API Reference#
This page contains auto-generated API reference documentation [1].
corecore.tokenizerscore.tokenizers.textcore.tokenizers.text.librariescore.tokenizers.text.libraries.bytelevel_tokenizercore.tokenizers.text.libraries.null_tokenizercore.tokenizers.text.libraries.huggingface_tokenizercore.tokenizers.text.libraries.abstract_tokenizercore.tokenizers.text.libraries.sft_tokenizercore.tokenizers.text.libraries.megatron_hf_tokenizercore.tokenizers.text.libraries.tiktoken_tokenizercore.tokenizers.text.libraries.sentencepiece_tokenizercore.tokenizers.text.libraries.chat_template
core.tokenizers.text.parserscore.tokenizers.text.text_tokenizer
core.tokenizers.utilscore.tokenizers.visioncore.tokenizers.base_tokenizercore.tokenizers.megatron_tokenizer
core.pipeline_parallelcore.pipeline_parallel.p2p_communicationcore.pipeline_parallel.utilscore.pipeline_parallel.schedulescore.pipeline_parallel.fine_grained_activation_offloadcore.pipeline_parallel.combined_1f1bcore.pipeline_parallel.hybrid_cp_schedulecore.pipeline_parallel.bridge_communicatorcore.pipeline_parallel.multimodule_communicator
core.ssmcore.ssm.gated_delta_netcore.ssm.opscore.ssm.context_parallelcore.ssm.mamba_hybrid_layer_allocationcore.ssm.packed_seq_helperscore.ssm.mamba_layercore.ssm.ssm_inferencecore.ssm.mamba_context_parallelcore.ssm.triton_cache_managercore.ssm.mamba_mixercore.ssm.utilscore.ssm.mamba_layer_configcore.ssm.mlp_layer_configcore.ssm.gdp_context_parallelcore.ssm.gated_delta_productcore.ssm.gdn_layer_configcore.ssm.mlp_layercore.ssm.causal_conv1dcore.ssm.mamba_block
core.exportcore.export.trtllmcore.export.export_configcore.export.model_typecore.export.data_type
core.inferencecore.inference.text_generation_servercore.inference.text_generation_server.dynamic_text_gen_servercore.inference.text_generation_server.dynamic_text_gen_server.endpointscore.inference.text_generation_server.dynamic_text_gen_server.endpoints.commoncore.inference.text_generation_server.dynamic_text_gen_server.endpoints.chat_completionscore.inference.text_generation_server.dynamic_text_gen_server.endpoints.profilecore.inference.text_generation_server.dynamic_text_gen_server.endpoints.healthcore.inference.text_generation_server.dynamic_text_gen_server.endpoints.completions
core.inference.text_generation_server.dynamic_text_gen_server.text_generation_servercore.inference.text_generation_server.dynamic_text_gen_server.image_preprocessingcore.inference.text_generation_server.dynamic_text_gen_server.incremental_detokenizercore.inference.text_generation_server.dynamic_text_gen_server.vlm_dynamic_inferencecore.inference.text_generation_server.dynamic_text_gen_server.openai_streamingcore.inference.text_generation_server.dynamic_text_gen_server.tokenization
core.inference.text_generation_server.text_generation_servercore.inference.text_generation_server.run_mcore_enginecore.inference.text_generation_server.tokenization
core.inference.moecore.inference.contextscore.inference.contexts.base_contextcore.inference.contexts.dynamic_contextcore.inference.contexts.gpu_viewcore.inference.contexts.mamba_slot_allocatorcore.inference.contexts.kv_block_allocatorcore.inference.contexts.static_contextcore.inference.contexts.mtp_metadatacore.inference.contexts.fused_kv_append_kernelcore.inference.contexts.mtp_context_mixincore.inference.contexts.routing_metadata
core.inference.text_generation_controllerscore.inference.text_generation_controllers.vlm_text_generation_controllercore.inference.text_generation_controllers.mtp_controller_mixincore.inference.text_generation_controllers.encoder_decoder_text_generation_controllercore.inference.text_generation_controllers.text_generation_controllercore.inference.text_generation_controllers.mtp_utils_pytorchcore.inference.text_generation_controllers.mtp_utils_triton
core.inference.apiscore.inference.quantizationcore.inference.samplingcore.inference.data_parallel_inference_coordinatorcore.inference.enginescore.inference.disaggregationcore.inference.disaggregation.transfer_backendscore.inference.disaggregation.ssm_reshardcore.inference.disaggregation.pending_handoff_importscore.inference.disaggregation.utilscore.inference.disaggregation.inference_state_handoffcore.inference.disaggregation.enginecore.inference.disaggregation.kv_reshardcore.inference.disaggregation.handoff_completion_trackercore.inference.disaggregation.decode_admission
core.inference.model_inference_wrapperscore.inference.shards_speccore.inference.inference_requestcore.inference.configcore.inference.headerscore.inference.symmetric_memorycore.inference.utilscore.inference.communication_utilscore.inference.async_streamcore.inference.batch_dimensions_utilscore.inference.common_inference_paramscore.inference.inference_clientcore.inference.shardscore.inference.unified_memorycore.inference.schedulercore.inference.sampling_params
core.dist_checkpointingcore.dist_checkpointing.strategiescore.dist_checkpointing.optimizