloadStats（torchrec.distributed.planner.shard_estimators 中的類別） EmbeddingPerfEstimator（torchrec.distributed.planner.shard_estimators 中的類別） embeddings_cat_empty_rank_handle()（位於模組 torchrec.distributed.embedding_lookup 中） embeddings_cat_empty_rank_handle_inference()（位於模組 torchrec.distributed.embedding_lookup 中） EmbeddingsAllToOne（torchrec.distributed.dist_data 中的類別），[1] EmbeddingsAllToOneReduce（torchrec.distributed.dist_data 中的類別），[1] EmbeddingSharding（torchrec.distributed.embedding_sharding 中的類別） EmbeddingShardingContext（torchrec.distributed.embedding_sharding 中的類別） EmbeddingShardingInfo (torchrec.distributed.embedding_sharding 中的類別) EmbeddingShardingPlanner (torchrec.distributed.planner.planners 中的類別) EmbeddingStats (torchrec.distributed.planner.stats 中的類別) EmbeddingStorageEstimator (torchrec.distributed.planner.shard_estimators 中的類別) EmbeddingTableConfig (torchrec.modules.embedding_configs 中的類別) empty() (torchrec.sparse.jagged_tensor.JaggedTensor 的靜態方法) (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 的靜態方法) empty_like() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 的靜態方法) EmptyFusedOptimizer (torchrec.optim.fused 中的類別) encode() (torchrec.distributed.types.NoOpQuantizedCommCodec 的方法) (torchrec.distributed.types.QuantizedCommCodec 方法) enforce_hbm (torchrec.distributed.planner.types.ParameterConstraints 的屬性), [1] (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.types.ParameterSharding 屬性) enumerate() (torchrec.distributed.planner.enumerators.EmbeddingEnumerator 的方法) (torchrec.distributed.planner.types.Enumerator 的方法) Enumerator (torchrec.distributed.planner.types 中的類別) equal_splits (torchrec.distributed.comm_ops.ReduceScatterVInfo 的屬性), [1] estimate() (torchrec.distributed.planner.shard_estimators.EmbeddingPerfEstimator 的方法) (torchrec.distributed.planner.shard_estimators.EmbeddingStorageEstimator 的方法) (torchrec.distributed.planner.types.ShardEstimator 的方法) estimate_cache_miss_rate() (torchrec.distributed.planner.shard_estimators.EmbeddingOffloadStats 的靜態方法) evict() (在 torchrec.modules.mc_embedding_modules 模組中) （torchrec.distributed.mc_modules.ShardedManagedCollisionCollection 方法） （torchrec.modules.mc_modules.ManagedCollisionCollection 方法） (torchrec.modules.mc_modules.ManagedCollisionModule 的方法) (torchrec.modules.mc_modules.MCHManagedCollisionModule 的方法) evictions_per_table (torchrec.distributed.mc_embeddingbag.ManagedCollisionEmbeddingBagCollectionContext 的屬性) expected_lookups (torchrec.distributed.planner.shard_estimators.EmbeddingOffloadStats 的屬性) (torchrec.distributed.types.CacheStatistics 屬性) expected_miss_rate() (torchrec.distributed.planner.shard_estimators.EmbeddingOffloadStats 的方法) (torchrec.distributed.types.CacheStatistics 的方法) extra_repr() (torchrec.distributed.embedding_types.ShardedEmbeddingModule 的方法) extract_module_or_tensor_callable() (在 torchrec.modules.utils 模組中) F FactorizationMachine (torchrec.modules.deepfm 中的類別) feature_hash_sizes() (torchrec.distributed.embedding_types.GroupedEmbeddingConfig 的方法) feature_names (torchrec.distributed.planner.types.ParameterConstraints 的屬性), [1] (torchrec.distributed.planner.types.ShardingOption 屬性) （torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性） （torchrec.modules.embedding_configs.EmbeddingConfig 屬性） feature_names() (torchrec.distributed.embedding_types.FeatureShardingMixIn 的方法) （torchrec.distributed.embedding_types.GroupedEmbeddingConfig 方法） （torchrec.distributed.sharding.dp_sharding.BaseDpEmbeddingSharding 方法） （torchrec.distributed.sharding.rw_sharding.BaseRwEmbeddingSharding 方法） （torchrec.distributed.sharding.tw_sharding.BaseTwEmbeddingSharding 方法） （torchrec.distributed.sharding.twrw_sharding.BaseTwRwEmbeddingSharding 方法） feature_names_per_rank() (torchrec.distributed.embedding_types.FeatureShardingMixIn 的方法) （torchrec.distributed.sharding.tw_sharding.BaseTwEmbeddingSharding 方法） FeatureProcessedEmbeddingBagCollection (torchrec.quant.embedding_modules 中的類別) features (torchrec.distributed.embedding_types.InputDistOutputs 的屬性) features_per_rank() (torchrec.distributed.embedding_types.FeatureShardingMixIn 的方法) （torchrec.distributed.sharding.tw_sharding.BaseTwEmbeddingSharding 方法） FeatureShardingMixIn (torchrec.distributed.embedding_types 中的類別) feedback() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 的方法) (torchrec.distributed.planner.proposers.GreedyProposer 的方法) (torchrec.distributed.planner.proposers.GridSearchProposer 的方法) (torchrec.distributed.planner.proposers.UniformProposer 的方法) (torchrec.distributed.planner.types.Proposer 的方法) filter_state_dict() (在 torchrec.distributed.utils 模組中) fits_in() (torchrec.distributed.planner.types.Storage 的方法) FixedPercentageStorageReservation (torchrec.distributed.planner.storage_reservations 中的類別) flatten_feature_lengths() (在 torchrec.distributed.quant_embeddingbag 模組中) flatten_kjt_list() (在 torchrec.sparse.jagged_tensor 模組中) flatten_lengths() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 的方法) flush() (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedPooledEmbeddingsLookup 的方法) FMInteractionArch (torchrec.models.deepfm 中的類別) for_each_module_of_type_do() (在 torchrec.quant.embedding_modules 模組中) ForkedPdb (torchrec.distributed.utils 中的類別) forward (torchrec.distributed.types.QuantizedCommCodecs 的屬性) forward() (torchrec.distributed.comm_ops.All2All_Pooled_Req 的靜態方法) (torchrec.distributed.comm_ops.All2All_Pooled_Wait 靜態方法) (torchrec.distributed.comm_ops.All2All_Seq_Req 靜態方法) (torchrec.distributed.comm_ops.All2All_Seq_Req_Wait 靜態方法) (torchrec.distributed.comm_ops.All2Allv_Req 靜態方法) (torchrec.distributed.comm_ops.All2Allv_Wait 靜態方法) (torchrec.distributed.comm_ops.AllGatherBase_Req 靜態方法) (torchrec.distributed.comm_ops.AllGatherBase_Wait 靜態方法) (torchrec.distributed.comm_ops.ReduceScatter_Req 靜態方法) (torchrec.distributed.comm_ops.ReduceScatter_Wait 靜態方法) (torchrec.distributed.comm_ops.ReduceScatterBase_Req 靜態方法) (torchrec.distributed.comm_ops.ReduceScatterBase_Wait 靜態方法) (torchrec.distributed.comm_ops.ReduceScatterV_Req 靜態方法) (torchrec.distributed.comm_ops.ReduceScatterV_Wait 靜態方法) (torchrec.distributed.comm_ops.Variable_Batch_All2All_Pooled_Req 靜態方法) (torchrec.distributed.comm_ops.Variable_Batch_All2All_Pooled_Wait 靜態方法) (torchrec.distributed.dist_data.EmbeddingsAllToOne 的方法), [1] (torchrec.distributed.dist_data.EmbeddingsAllToOneReduce 的方法), [1] (torchrec.distributed.dist_data.KJTAllToAll 的方法), [1] (torchrec.distributed.dist_data.KJTOneToAll 的方法), [1] (torchrec.distributed.dist_data.MergePooledEmbeddingsModule 的方法), [1] (torchrec.distributed.dist_data.PooledEmbeddingsAllGather 的方法), [1] (torchrec.distributed.dist_data.PooledEmbeddingsAllToAll 的方法), [1] (torchrec.distributed.dist_data.PooledEmbeddingsReduceScatter 的方法), [1] (torchrec.distributed.dist_data.SeqEmbeddingsAllToOne 的方法), [1] (torchrec.distributed.dist_data.SequenceEmbeddingsAllToAll 的方法), [1] (torchrec.distributed.dist_data.TensorAllToAll 的方法), [1] (torchrec.distributed.dist_data.TensorValuesAllToAll 的方法), [1] (torchrec.distributed.dist_data.VariableBatchPooledEmbeddingsAllToAll 的方法), [1] (torchrec.distributed.dist_data.VariableBatchPooledEmbeddingsReduceScatter 的方法), [1] (torchrec.distributed.embedding_lookup.CommOpGradientScaling 靜態方法) (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.InferGroupedLookupMixin 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_sharding.BaseEmbeddingDist 的方法) (torchrec.distributed.embedding_sharding.BaseSparseFeaturesDist 的方法) (torchrec.distributed.embedding_types.BaseEmbeddingLookup 的方法) (torchrec.distributed.embedding_types.BaseGroupedFeatureProcessor 的方法) (torchrec.distributed.grouped_position_weighted.GroupedPositionWeightedModule 的方法) (torchrec.distributed.model_parallel.DistributedModelParallel 的方法) (torchrec.distributed.quant_embeddingbag.ShardedQuantEbcInputDist 的方法) （torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法） (torchrec.distributed.sharding.cw_sharding.InferCwPooledEmbeddingDist 的方法) (torchrec.distributed.sharding.cw_sharding.InferCwPooledEmbeddingDistWithPermute 的方法) (torchrec.distributed.sharding.dp_sharding.DpPooledEmbeddingDist 的方法) (torchrec.distributed.sharding.dp_sharding.DpSparseFeaturesDist 的方法) (torchrec.distributed.sharding.rw_sharding.InferRwPooledEmbeddingDist 的方法) (torchrec.distributed.sharding.rw_sharding.InferRwSparseFeaturesDist 的方法) (torchrec.distributed.sharding.rw_sharding.RwPooledEmbeddingDist 的方法) (torchrec.distributed.sharding.rw_sharding.RwSparseFeaturesDist 的方法) (torchrec.distributed.sharding.tw_sharding.InferTwPooledEmbeddingDist 的方法) (torchrec.distributed.sharding.tw_sharding.InferTwSparseFeaturesDist 的方法) (torchrec.distributed.sharding.tw_sharding.TwPooledEmbeddingDist 的方法) (torchrec.distributed.sharding.tw_sharding.TwSparseFeaturesDist 的方法) (torchrec.distributed.sharding.twrw_sharding.TwRwPooledEmbeddingDist 的方法) (torchrec.distributed.sharding.twrw_sharding.TwRwSparseFeaturesDist 的方法) （torchrec.distributed.types.ShardedModule 方法） (torchrec.inference.modules.PredictModule 的方法) (torchrec.models.deepfm.DenseArch 的方法) (torchrec.models.deepfm.FMInteractionArch 的方法) (torchrec.models.deepfm.OverArch 的方法) (torchrec.models.deepfm.SimpleDeepFMNN 的方法) (torchrec.models.deepfm.SparseArch 的方法) (torchrec.modules.activation.SwishLayerNorm 的方法) (torchrec.modules.crossnet.CrossNet 的方法) (torchrec.modules.crossnet.LowRankCrossNet 的方法) (torchrec.modules.crossnet.LowRankMixtureCrossNet 的方法) (torchrec.modules.crossnet.VectorCrossNet 的方法) (torchrec.modules.deepfm.DeepFM 的方法) (torchrec.modules.deepfm.FactorizationMachine 的方法) （torchrec.modules.embedding_modules.EmbeddingBagCollection 方法） （torchrec.modules.embedding_modules.EmbeddingBagCollectionInterface 方法） (torchrec.modules.embedding_modules.EmbeddingCollection 的方法) （torchrec.modules.embedding_modules.EmbeddingCollectionInterface 方法） (torchrec.modules.feature_processor.BaseFeatureProcessor 的方法) (torchrec.modules.feature_processor.BaseGroupedFeatureProcessor 的方法) (torchrec.modules.feature_processor.PositionWeightedModule 的方法) (torchrec.modules.feature_processor.PositionWeightedProcessor 的方法) (torchrec.modules.mc_embedding_modules.BaseManagedCollisionEmbeddingCollection 的方法) （torchrec.modules.mc_modules.ManagedCollisionCollection 方法） (torchrec.modules.mc_modules.ManagedCollisionModule 的方法) (torchrec.modules.mc_modules.MCHManagedCollisionModule 的方法) (torchrec.modules.mlp.MLP 的方法) (torchrec.modules.mlp.Perceptron 的方法) （torchrec.quant.embedding_modules.EmbeddingBagCollection 方法） （torchrec.quant.embedding_modules.EmbeddingCollection 方法） (torchrec.quant.embedding_modules.FeatureProcessedEmbeddingBagCollection 的方法) (torchrec.sparse.jagged_tensor.ComputeJTDictToKJT 的方法) (torchrec.sparse.jagged_tensor.ComputeKJTToJTDict 的方法) forward_recat_tensor (torchrec.distributed.comm_ops.All2AllSequenceInfo 的屬性), [1] fqn (torchrec.distributed.planner.types.ShardingOption 的屬性) from_dense() (torchrec.sparse.jagged_tensor.JaggedTensor 的靜態方法) from_dense_lengths() (torchrec.sparse.jagged_tensor.JaggedTensor 的靜態方法) from_float() (torchrec.quant.embedding_modules.EmbeddingBagCollection 的類別方法) (torchrec.quant.embedding_modules.EmbeddingCollection 的類別方法) (torchrec.quant.embedding_modules.FeatureProcessedEmbeddingBagCollection 的類別方法) from_jt_dict() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 的靜態方法) from_lengths_sync() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 的靜態方法) from_local() (torchrec.distributed.types.ShardingEnv 的類別方法) from_offsets_sync() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 的靜態方法) from_process_group() (torchrec.distributed.types.ShardingEnv 的類別方法) from_tensor_list() (torchrec.sparse.jagged_tensor.KeyedTensor 的靜態方法) FUSED (torchrec.distributed.embedding_types.EmbeddingComputeKernel 的屬性) fused_optimizer (torchrec.distributed.embedding.ShardedEmbeddingCollection 的屬性) (torchrec.distributed.embeddingbag.ShardedEmbeddingBag 的屬性) (torchrec.distributed.embeddingbag.ShardedEmbeddingBagCollection 的屬性) (torchrec.distributed.model_parallel.DistributedModelParallel 的屬性) (torchrec.optim.fused.FusedOptimizerModule 的屬性) fused_params (torchrec.distributed.embedding_sharding.EmbeddingShardingInfo 的屬性) (torchrec.distributed.embedding_types.BaseEmbeddingSharder 的屬性) (torchrec.distributed.embedding_types.BaseQuantEmbeddingSharder 的屬性) （torchrec.distributed.embedding_types.GroupedEmbeddingConfig 屬性） (torchrec.distributed.embedding_types.ShardedEmbeddingTable 的屬性) FUSED_UVM (torchrec.distributed.embedding_types.EmbeddingComputeKernel 的屬性) FUSED_UVM_CACHING (torchrec.distributed.embedding_types.EmbeddingComputeKernel 的屬性) FusedKJTListSplitsAwaitable (torchrec.distributed.embedding_sharding 中的類別) FusedOptimizer (torchrec.optim.fused 中的類別) FusedOptimizerModule (torchrec.optim.fused 中的類別) fwd_comms (torchrec.distributed.planner.types.Perf 的屬性) fwd_compute (torchrec.distributed.planner.types.Perf 的屬性) fx_wrap_tensor_view2d() (在 torchrec.distributed.embedding_lookup 模組中) G gb_to_bytes() (在 torchrec.distributed.planner.utils 模組中) GenericMeta (torchrec.distributed.types 中的類別) get_block_sizes_runtime_device() (在 torchrec.distributed.sharding.rw_sharding 模組中) get_budget() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 的靜態方法) get_cacheability() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 的靜態方法) get_data() (torchrec.distributed.planner.types.CustomTopologyData 的方法) get_device_from_parameter_sharding() (在 torchrec.distributed.embedding 模組中) （位於模組 torchrec.distributed.embeddingbag 中） (在 torchrec.distributed.quant_embeddingbag 模組中) get_device_from_sharding_infos() (在 torchrec.distributed.quant_embeddingbag 模組中) get_ec_index_dedup() (在 torchrec.distributed.embedding 模組中) get_embedding_names_by_table() (在 torchrec.modules.embedding_modules 模組中) get_embedding_shard_metadata() (在 torchrec.distributed.sharding.rw_sharding 模組中) get_expected_lookups() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 的靜態方法) get_gradient_division() (在 torchrec.distributed.comm_ops 模組中) get_group_rank() (在 torchrec.distributed.comm 模組中) get_local_rank() (在 torchrec.distributed.comm 模組中) get_local_size() (在 torchrec.distributed.comm 模組中) get_module() (在 torchrec.distributed.model_parallel 模組中) get_module_output_dimension() (在 torchrec.modules.utils 模組中) get_num_groups() (在 torchrec.distributed.comm 模組中) get_partition_by_type() (在 torchrec.distributed.planner.enumerators 模組中) get_plan_for_module() (torchrec.distributed.types.ShardingPlan 的方法) get_tbes_to_register() (torchrec.distributed.embedding_lookup.InferCPUGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.InferGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.InferGroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedPooledEmbeddingsLookup 的方法) get_tensor_size_bytes() (在 torchrec.distributed.types 模組中) get_unsharded_module_names() (在 torchrec.distributed.utils 模組中) get_unwrapped_module() (在 torchrec.distributed.model_parallel 模組中) get_use_sync_collectives() (在 torchrec.distributed.comm_ops 模組中) get_weight_init_max() (torchrec.modules.embedding_configs.BaseEmbeddingConfig 方法) get_weight_init_min() (torchrec.modules.embedding_configs.BaseEmbeddingConfig 方法) GetItemLazyAwaitable (torchrec.distributed.types 中的類別) global_metadata (torchrec.distributed.embedding_types.ShardedMetaConfig 屬性) GradientClipping (torchrec.optim.clipping 中的類別) GradientClippingOptimizer (torchrec.optim.clipping 中的類別) GreedyPerfPartitioner (torchrec.distributed.planner.partitioners 中的類別) GreedyProposer (torchrec.distributed.planner.proposers 中的類別) GridSearchProposer (torchrec.distributed.planner.proposers 中的類別) group_tables() (在 torchrec.distributed.embedding_sharding 模組中) GroupedEmbeddingConfig (torchrec.distributed.embedding_types 中的類別) GroupedEmbeddingsLookup (torchrec.distributed.embedding_lookup 中的類別) GroupedPooledEmbeddingsLookup (torchrec.distributed.embedding_lookup 中的類別) GroupedPositionWeightedModule (torchrec.distributed.grouped_position_weighted 中的類別) H has_data() (torchrec.distributed.planner.types.CustomTopologyData 方法) has_feature_processor (torchrec.distributed.embedding_types.GroupedEmbeddingConfig 屬性) (torchrec.modules.embedding_configs.EmbeddingTableConfig 屬性) hbm (torchrec.distributed.planner.types.Storage 屬性) HBM (torchrec.distributed.types.ParameterStorage 屬性) hbm_mem_bw (torchrec.distributed.planner.types.Topology 屬性) HeteroEmbeddingShardingPlanner (torchrec.distributed.planner.planners 中的類別) HeuristicalStorageReservation (torchrec.distributed.planner.storage_reservations 中的類別) HOST (torchrec.distributed.planner.types.PartitionByType 屬性) I index_per_key() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) InferCPUGroupedEmbeddingsLookup (torchrec.distributed.embedding_lookup 中的類別) InferCwPooledEmbeddingDist (torchrec.distributed.sharding.cw_sharding 中的類別) InferCwPooledEmbeddingDistWithPermute (torchrec.distributed.sharding.cw_sharding 中的類別) InferCwPooledEmbeddingSharding (torchrec.distributed.sharding.cw_sharding 中的類別) inference (torchrec.distributed.types.ObjectPoolShardingPlan 屬性) InferenceStorageReservation (torchrec.distributed.planner.storage_reservations 中的類別) InferGroupedEmbeddingsLookup (torchrec.distributed.embedding_lookup 中的類別) InferGroupedLookupMixin (torchrec.distributed.embedding_lookup 中的類別) InferGroupedPooledEmbeddingsLookup (torchrec.distributed.embedding_lookup 中的類別) InferRwPooledEmbeddingDist (torchrec.distributed.sharding.rw_sharding 中的類別) InferRwPooledEmbeddingSharding (torchrec.distributed.sharding.rw_sharding 中的類別) InferRwSparseFeaturesDist (torchrec.distributed.sharding.rw_sharding 中的類別) InferTwEmbeddingSharding (torchrec.distributed.sharding.tw_sharding 中的類別) InferTwPooledEmbeddingDist (torchrec.distributed.sharding.tw_sharding 中的類別) InferTwSparseFeaturesDist (torchrec.distributed.sharding.tw_sharding 中的類別) init_data_parallel() (torchrec.distributed.model_parallel.DistributedModelParallel 方法) init_fn (torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性) init_mlp_weights_xavier_uniform() (在 torchrec.modules.utils 模組中) init_parameters() (在 torchrec.distributed.utils 模組中) init_state() (torchrec.optim.keyed.KeyedOptimizer 方法) input_dist() (torchrec.distributed.embedding.ShardedEmbeddingCollection 方法) （torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法） （torchrec.distributed.embeddingbag.ShardedEmbeddingBagCollection 方法） （torchrec.distributed.mc_modules.ShardedManagedCollisionCollection 方法） （torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法） （torchrec.distributed.types.ShardedModule 方法） input_lengths (torchrec.distributed.planner.types.ShardingOption 屬性) input_shape (torchrec.distributed.comm_ops.All2AllDenseInfo 屬性) input_size (torchrec.distributed.comm_ops.AllGatherBaseInfo 屬性), [1] input_size() (torchrec.modules.mc_modules.ManagedCollisionModule 方法) (torchrec.modules.mc_modules.MCHManagedCollisionModule 的方法) input_sizes (torchrec.distributed.comm_ops.ReduceScatterBaseInfo 屬性), [1] (torchrec.distributed.comm_ops.ReduceScatterInfo 屬性), [1] (torchrec.distributed.comm_ops.ReduceScatterVInfo 屬性), [1] input_split_sizes (torchrec.distributed.comm_ops.All2AllVInfo 屬性), [1] input_splits (torchrec.distributed.comm_ops.All2AllDenseInfo 屬性) (torchrec.distributed.comm_ops.All2AllSequenceInfo 屬性), [1] (torchrec.distributed.comm_ops.ReduceScatterVInfo 屬性), [1] (torchrec.distributed.comm_ops.VariableBatchAll2AllPooledInfo 屬性), [1] (torchrec.distributed.embedding_sharding.KJTSplitsAllToAllMeta 屬性) input_tensors (torchrec.distributed.embedding_sharding.KJTSplitsAllToAllMeta 屬性) InputDistOutputs (torchrec.distributed.embedding_types 中的類別) INSUFFICIENT_STORAGE (torchrec.distributed.planner.types.PlannerErrorType 屬性) inter_host_bw (torchrec.distributed.planner.types.Topology 屬性) interaction() (torchrec.distributed.utils.ForkedPdb 方法) intra_and_cross_node_pg() (在 torchrec.distributed.comm 模組中) intra_host_bw (torchrec.distributed.planner.types.Topology 屬性) inverse_indices (torchrec.distributed.embeddingbag.EmbeddingBagCollectionContext 屬性) inverse_indices() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) inverse_indices_or_none() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) invoke_on_rank_and_broadcast_result() (在 torchrec.distributed.collective_utils 模組中) INVSQRT (torchrec.optim.warmup.WarmupPolicy 屬性) is_fx_tracing() (在 torchrec.fx.tracer 模組中) is_history_metadata (torchrec.modules.mc_modules.MCHEvictionPolicyMetadataInfo 屬性) is_leader() (在 torchrec.distributed.collective_utils 模組中) is_leaf_module() (torchrec.fx.tracer.Tracer 方法) is_mch_metadata (torchrec.modules.mc_modules.MCHEvictionPolicyMetadataInfo 屬性) is_pooled (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.planner.types.ShardingOption 屬性) is_weighted (torchrec.distributed.embedding_types.GroupedEmbeddingConfig 屬性) (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] (torchrec.modules.embedding_configs.EmbeddingTableConfig 屬性) is_weighted() (torchrec.modules.embedding_modules.EmbeddingBagCollection 方法) （torchrec.modules.embedding_modules.EmbeddingBagCollectionInterface 方法） （torchrec.quant.embedding_modules.EmbeddingBagCollection 方法） J jagged_index_select_with_empty() (在 torchrec.modules.utils 模組中) JaggedTensor (torchrec.sparse.jagged_tensor 中的類別) JaggedTensorAllToAll (torchrec.distributed.dist_data 中的類別), [1] JaggedTensorMeta (torchrec.sparse.jagged_tensor 中的類別) jt_is_equal() (在 torchrec.sparse.jagged_tensor 模組中) K kernel_bw_lookup() (在 torchrec.distributed.planner.constants 模組中) key_dim() (torchrec.sparse.jagged_tensor.KeyedTensor 方法) KEY_VALUE (torchrec.distributed.embedding_types.EmbeddingComputeKernel 屬性) key_value_params (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.types.ParameterSharding 屬性) KeyedJaggedTensor (torchrec.sparse.jagged_tensor 中的類別) KeyedOptimizer (torchrec.optim.keyed 中的類別) KeyedOptimizerWrapper (torchrec.optim.keyed 中的類別) KeyedTensor (torchrec.sparse.jagged_tensor 中的類別) keys (torchrec.distributed.embedding_sharding.KJTSplitsAllToAllMeta 屬性) keys() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedTensor 方法） KeyValueParams (torchrec.distributed.types 中的類別) kjt_is_equal() (在 torchrec.sparse.jagged_tensor 模組中) KJTAllToAll (torchrec.distributed.dist_data 中的類別), [1] KJTAllToAllSplitsAwaitable (torchrec.distributed.dist_data 中的類別), [1] KJTAllToAllTensorsAwaitable (torchrec.distributed.dist_data 中的類別), [1] KJTList (torchrec.distributed.embedding_types 中的類別) KJTListAwaitable (torchrec.distributed.embedding_sharding 中的類別) KJTListSplitsAwaitable (torchrec.distributed.embedding_sharding 中的類別) KJTOneToAll (torchrec.distributed.dist_data 中的類別), [1] KJTSplitsAllToAllMeta (torchrec.distributed.embedding_sharding 中的類別) L labels (torchrec.distributed.embedding_sharding.KJTSplitsAllToAllMeta 屬性) LAMB (torchrec.distributed.embedding_types.OptimType 屬性) LARS_SGD (torchrec.distributed.embedding_types.OptimType 屬性) lazy_apply() (在 torchrec.modules.lazy_extension 模組中) LazyAwaitable (torchrec.distributed.types 中的類別) LazyGetItemMixin (torchrec.distributed.types 中的類別) LazyModuleExtensionMixin (torchrec.modules.lazy_extension 中的類別) LazyNoWait (torchrec.distributed.types 中的類別) length_per_key() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedTensor 方法） length_per_key_or_none() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) lengths() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） lengths_after_sparse_data_all2all (torchrec.distributed.comm_ops.All2AllSequenceInfo 屬性), [1] lengths_offset_per_key() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) lengths_or_none() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） LFU_EvictionPolicy (torchrec.modules.mc_modules 中的類別) LINEAR (torchrec.optim.warmup.WarmupPolicy 屬性) LION (torchrec.distributed.embedding_types.OptimType 屬性) ListOfKJTList (torchrec.distributed.embedding_types 中的類別) ListOfKJTListAwaitable (torchrec.distributed.embedding_sharding 中的類別) ListOfKJTListSplitsAwaitable (torchrec.distributed.embedding_sharding 中的類別) load() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 方法) (torchrec.distributed.planner.proposers.GreedyProposer 的方法) (torchrec.distributed.planner.proposers.GridSearchProposer 的方法) (torchrec.distributed.planner.proposers.UniformProposer 的方法) (torchrec.distributed.planner.types.Proposer 的方法) load_config_text() (在 torchrec.inference.model_packager 模組中) load_factor (torchrec.distributed.types.CacheParams 屬性), [1] load_pickle_config() (在 torchrec.inference.model_packager 模組中) load_state_dict() (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 方法) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.InferGroupedLookupMixin 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedPooledEmbeddingsLookup 的方法) （torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法） (torchrec.distributed.model_parallel.DistributedModelParallel 的方法) (torchrec.optim.keyed.KeyedOptimizer 方法) (torchrec.optim.keyed.OptimizerWrapper 方法) local_cols (torchrec.distributed.embedding_types.ShardedConfig 屬性) local_metadata (torchrec.distributed.embedding_types.ShardedMetaConfig 屬性) local_rows (torchrec.distributed.embedding_types.ShardedConfig 屬性) local_world_size (torchrec.distributed.planner.partitioners.OrderedDeviceHardware 屬性) (torchrec.distributed.planner.types.Topology 屬性) log() (torchrec.distributed.planner.stats.EmbeddingStats 方法) (torchrec.distributed.planner.stats.NoopEmbeddingStats 方法) (torchrec.distributed.planner.types.Stats 方法) LowRankCrossNet (torchrec.modules.crossnet 中的類別) LowRankMixtureCrossNet (torchrec.modules.crossnet 中的類別) lr_scale (torchrec.optim.warmup.WarmupStage 屬性) LRU_EvictionPolicy (torchrec.modules.mc_modules 中的類別) LuusJaakolaSearch (torchrec.distributed.planner.utils 中的類別) M ManagedCollisionCollection (torchrec.modules.mc_modules 中的類別) ManagedCollisionCollectionAwaitable (torchrec.distributed.mc_modules 中的類別) ManagedCollisionCollectionContext (torchrec.distributed.mc_modules 中的類別) ManagedCollisionCollectionSharder (torchrec.distributed.mc_modules 中的類別) ManagedCollisionEmbeddingBagCollection (torchrec.modules.mc_embedding_modules 中的類別) ManagedCollisionEmbeddingBagCollectionContext (torchrec.distributed.mc_embeddingbag 中的類別) ManagedCollisionEmbeddingBagCollectionSharder (torchrec.distributed.mc_embeddingbag 中的類別) ManagedCollisionEmbeddingCollection (torchrec.modules.mc_embedding_modules 中的類別) ManagedCollisionEmbeddingCollectionContext (torchrec.distributed.mc_embedding 中的類別) ManagedCollisionEmbeddingCollectionSharder (torchrec.distributed.mc_embedding 中的類別) ManagedCollisionModule (torchrec.modules.mc_modules 中的類別) max_iters (torchrec.optim.warmup.WarmupStage 屬性) MCHEvictionPolicy (torchrec.modules.mc_modules 中的類別) MCHEvictionPolicyMetadataInfo (torchrec.modules.mc_modules 中的類別) MCHManagedCollisionModule (torchrec.modules.mc_modules 中的類別) MEAN (torchrec.modules.embedding_configs.PoolingType 屬性) MemoryBalancedPartitioner (torchrec.distributed.planner.partitioners 中的類別) merge_fused_params() (在 torchrec.distributed.utils 模組中) MergePooledEmbeddingsModule (torchrec.distributed.dist_data 中的類別), [1] mesh (torchrec.distributed.embedding_types.DTensorMetadata 屬性) metadata_info (torchrec.modules.mc_modules.DistanceLFU_EvictionPolicy 屬性) (torchrec.modules.mc_modules.LFU_EvictionPolicy 屬性) (torchrec.modules.mc_modules.LRU_EvictionPolicy 屬性) (torchrec.modules.mc_modules.MCHEvictionPolicy 屬性) metadata_name (torchrec.modules.mc_modules.MCHEvictionPolicyMetadataInfo 屬性) MetaInferGroupedEmbeddingsLookup (torchrec.distributed.embedding_lookup 中的類別) MetaInferGroupedPooledEmbeddingsLookup (torchrec.distributed.embedding_lookup 中的類別) min_partition (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] MLP (torchrec.modules.mlp 中的類別) model_inputs_data() (torchrec.inference.modules.PredictFactory 方法) 模組 torchrec.distributed torchrec.distributed.collective_utils torchrec.distributed.comm torchrec.distributed.comm_ops torchrec.distributed.dist_data, [1] torchrec.distributed.embedding torchrec.distributed.embedding_lookup torchrec.distributed.embedding_sharding torchrec.distributed.embedding_types torchrec.distributed.embeddingbag torchrec.distributed.grouped_position_weighted torchrec.distributed.mc_embedding torchrec.distributed.mc_embeddingbag torchrec.distributed.mc_modules torchrec.distributed.model_parallel torchrec.distributed.planner torchrec.distributed.planner.constants torchrec.distributed.planner.enumerators torchrec.distributed.planner.partitioners torchrec.distributed.planner.perf_models torchrec.distributed.planner.planners torchrec.distributed.planner.proposers torchrec.distributed.planner.shard_estimators torchrec.distributed.planner.stats torchrec.distributed.planner.storage_reservations torchrec.distributed.planner.types torchrec.distributed.planner.utils torchrec.distributed.quant_embeddingbag torchrec.distributed.sharding torchrec.distributed.sharding.cw_sharding torchrec.distributed.sharding.dp_sharding torchrec.distributed.sharding.rw_sharding torchrec.distributed.sharding.tw_sharding torchrec.distributed.sharding.twcw_sharding torchrec.distributed.sharding.twrw_sharding torchrec.distributed.train_pipeline torchrec.distributed.types torchrec.distributed.utils torchrec.fx, [1] torchrec.fx.tracer torchrec.inference, [1] torchrec.inference.model_packager torchrec.inference.modules torchrec.models.deepfm torchrec.modules torchrec.modules.activation torchrec.modules.crossnet torchrec.modules.deepfm torchrec.modules.embedding_configs torchrec.modules.embedding_modules torchrec.modules.feature_processor torchrec.modules.lazy_extension torchrec.modules.mc_embedding_modules torchrec.modules.mc_modules torchrec.modules.mlp torchrec.modules.utils torchrec.optim, [1] torchrec.optim.clipping torchrec.optim.fused torchrec.optim.keyed torchrec.optim.warmup torchrec.quant, [1] torchrec.quant.embedding_modules torchrec.sparse, [1] torchrec.sparse.jagged_tensor 模組 (torchrec.distributed.model_parallel.DistributedModelParallel 屬性) (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.planner.types.ShardingOption 屬性) module_pooled() (torchrec.distributed.planner.types.ShardingOption 静態方法) module_type (torchrec.distributed.embedding.EmbeddingCollectionSharder 屬性) (torchrec.distributed.embeddingbag.EmbeddingBagCollectionSharder 屬性) (torchrec.distributed.embeddingbag.EmbeddingBagSharder 屬性) (torchrec.distributed.mc_embedding.ManagedCollisionEmbeddingCollectionSharder 屬性) (torchrec.distributed.mc_embeddingbag.ManagedCollisionEmbeddingBagCollectionSharder 屬性) (torchrec.distributed.mc_modules.ManagedCollisionCollectionSharder 屬性) (torchrec.distributed.quant_embeddingbag.QuantEmbeddingBagCollectionSharder 屬性) (torchrec.distributed.quant_embeddingbag.QuantFeatureProcessedEmbeddingBagCollectionSharder 屬性) (torchrec.distributed.types.ModuleSharder 屬性) ModuleSharder (torchrec.distributed.types 中的類別) ModuleShardingMixIn (torchrec.distributed.embedding_types 中的類別) ModuleShardingPlan (torchrec.distributed.types 中的類別) multipass_prefetch_config (torchrec.distributed.types.CacheParams 屬性) N name (torchrec.distributed.planner.types.ShardingOption 屬性) （torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性） named_buffers() (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 方法) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.InferGroupedLookupMixin 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedPooledEmbeddingsLookup 的方法) （torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法） (torchrec.distributed.grouped_position_weighted.GroupedPositionWeightedModule 的方法) (torchrec.distributed.model_parallel.DistributedModelParallel 的方法) (torchrec.modules.feature_processor.PositionWeightedProcessor 的方法) named_modules() (torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法) named_parameters() (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 方法) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.InferGroupedLookupMixin 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedPooledEmbeddingsLookup 的方法) （torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法） (torchrec.distributed.grouped_position_weighted.GroupedPositionWeightedModule 的方法) (torchrec.distributed.model_parallel.DistributedModelParallel 的方法) named_parameters_by_table() (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 方法) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 的方法) need_indices() (torchrec.modules.embedding_modules.EmbeddingCollection 方法) （torchrec.modules.embedding_modules.EmbeddingCollectionInterface 方法） （torchrec.quant.embedding_modules.EmbeddingCollection 方法） need_pos (torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性) need_preproc (torchrec.inference.modules.QualNameMetadata 屬性) next() (torchrec.distributed.planner.utils.BinarySearchPredicate 方法) (torchrec.distributed.planner.utils.LuusJaakolaSearch 方法) next_plan() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 静態方法) NONE (torchrec.distributed.types.PipelineType 屬性) (torchrec.modules.embedding_configs.PoolingType 屬性) (torchrec.optim.clipping.GradientClipping 屬性) (torchrec.optim.warmup.WarmupPolicy 屬性) none_throws() (在 torchrec.distributed.utils 模組中) NoopEmbeddingStats (torchrec.distributed.planner.stats 中的類別) NoopPerfModel (torchrec.distributed.planner.perf_models 中的類別) NoOpQuantizedCommCodec (torchrec.distributed.types 中的類別) NoopStorageModel (torchrec.distributed.planner.perf_models 中的類別) NORM (torchrec.optim.clipping.GradientClipping 屬性) NoWait (torchrec.distributed.types 中的類別) NullShardedModuleContext (torchrec.distributed.types 中的類別) NullShardingContext (torchrec.distributed.types 中的類別) num_embeddings (torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性) （torchrec.modules.embedding_configs.EmbeddingConfig 屬性） num_features() (torchrec.distributed.embedding_types.GroupedEmbeddingConfig 方法) (torchrec.modules.embedding_configs.BaseEmbeddingConfig 方法) num_inputs (torchrec.distributed.planner.types.ShardingOption 屬性) num_poolings (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] num_shards (torchrec.distributed.planner.types.ShardingOption 屬性) O ObjectPoolShardingPlan (torchrec.distributed.types 中的類別) ObjectPoolShardingType (torchrec.distributed.types 中的類別) offset (torchrec.distributed.planner.types.Shard 屬性) offset_per_key() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedTensor 方法） offset_per_key_or_none() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) offsets() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） offsets_or_none() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） offsets_to_range_traceble() (在 torchrec.modules.feature_processor 模組中) optimizer_type_to_emb_opt_type() (在 torchrec.distributed.utils 模組中) optimizers (torchrec.optim.keyed.CombinedOptimizer 屬性) OptimizerWrapper (torchrec.optim.keyed 中的類別) OptimType (torchrec.distributed.embedding_types 中的類別) OrderedDeviceHardware (torchrec.distributed.planner.partitioners 中的類別) OTHER (torchrec.distributed.planner.types.PlannerErrorType 屬性) output_dist() (torchrec.distributed.embedding.ShardedEmbeddingCollection 方法) （torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法） （torchrec.distributed.embeddingbag.ShardedEmbeddingBagCollection 方法） （torchrec.distributed.mc_modules.ShardedManagedCollisionCollection 方法） （torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法） （torchrec.distributed.types.ShardedModule 方法） output_dtype (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.types.ParameterSharding 屬性) output_dtype() (torchrec.quant.embedding_modules.EmbeddingBagCollection 方法) （torchrec.quant.embedding_modules.EmbeddingCollection 方法） output_size() (torchrec.modules.mc_modules.ManagedCollisionModule 方法) (torchrec.modules.mc_modules.MCHManagedCollisionModule 的方法) output_split_sizes (torchrec.distributed.comm_ops.All2AllVInfo 屬性), [1] output_splits (torchrec.distributed.comm_ops.All2AllDenseInfo 屬性) (torchrec.distributed.comm_ops.All2AllSequenceInfo 屬性), [1] (torchrec.distributed.comm_ops.VariableBatchAll2AllPooledInfo 屬性), [1] OverArch (torchrec.models.deepfm 中的類別) P pad_vbe_kjt_lengths() (在 torchrec.distributed.embedding 模組中) param (torchrec.distributed.embedding_sharding.EmbeddingShardingInfo 屬性) param_count (torchrec.distributed.planner.partitioners.ShardingOptionGroup 屬性) param_groups (torchrec.optim.keyed.CombinedOptimizer 屬性) param_sharding (torchrec.distributed.embedding_sharding.EmbeddingShardingInfo 屬性) ParameterConstraints (torchrec.distributed.planner.types 中的類別) ParameterSharding (torchrec.distributed.types 中的類別) ParameterStorage (torchrec.distributed.types 中的類別) params (torchrec.optim.keyed.CombinedOptimizer 屬性) PARTIAL_ROWWISE_ADAM (torchrec.distributed.embedding_types.OptimType 屬性) PARTIAL_ROWWISE_LAMB (torchrec.distributed.embedding_types.OptimType 屬性) PARTITION (torchrec.distributed.planner.types.PlannerErrorType 屬性) partition() (torchrec.distributed.planner.partitioners.GreedyPerfPartitioner 方法) (torchrec.distributed.planner.partitioners.MemoryBalancedPartitioner 方法) (torchrec.distributed.planner.types.Partitioner 方法) PartitionByType (torchrec.distributed.planner.types 中的類別) Partitioner (torchrec.distributed.planner.types 中的類別) path (torchrec.distributed.planner.types.ShardingOption 屬性) path_of_module() (torchrec.fx.tracer.Tracer 方法) per_table_weight_dtype (torchrec.modules.embedding_configs.QuantConfig 屬性) Perceptron (torchrec.modules.mlp 中的類別) Perf (torchrec.distributed.planner.types 中的類別) PERF (torchrec.distributed.planner.partitioners.SortBy 屬性) perf (torchrec.distributed.planner.types.DeviceHardware 屬性) (torchrec.distributed.planner.types.Shard 屬性) perf_func_emb_wall_time() (torchrec.distributed.planner.shard_estimators.EmbeddingPerfEstimator 類別方法) perf_sum (torchrec.distributed.planner.partitioners.ShardingOptionGroup 屬性) PerfModel (torchrec.distributed.planner.types 中的類別) permute() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) permute_multi_embedding() (在 torchrec.sparse.jagged_tensor 模組中) permuted_lengths_after_sparse_data_all2all (torchrec.distributed.comm_ops.All2AllSequenceInfo 屬性), [1] pg (torchrec.distributed.embedding_sharding.KJTSplitsAllToAllMeta 屬性) pg_name() (在 torchrec.distributed.comm_ops 模組中) pin_memory() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) pinned (torchrec.inference.modules.BatchingMetadata 屬性) PipelineType (torchrec.distributed.types 中的類別) placement() (在 torchrec.distributed.planner.utils 模組中) placements (torchrec.distributed.embedding_types.DTensorMetadata 屬性) plan (torchrec.distributed.model_parallel.DistributedModelParallel 屬性) (torchrec.distributed.types.ShardingPlan 屬性), [1] plan() (torchrec.distributed.planner.planners.EmbeddingShardingPlanner 方法) （torchrec.distributed.planner.planners.HeteroEmbeddingShardingPlanner 方法） （torchrec.distributed.types.ShardingPlanner 方法） PlannerError PlannerErrorType (torchrec.distributed.planner.types 中的類別) policy (torchrec.optim.warmup.WarmupStage 屬性) POLY (torchrec.optim.warmup.WarmupPolicy 屬性) POOLED_EMBEDDINGS_ALL_TO_ALL (torchrec.distributed.types.CommOp 屬性) POOLED_EMBEDDINGS_REDUCE_SCATTER (torchrec.distributed.types.CommOp 屬性) PooledEmbeddingsAllGather (torchrec.distributed.dist_data 中的類別), [1] PooledEmbeddingsAllToAll (torchrec.distributed.dist_data 中的類別), [1] PooledEmbeddingsAwaitable (torchrec.distributed.dist_data 中的類別), [1] PooledEmbeddingsReduceScatter (torchrec.distributed.dist_data 中的類別), [1] pooling (torchrec.distributed.embedding_types.GroupedEmbeddingConfig 屬性) (torchrec.modules.embedding_configs.EmbeddingBagConfig 屬性) (torchrec.modules.embedding_configs.EmbeddingTableConfig 屬性) pooling_factors (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] pooling_type_to_pooling_mode() (在 torchrec.modules.embedding_configs 模組中) pooling_type_to_str() (在 torchrec.modules.embedding_configs 模組中) PoolingType (torchrec.modules.embedding_configs 中的類別) populate_estimates() (torchrec.distributed.planner.enumerators.EmbeddingEnumerator 方法) (torchrec.distributed.planner.types.Enumerator 的方法) position_weighted_module_update_features() (在 torchrec.modules.feature_processor 模組中) PositionWeightedModule (torchrec.modules.feature_processor 中的類別) PositionWeightedProcessor (torchrec.modules.feature_processor 中的類別) post_load_state_dict() (torchrec.optim.keyed.CombinedOptimizer 方法) (torchrec.optim.keyed.KeyedOptimizer 方法) (torchrec.optim.keyed.OptimizerWrapper 方法) (torchrec.optim.warmup.WarmupOptimizer 方法) precision (torchrec.distributed.types.CacheParams 屬性), [1] predict_forward() (torchrec.inference.modules.PredictModule 方法) predict_module (torchrec.inference.modules.PredictModule 屬性) PredictFactory (torchrec.inference.modules 中的類別) PredictFactoryPackager (torchrec.inference.model_packager 中的類別) PredictModule (torchrec.inference.modules 中的類別) prefetch() (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 方法) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_types.ShardedEmbeddingModule 方法) prefetch_compute (torchrec.distributed.planner.types.Perf 屬性) prefetch_pipeline (torchrec.distributed.types.CacheParams 屬性), [1] prepend_opt_key() (torchrec.optim.keyed.CombinedOptimizer 静態方法) preprocess() (torchrec.modules.mc_modules.ManagedCollisionModule 方法) (torchrec.modules.mc_modules.MCHManagedCollisionModule 的方法) probabilistic_threshold_filter() (在 torchrec.modules.mc_modules 模組中) process_pooled_embeddings() (在 torchrec.modules.embedding_modules 模組中) prod() (在 torchrec.distributed.planner.utils 模組中) profile() (torchrec.modules.mc_modules.MCHManagedCollisionModule 方法) promote_high_prefetch_overheaad_table_to_hbm() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 静態方法) propose() (torchrec.distributed.planner.proposers.EmbeddingOffloadScaleupProposer 方法) (torchrec.distributed.planner.proposers.GreedyProposer 的方法) (torchrec.distributed.planner.proposers.GridSearchProposer 的方法) (torchrec.distributed.planner.proposers.UniformProposer 的方法) (torchrec.distributed.planner.types.Proposer 的方法) Proposer (torchrec.distributed.planner.types 中的類別) proposers_to_proposals_list() (在 torchrec.distributed.planner.proposers 模組中) pruned_num_embeddings() (在 torchrec.quant.embedding_modules 模組中) pruning_indices_remapping (torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性) ps_hosts (torchrec.distributed.types.KeyValueParams 屬性), [1] purge() (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 方法) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedPooledEmbeddingsLookup 的方法) Q qcomm_codecs_registry (torchrec.distributed.embedding_sharding.EmbeddingSharding 屬性) (torchrec.distributed.types.ModuleSharder 屬性) (torchrec.distributed.types.ShardedModule 屬性) qualname_metadata() (torchrec.inference.modules.PredictFactory 方法) qualname_metadata_json() (torchrec.inference.modules.PredictFactory 方法) QualNameMetadata (torchrec.inference.modules 中的類別) QUANT (torchrec.distributed.embedding_types.EmbeddingComputeKernel 屬性) quant_prep_customize_row_alignment() (在 torchrec.quant.embedding_modules 模組中) quant_prep_enable_quant_state_dict_split_scale_bias() (在 torchrec.quant.embedding_modules 模組中) quant_prep_enable_quant_state_dict_split_scale_bias_for_types() (在 torchrec.quant.embedding_modules 模組中) quant_prep_enable_register_tbes() (在 torchrec.quant.embedding_modules 模組中) QUANT_UVM (torchrec.distributed.embedding_types.EmbeddingComputeKernel 屬性) QUANT_UVM_CACHING (torchrec.distributed.embedding_types.EmbeddingComputeKernel 屬性) QuantConfig (torchrec.modules.embedding_configs 中的類別) QuantEmbeddingBagCollectionSharder (torchrec.distributed.quant_embeddingbag 中的類別) QuantFeatureProcessedEmbeddingBagCollectionSharder (torchrec.distributed.quant_embeddingbag 中的類別) quantize_dense() (在 torchrec.inference.modules 模組中) quantize_embeddings() (在 torchrec.inference.modules 模組中) quantize_feature() (在 torchrec.inference.modules 模組中) quantize_inference_model() (在 torchrec.inference.modules 模組中) quantize_state_dict() (在 torchrec.quant.embedding_modules 模組中) quantized_dtype (torchrec.distributed.types.QuantizedCommCodec 屬性) quantized_dtype() (torchrec.distributed.types.NoOpQuantizedCommCodec 方法) QuantizedCommCodec (torchrec.distributed.types 中的類別) QuantizedCommCodecs (torchrec.distributed.types 中的類別) R rank (torchrec.distributed.planner.types.DeviceHardware 屬性) (torchrec.distributed.planner.types.Shard 屬性) rank_device() (在 torchrec.distributed.types 模組中) ranks (torchrec.distributed.types.ParameterSharding 屬性) rate() (torchrec.distributed.planner.perf_models.NoopPerfModel 方法) (torchrec.distributed.planner.perf_models.NoopStorageModel 方法) (torchrec.distributed.planner.types.PerfModel 方法) rebuild_with_output_id_range() (torchrec.modules.mc_modules.ManagedCollisionModule 方法) (torchrec.modules.mc_modules.MCHManagedCollisionModule 的方法) recat (torchrec.modules.utils.SequenceVBEContext 屬性) record_history_metadata() (torchrec.modules.mc_modules.DistanceLFU_EvictionPolicy 方法) (torchrec.modules.mc_modules.LFU_EvictionPolicy 方法) （torchrec.modules.mc_modules.LRU_EvictionPolicy 方法） （torchrec.modules.mc_modules.MCHEvictionPolicy 方法） record_stream() (torchrec.distributed.embedding.EmbeddingCollectionContext 方法) (torchrec.distributed.embedding_sharding.EmbeddingShardingContext 方法) (torchrec.distributed.embedding_types.InputDistOutputs 方法) (torchrec.distributed.embedding_types.KJTList 方法) (torchrec.distributed.embedding_types.ListOfKJTList 方法) (torchrec.distributed.embeddingbag.EmbeddingBagCollectionContext 方法) (torchrec.distributed.mc_embedding.ManagedCollisionEmbeddingCollectionContext 方法) (torchrec.distributed.mc_embeddingbag.ManagedCollisionEmbeddingBagCollectionContext 方法) (torchrec.distributed.types.NullShardedModuleContext 方法) (torchrec.distributed.types.NullShardingContext 方法) (torchrec.modules.utils.SequenceVBEContext 方法) (torchrec.sparse.jagged_tensor.JaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） （torchrec.sparse.jagged_tensor.KeyedTensor 方法） reduce_scatter_base_pooled() (在 torchrec.distributed.comm_ops 模組中) reduce_scatter_base_sync() (在 torchrec.distributed.comm_ops 模組中) reduce_scatter_pooled() (在 torchrec.distributed.comm_ops 模組中) reduce_scatter_sync() (在 torchrec.distributed.comm_ops 模組中) reduce_scatter_tensor_backward() (在 torchrec.distributed.comm_ops 模組中) reduce_scatter_tensor_fake() (在 torchrec.distributed.comm_ops 模組中) reduce_scatter_tensor_setup_context() (在 torchrec.distributed.comm_ops 模組中) reduce_scatter_v_per_feature_pooled() (在 torchrec.distributed.comm_ops 模組中) reduce_scatter_v_pooled() (在 torchrec.distributed.comm_ops 模組中) reduce_scatter_v_sync() (在 torchrec.distributed.comm_ops 模組中) ReduceScatter_Req (torchrec.distributed.comm_ops 中的類別) ReduceScatter_Wait (torchrec.distributed.comm_ops 中的類別) ReduceScatterBase_Req (torchrec.distributed.comm_ops 中的類別) ReduceScatterBase_Wait (torchrec.distributed.comm_ops 中的類別) ReduceScatterBaseInfo (torchrec.distributed.comm_ops 中的類別) ReduceScatterInfo (torchrec.distributed.comm_ops 中的類別) ReduceScatterV_Req (torchrec.distributed.comm_ops 中的類別) ReduceScatterV_Wait (torchrec.distributed.comm_ops 中的類別) ReduceScatterVInfo (torchrec.distributed.comm_ops 中的類別) regroup() (torchrec.sparse.jagged_tensor.KeyedTensor 静態方法) regroup_as_dict() (torchrec.sparse.jagged_tensor.KeyedTensor 静態方法) regroup_kts() (在 torchrec.sparse.jagged_tensor 模組中) reindexed_length_per_key (torchrec.modules.utils.SequenceVBEContext 屬性) reindexed_lengths (torchrec.modules.utils.SequenceVBEContext 屬性) reindexed_values (torchrec.modules.utils.SequenceVBEContext 屬性) remap() (torchrec.modules.mc_modules.MCHManagedCollisionModule 方法) remapped_kjt (torchrec.distributed.mc_embeddingbag.ManagedCollisionEmbeddingBagCollectionContext 屬性) reorder_inverse_indices() (在 torchrec.modules.embedding_modules 模組中) replace_placement_with_meta_device() (在 torchrec.distributed.embeddingbag 模組中) REPLICATED_ROW_WISE (torchrec.distributed.types.ObjectPoolShardingType 屬性) Request (torchrec.distributed.comm_ops 中的類別) reserve() (torchrec.distributed.planner.storage_reservations.FixedPercentageStorageReservation 方法) (torchrec.distributed.planner.storage_reservations.HeuristicalStorageReservation 方法) (torchrec.distributed.planner.storage_reservations.InferenceStorageReservation 方法) (torchrec.distributed.planner.types.StorageReservation 方法) reserved_memory (torchrec.distributed.types.CacheParams 屬性), [1] reset_parameters() (torchrec.distributed.embedding.ShardedEmbeddingCollection 方法) （torchrec.distributed.embeddingbag.ShardedEmbeddingBagCollection 方法） （torchrec.modules.embedding_modules.EmbeddingBagCollection 方法） (torchrec.modules.embedding_modules.EmbeddingCollection 的方法) (torchrec.modules.feature_processor.PositionWeightedModule 的方法) reset_shard_rank() (在 torchrec.distributed.planner.utils 模組中) result_metadata() (torchrec.inference.modules.PredictFactory 方法) round_to_one_sigfig() (在 torchrec.distributed.planner.stats 模組中) ROW_WISE (torchrec.distributed.types.ObjectPoolShardingType 屬性) (torchrec.distributed.types.ShardingType 屬性) （torchrec.modules.embedding_configs.ShardingType 屬性） ROWWISE_ADAGRAD (torchrec.distributed.embedding_types.OptimType 屬性) run_on_leader() (在 torchrec.distributed.collective_utils 模組中) run_weights_dependent_transformations() (torchrec.inference.modules.PredictFactory 方法) run_weights_independent_tranformations() (torchrec.inference.modules.PredictFactory 方法) RwPooledEmbeddingDist (torchrec.distributed.sharding.rw_sharding 中的類別) RwPooledEmbeddingSharding (torchrec.distributed.sharding.rw_sharding 中的類別) RwSparseFeaturesDist (torchrec.distributed.sharding.rw_sharding 中的類別) S save_param_groups() (torchrec.optim.keyed.CombinedOptimizer 方法) (torchrec.optim.keyed.KeyedOptimizer 方法) (torchrec.optim.keyed.OptimizerWrapper 方法) save_predict_factory() (torchrec.inference.model_packager.PredictFactoryPackager 類別方法) scope() (在 torchrec.distributed.types 模組中) SeqEmbeddingsAllToOne (torchrec.distributed.dist_data 中的類別), [1] SEQUENCE_EMBEDDINGS_ALL_TO_ALL (torchrec.distributed.types.CommOp 屬性) SequenceEmbeddingsAllToAll (torchrec.distributed.dist_data 中的類別), [1] SequenceEmbeddingsAwaitable (torchrec.distributed.dist_data 中的類別), [1] SequenceVBEContext (torchrec.modules.utils 中的類別) set_device() (torchrec.distributed.dist_data.EmbeddingsAllToOne 方法), [1] (torchrec.distributed.dist_data.EmbeddingsAllToOneReduce 方法), [1] (torchrec.distributed.dist_data.MergePooledEmbeddingsModule 方法), [1] (torchrec.distributed.dist_data.SeqEmbeddingsAllToOne 方法), [1] set_ec_index_dedup() (在 torchrec.distributed.embedding 模組中) set_extern_modules() (torchrec.inference.model_packager.PredictFactoryPackager 類別方法) set_gradient_division() (在 torchrec.distributed.comm_ops 模組中) set_hbm_per_device() (在 torchrec.distributed.planner.partitioners 模組中) set_mocked_modules() (torchrec.inference.model_packager.PredictFactoryPackager 類別方法) set_use_sync_collectives() (在 torchrec.distributed.comm_ops 模組中) SGD (torchrec.distributed.embedding_types.OptimType 屬性) sgdr_period (torchrec.optim.warmup.WarmupStage 屬性) SHAMPOO (torchrec.distributed.embedding_types.OptimType 屬性) SHAMPOO_V2 (torchrec.distributed.embedding_types.OptimType 屬性) Shard (torchrec.distributed.planner.types 中的類別) shard() (torchrec.distributed.embedding.EmbeddingCollectionSharder 方法) (torchrec.distributed.embeddingbag.EmbeddingBagCollectionSharder 方法) (torchrec.distributed.embeddingbag.EmbeddingBagSharder 方法) (torchrec.distributed.mc_embedding.ManagedCollisionEmbeddingCollectionSharder 方法) (torchrec.distributed.mc_embeddingbag.ManagedCollisionEmbeddingBagCollectionSharder 方法) (torchrec.distributed.mc_modules.ManagedCollisionCollectionSharder 方法) (torchrec.distributed.quant_embeddingbag.QuantEmbeddingBagCollectionSharder 方法) （torchrec.distributed.quant_embeddingbag.QuantFeatureProcessedEmbeddingBagCollectionSharder 方法） (torchrec.distributed.types.ModuleSharder 類別方法) shard_quant_model() (在 torchrec.inference.modules 模組中) shardable_parameters() (torchrec.distributed.embedding.EmbeddingCollectionSharder 方法) （torchrec.distributed.embedding_types.BaseQuantEmbeddingSharder 方法） (torchrec.distributed.embeddingbag.EmbeddingBagCollectionSharder 方法) (torchrec.distributed.embeddingbag.EmbeddingBagSharder 方法) (torchrec.distributed.mc_modules.ManagedCollisionCollectionSharder 方法) （torchrec.distributed.types.ModuleSharder 方法） sharded_model_copy (torchrec.distributed.utils 中的類別) sharded_parameter_names() (torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法) （torchrec.distributed.mc_modules.ShardedManagedCollisionCollection 方法） （torchrec.distributed.types.ShardedModule 方法） ShardedConfig (torchrec.distributed.embedding_types 中的類別) ShardedEmbeddingBag (torchrec.distributed.embeddingbag 中的類別) ShardedEmbeddingBagCollection (torchrec.distributed.embeddingbag 中的類別) ShardedEmbeddingCollection (torchrec.distributed.embedding 中的類別) ShardedEmbeddingModule (torchrec.distributed.embedding_types 中的類別) ShardedEmbeddingTable (torchrec.distributed.embedding_types 中的類別) ShardedManagedCollisionCollection (torchrec.distributed.mc_modules 中的類別) ShardedManagedCollisionEmbeddingBagCollection (torchrec.distributed.mc_embeddingbag 中的類別) ShardedManagedCollisionEmbeddingCollection (torchrec.distributed.mc_embedding 中的類別) ShardedMetaConfig (torchrec.distributed.embedding_types 中的類別) ShardedModule (torchrec.distributed.types 中的類別) ShardedQuantEbcInputDist (torchrec.distributed.quant_embeddingbag 中的類別) ShardedQuantEmbeddingBagCollection (torchrec.distributed.quant_embeddingbag 中的類別) ShardedQuantFeatureProcessedEmbeddingBagCollection (torchrec.distributed.quant_embeddingbag 中的類別) sharder_name() (在 torchrec.distributed.planner.utils 模組中) ShardEstimator (torchrec.distributed.planner.types 中的類別) sharding_contexts (torchrec.distributed.embeddingbag.EmbeddingBagCollectionContext 屬性) sharding_options (torchrec.distributed.planner.partitioners.ShardingOptionGroup 屬性) sharding_spec (torchrec.distributed.types.ParameterSharding 屬性) sharding_type (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.types.ObjectPoolShardingPlan 屬性) (torchrec.distributed.types.ParameterSharding 屬性) sharding_type_device_group_to_sharding_infos() (torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法) sharding_types (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] sharding_types() (torchrec.distributed.embedding.EmbeddingCollectionSharder 方法) (torchrec.distributed.embedding_types.BaseEmbeddingSharder 方法) （torchrec.distributed.embedding_types.BaseQuantEmbeddingSharder 方法） (torchrec.distributed.mc_modules.ManagedCollisionCollectionSharder 方法) （torchrec.distributed.quant_embeddingbag.QuantFeatureProcessedEmbeddingBagCollectionSharder 方法） （torchrec.distributed.types.ModuleSharder 方法） ShardingEnv (torchrec.distributed.types 中的類別) ShardingOption (torchrec.distributed.planner.types 中的類別) ShardingOptionGroup (torchrec.distributed.planner.partitioners 中的類別) ShardingPlan (torchrec.distributed.types 中的類別) ShardingPlanner (torchrec.distributed.types 中的類別) shardings (torchrec.distributed.embedding_types.ModuleShardingMixIn 屬性) (torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 屬性) ShardingType (torchrec.distributed.types 中的類別) (torchrec.modules.embedding_configs 中的類別) shards (torchrec.distributed.planner.types.ShardingOption 屬性) shrink_right() (torchrec.distributed.planner.utils.LuusJaakolaSearch 方法) SimpleDeepFMNN (torchrec.models.deepfm 中的類別) size (torchrec.distributed.embedding_types.DTensorMetadata 屬性) (torchrec.distributed.planner.types.Shard 屬性) SortBy (torchrec.distributed.planner.partitioners 中的類別) sparse_grad_parameter_names() (torchrec.distributed.model_parallel.DistributedModelParallel 方法) SparseArch (torchrec.models.deepfm 中的類別) split() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) splits (torchrec.distributed.embedding_sharding.KJTSplitsAllToAllMeta 屬性) splits_tensors (torchrec.distributed.embedding_sharding.KJTSplitsAllToAllMeta 屬性) SplitsAllToAllAwaitable (torchrec.distributed.dist_data 中的類別), [1] ssd_storage_directory (torchrec.distributed.types.KeyValueParams 屬性), [1] stagger (torchrec.distributed.embedding_sharding.KJTSplitsAllToAllMeta 屬性) state (torchrec.optim.keyed.CombinedOptimizer 屬性) state_dict() (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 方法) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.InferGroupedLookupMixin 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedEmbeddingsLookup 的方法) (torchrec.distributed.embedding_lookup.MetaInferGroupedPooledEmbeddingsLookup 的方法) （torchrec.distributed.embeddingbag.ShardedEmbeddingBag 方法） (torchrec.distributed.grouped_position_weighted.GroupedPositionWeightedModule 的方法) (torchrec.distributed.model_parallel.DistributedModelParallel 的方法) (torchrec.inference.modules.PredictModule 的方法) (torchrec.modules.feature_processor.PositionWeightedProcessor 的方法) (torchrec.optim.keyed.KeyedOptimizer 方法) (torchrec.optim.keyed.OptimizerWrapper 方法) Stats (torchrec.distributed.planner.types 中的類別) stats (torchrec.distributed.types.CacheParams 屬性), [1] STEP (torchrec.optim.warmup.WarmupPolicy 屬性) step() (torchrec.optim.clipping.GradientClippingOptimizer 方法) (torchrec.optim.fused.EmptyFusedOptimizer 方法) (torchrec.optim.fused.FusedOptimizer 方法) (torchrec.optim.keyed.CombinedOptimizer 方法) (torchrec.optim.keyed.KeyedOptimizerWrapper 方法) (torchrec.optim.keyed.OptimizerWrapper 方法) (torchrec.optim.warmup.WarmupOptimizer 方法) stochastic_rounding (torchrec.distributed.planner.types.ParameterConstraints 屬性), [1] (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.types.ParameterSharding 屬性) Storage (torchrec.distributed.planner.types 中的類別) STORAGE (torchrec.distributed.planner.partitioners.SortBy 屬性) storage (torchrec.distributed.planner.types.DeviceHardware 屬性) (torchrec.distributed.planner.types.Shard 屬性) storage_repr_in_gb() (在 torchrec.distributed.planner.utils 模組中) storage_sum (torchrec.distributed.planner.partitioners.ShardingOptionGroup 屬性) storage_usage() (torchrec.distributed.embedding_types.BaseEmbeddingSharder 方法) （torchrec.distributed.embedding_types.BaseQuantEmbeddingSharder 方法） （torchrec.distributed.types.ModuleSharder 方法） StorageReservation (torchrec.distributed.planner.types 中的類別) STRICT_CONSTRAINTS (torchrec.distributed.planner.types.PlannerErrorType 屬性) stride (torchrec.distributed.embedding_types.DTensorMetadata 屬性) stride() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) stride_per_key() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) stride_per_key_per_rank() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) SUM (torchrec.modules.embedding_configs.PoolingType 屬性) supported_fields (torchrec.distributed.planner.types.CustomTopologyData 屬性) SwishLayerNorm (torchrec.modules.activation 中的類別) symbolic_trace() (在 torchrec.fx.tracer 模組中) sync() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) T TABLE_COLUMN_WISE (torchrec.distributed.types.ShardingType 屬性) （torchrec.modules.embedding_configs.ShardingType 屬性） table_names() (torchrec.distributed.embedding_types.GroupedEmbeddingConfig 方法) TABLE_ROW_WISE (torchrec.distributed.types.ShardingType 屬性) （torchrec.modules.embedding_configs.ShardingType 屬性） TABLE_WISE (torchrec.distributed.types.ShardingType 屬性) （torchrec.modules.embedding_configs.ShardingType 屬性） tbes (torchrec.distributed.quant_embeddingbag.ShardedQuantFeatureProcessedEmbeddingBagCollection 屬性) （torchrec.quant.embedding_modules.FeatureProcessedEmbeddingBagCollection 屬性） tbes_configs() (torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 方法) tensor (torchrec.distributed.planner.types.ShardingOption 屬性) (torchrec.distributed.planner.types.ShardingOption 屬性) TensorAllToAll (torchrec.distributed.dist_data 中的類別), [1] TensorAllToAllSplitsAwaitable (torchrec.distributed.dist_data 中的類別), [1] TensorAllToAllValuesAwaitable (torchrec.distributed.dist_data 中的類別), [1] TensorValuesAllToAll (torchrec.distributed.dist_data 中的類別), [1] to() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） （torchrec.sparse.jagged_tensor.KeyedTensor 方法） to_dense() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) to_dense_weights() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) to_dict() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedTensor 方法） to_padded_dense() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) to_padded_dense_weights() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) Topology (torchrec.distributed.planner.types 中的類別) torchrec.distributed 模組 torchrec.distributed.collective_utils 模組 torchrec.distributed.comm 模組 torchrec.distributed.comm_ops 模組 torchrec.distributed.dist_data 模組, [1] torchrec.distributed.embedding 模組 torchrec.distributed.embedding_lookup 模組 torchrec.distributed.embedding_sharding 模組 torchrec.distributed.embedding_types 模組 torchrec.distributed.embeddingbag 模組 torchrec.distributed.grouped_position_weighted 模組 torchrec.distributed.mc_embedding 模組 torchrec.distributed.mc_embeddingbag 模組 torchrec.distributed.mc_modules 模組 torchrec.distributed.model_parallel 模組 torchrec.distributed.planner 模組 torchrec.distributed.planner.constants 模組 torchrec.distributed.planner.enumerators 模組 torchrec.distributed.planner.partitioners 模組 torchrec.distributed.planner.perf_models 模組 torchrec.distributed.planner.planners 模組 torchrec.distributed.planner.proposers 模組 torchrec.distributed.planner.shard_estimators 模組 torchrec.distributed.planner.stats 模組 torchrec.distributed.planner.storage_reservations 模組 torchrec.distributed.planner.types 模組 torchrec.distributed.planner.utils 模組 torchrec.distributed.quant_embeddingbag 模組 torchrec.distributed.sharding 模組 torchrec.distributed.sharding.cw_sharding 模組 torchrec.distributed.sharding.dp_sharding 模組 torchrec.distributed.sharding.rw_sharding 模組 torchrec.distributed.sharding.tw_sharding 模組 torchrec.distributed.sharding.twcw_sharding 模組 torchrec.distributed.sharding.twrw_sharding 模組 torchrec.distributed.train_pipeline 模組 torchrec.distributed.types 模組 torchrec.distributed.utils 模組 torchrec.fx 模組, [1] torchrec.fx.tracer 模組 torchrec.inference 模組, [1] torchrec.inference.model_packager 模組 torchrec.inference.modules 模組 torchrec.models.deepfm 模組 torchrec.modules 模組 torchrec.modules.activation 模組 torchrec.modules.crossnet 模組 torchrec.modules.deepfm 模組 torchrec.modules.embedding_configs 模組 torchrec.modules.embedding_modules 模組 torchrec.modules.feature_processor 模組 torchrec.modules.lazy_extension 模組 torchrec.modules.mc_embedding_modules 模組 torchrec.modules.mc_modules 模組 torchrec.modules.mlp 模組 torchrec.modules.utils 模組 torchrec.optim 模組, [1] torchrec.optim.clipping 模組 torchrec.optim.fused 模組 torchrec.optim.keyed 模組 torchrec.optim.warmup 模組 torchrec.quant 模組, [1] torchrec.quant.embedding_modules 模組 torchrec.sparse 模組, [1] torchrec.sparse.jagged_tensor 模組 torchrec_use_sync_collectives() (在 torchrec.distributed.comm_ops 模組中) total (torchrec.distributed.planner.types.Perf 屬性) total_input_size (torchrec.distributed.comm_ops.ReduceScatterVInfo 屬性), [1] total_perf (torchrec.distributed.planner.types.ShardingOption 屬性) total_storage (torchrec.distributed.planner.types.ShardingOption 屬性) trace() (torchrec.fx.tracer.Tracer 方法) Tracer (torchrec.fx.tracer 中的類別) TRAIN_BASE (torchrec.distributed.types.PipelineType 屬性) TRAIN_PREFETCH_SPARSE_DIST (torchrec.distributed.types.PipelineType 屬性) TRAIN_SPARSE_DIST (torchrec.distributed.types.PipelineType 屬性) training (torchrec.distributed.dist_data.EmbeddingsAllToOne 屬性), [1] (torchrec.distributed.dist_data.EmbeddingsAllToOneReduce 屬性), [1] (torchrec.distributed.dist_data.KJTAllToAll 屬性), [1] (torchrec.distributed.dist_data.KJTOneToAll 屬性), [1] (torchrec.distributed.dist_data.MergePooledEmbeddingsModule 屬性), [1] (torchrec.distributed.dist_data.PooledEmbeddingsAllGather 屬性), [1] (torchrec.distributed.dist_data.PooledEmbeddingsAllToAll 屬性), [1] (torchrec.distributed.dist_data.PooledEmbeddingsReduceScatter 屬性), [1] (torchrec.distributed.dist_data.SeqEmbeddingsAllToOne 屬性), [1] (torchrec.distributed.dist_data.SequenceEmbeddingsAllToAll 屬性), [1] (torchrec.distributed.dist_data.TensorAllToAll 屬性), [1] (torchrec.distributed.dist_data.TensorValuesAllToAll 屬性), [1] (torchrec.distributed.dist_data.VariableBatchPooledEmbeddingsAllToAll 屬性), [1] (torchrec.distributed.dist_data.VariableBatchPooledEmbeddingsReduceScatter 屬性), [1] (torchrec.distributed.embedding.ShardedEmbeddingCollection 屬性) (torchrec.distributed.embedding_lookup.GroupedEmbeddingsLookup 屬性) (torchrec.distributed.embedding_lookup.GroupedPooledEmbeddingsLookup 屬性) (torchrec.distributed.embedding_lookup.InferCPUGroupedEmbeddingsLookup 屬性) (torchrec.distributed.embedding_lookup.InferGroupedEmbeddingsLookup 屬性) (torchrec.distributed.embedding_lookup.InferGroupedPooledEmbeddingsLookup 屬性) (torchrec.distributed.embedding_lookup.MetaInferGroupedEmbeddingsLookup 屬性) (torchrec.distributed.embedding_lookup.MetaInferGroupedPooledEmbeddingsLookup 屬性) (torchrec.distributed.embedding_sharding.BaseEmbeddingDist 屬性) (torchrec.distributed.embedding_sharding.BaseSparseFeaturesDist 屬性) (torchrec.distributed.embedding_types.BaseEmbeddingLookup 屬性) (torchrec.distributed.embedding_types.BaseGroupedFeatureProcessor 屬性) (torchrec.distributed.embedding_types.ShardedEmbeddingModule 屬性) (torchrec.distributed.embeddingbag.ShardedEmbeddingBag 屬性) (torchrec.distributed.embeddingbag.ShardedEmbeddingBagCollection 屬性) (torchrec.distributed.grouped_position_weighted.GroupedPositionWeightedModule 屬性) (torchrec.distributed.mc_embedding.ShardedManagedCollisionEmbeddingCollection 屬性) (torchrec.distributed.mc_embeddingbag.ShardedManagedCollisionEmbeddingBagCollection 屬性) (torchrec.distributed.mc_modules.ShardedManagedCollisionCollection 屬性) (torchrec.distributed.model_parallel.DistributedModelParallel 屬性) (torchrec.distributed.quant_embeddingbag.ShardedQuantEbcInputDist 屬性) (torchrec.distributed.quant_embeddingbag.ShardedQuantEmbeddingBagCollection 屬性) (torchrec.distributed.quant_embeddingbag.ShardedQuantFeatureProcessedEmbeddingBagCollection 屬性) (torchrec.distributed.sharding.cw_sharding.InferCwPooledEmbeddingDist 屬性) (torchrec.distributed.sharding.cw_sharding.InferCwPooledEmbeddingDistWithPermute 屬性) (torchrec.distributed.sharding.dp_sharding.DpPooledEmbeddingDist 屬性) (torchrec.distributed.sharding.dp_sharding.DpSparseFeaturesDist 屬性) (torchrec.distributed.sharding.rw_sharding.InferRwPooledEmbeddingDist 屬性) (torchrec.distributed.sharding.rw_sharding.InferRwSparseFeaturesDist 屬性) (torchrec.distributed.sharding.rw_sharding.RwPooledEmbeddingDist 屬性) (torchrec.distributed.sharding.rw_sharding.RwSparseFeaturesDist 屬性) (torchrec.distributed.sharding.tw_sharding.InferTwPooledEmbeddingDist 屬性) (torchrec.distributed.sharding.tw_sharding.InferTwSparseFeaturesDist 屬性) (torchrec.distributed.sharding.tw_sharding.TwPooledEmbeddingDist 屬性) (torchrec.distributed.sharding.tw_sharding.TwSparseFeaturesDist 屬性) (torchrec.distributed.sharding.twrw_sharding.TwRwPooledEmbeddingDist 屬性) (torchrec.distributed.sharding.twrw_sharding.TwRwSparseFeaturesDist 屬性) (torchrec.distributed.types.ShardedModule 屬性) (torchrec.distributed.utils.CopyableMixin 屬性) (torchrec.inference.modules.PredictModule 屬性) (torchrec.models.deepfm.DenseArch 屬性) (torchrec.models.deepfm.FMInteractionArch 屬性) (torchrec.models.deepfm.OverArch 屬性) (torchrec.models.deepfm.SimpleDeepFMNN 屬性) (torchrec.models.deepfm.SparseArch 屬性) (torchrec.modules.activation.SwishLayerNorm 屬性) (torchrec.modules.crossnet.CrossNet 屬性) (torchrec.modules.crossnet.LowRankCrossNet 屬性) (torchrec.modules.crossnet.LowRankMixtureCrossNet 屬性) (torchrec.modules.crossnet.VectorCrossNet 屬性) (torchrec.modules.deepfm.DeepFM 屬性) (torchrec.modules.deepfm.FactorizationMachine 屬性) (torchrec.modules.embedding_modules.EmbeddingBagCollection 屬性) (torchrec.modules.embedding_modules.EmbeddingBagCollectionInterface 屬性) (torchrec.modules.embedding_modules.EmbeddingCollection 屬性) (torchrec.modules.embedding_modules.EmbeddingCollectionInterface 屬性) (torchrec.modules.feature_processor.BaseFeatureProcessor 屬性) (torchrec.modules.feature_processor.BaseGroupedFeatureProcessor 屬性) (torchrec.modules.feature_processor.PositionWeightedModule 屬性) (torchrec.modules.feature_processor.PositionWeightedProcessor 屬性) (torchrec.modules.mc_embedding_modules.BaseManagedCollisionEmbeddingCollection 屬性) (torchrec.modules.mc_embedding_modules.ManagedCollisionEmbeddingBagCollection 屬性) (torchrec.modules.mc_embedding_modules.ManagedCollisionEmbeddingCollection 屬性) (torchrec.modules.mc_modules.ManagedCollisionCollection 屬性) (torchrec.modules.mc_modules.ManagedCollisionModule 屬性) (torchrec.modules.mc_modules.MCHManagedCollisionModule 屬性) (torchrec.modules.mlp.MLP 屬性) (torchrec.modules.mlp.Perceptron 屬性) (torchrec.quant.embedding_modules.EmbeddingBagCollection 屬性) (torchrec.quant.embedding_modules.EmbeddingCollection 屬性) （torchrec.quant.embedding_modules.FeatureProcessedEmbeddingBagCollection 屬性） (torchrec.sparse.jagged_tensor.ComputeJTDictToKJT 屬性) (torchrec.sparse.jagged_tensor.ComputeKJTToJTDict 屬性) trim_torch_package_prefix_from_typename() (在 torchrec.inference.modules 模組中) TwCwPooledEmbeddingSharding (torchrec.distributed.sharding.twcw_sharding 中的類別) TwPooledEmbeddingDist (torchrec.distributed.sharding.tw_sharding 中的類別) TwPooledEmbeddingSharding (torchrec.distributed.sharding.tw_sharding 中的類別) TwRwPooledEmbeddingDist (torchrec.distributed.sharding.twrw_sharding 中的類別) TwRwPooledEmbeddingSharding (torchrec.distributed.sharding.twrw_sharding 中的類別) TwRwSparseFeaturesDist (torchrec.distributed.sharding.twrw_sharding 中的類別) TwSparseFeaturesDist (torchrec.distributed.sharding.tw_sharding 中的類別) type (torchrec.inference.modules.BatchingMetadata 屬性) U unbucketize_permute_tensor (torchrec.distributed.embedding_types.InputDistOutputs 屬性) uncombined_embedding_dims() (torchrec.distributed.embedding_sharding.EmbeddingSharding 方法) （torchrec.distributed.sharding.cw_sharding.BaseCwEmbeddingSharding 方法） uncombined_embedding_names() (torchrec.distributed.embedding_sharding.EmbeddingSharding 方法) （torchrec.distributed.sharding.cw_sharding.BaseCwEmbeddingSharding 方法） uneven_sharding_perf_multiplier (torchrec.distributed.planner.types.Topology 屬性) unflatten_kjt_list() (在 torchrec.sparse.jagged_tensor 模組中) UNIFORM (torchrec.distributed.planner.types.PartitionByType 屬性) uniform() (torchrec.distributed.planner.utils.LuusJaakolaSearch 方法) UniformProposer (torchrec.distributed.planner.proposers 中的類別) unpadded_lengths (torchrec.modules.utils.SequenceVBEContext 屬性) unsync() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) update_metadata_and_generate_eviction_scores() (torchrec.modules.mc_modules.DistanceLFU_EvictionPolicy 方法) (torchrec.modules.mc_modules.LFU_EvictionPolicy 方法) （torchrec.modules.mc_modules.LRU_EvictionPolicy 方法） （torchrec.modules.mc_modules.MCHEvictionPolicy 方法） V VALUE (torchrec.optim.clipping.GradientClipping 屬性) value (torchrec.optim.warmup.WarmupStage 屬性) values() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） （torchrec.sparse.jagged_tensor.KeyedTensor 方法） Variable_Batch_All2All_Pooled_Req (torchrec.distributed.comm_ops 中的類別) variable_batch_all2all_pooled_sync() (在 torchrec.distributed.comm_ops 模組中) Variable_Batch_All2All_Pooled_Wait (torchrec.distributed.comm_ops 中的類別) variable_batch_alltoall_pooled() (在 torchrec.distributed.comm_ops 模組中) variable_batch_per_feature (torchrec.distributed.embeddingbag.EmbeddingBagCollectionContext 屬性) variable_batch_size (torchrec.distributed.comm_ops.All2AllSequenceInfo 屬性), [1] variable_stride_per_key() (torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法) VariableBatchAll2AllPooledInfo (torchrec.distributed.comm_ops 中的類別) VariableBatchEmbeddingBagCollectionAwaitable (torchrec.distributed.embeddingbag 中的類別) VariableBatchPooledEmbeddingsAllToAll (torchrec.distributed.dist_data 中的類別), [1] VariableBatchPooledEmbeddingsReduceScatter (torchrec.distributed.dist_data 中的類別), [1] VectorCrossNet (torchrec.modules.crossnet 中的類別) W wait() (torchrec.distributed.types.Awaitable 方法) WarmupOptimizer (torchrec.optim.warmup 中的類別) WarmupPolicy (torchrec.optim.warmup 中的類別) WarmupStage (torchrec.optim.warmup 中的類別) weight (torchrec.modules.embedding_configs.QuantConfig 屬性) weight_init_max (torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性) weight_init_min (torchrec.modules.embedding_configs.BaseEmbeddingConfig 屬性) weighted_feature_bwd_compute_multiplier (torchrec.distributed.planner.types.Topology 屬性) weights() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） weights_or_none() (torchrec.sparse.jagged_tensor.JaggedTensor 方法) （torchrec.sparse.jagged_tensor.KeyedJaggedTensor 方法） world_size (torchrec.distributed.planner.types.Topology 屬性) wrap() (torchrec.distributed.model_parallel.DataParallelWrapper 方法) (torchrec.distributed.model_parallel.DefaultDataParallelWrapper 方法) Z zero_grad() (torchrec.optim.fused.EmptyFusedOptimizer 方法) (torchrec.optim.fused.FusedOptimizer 方法) (torchrec.optim.keyed.CombinedOptimizer 方法) (torchrec.optim.keyed.KeyedOptimizerWrapper 方法) (torchrec.optim.keyed.OptimizerWrapper 方法)
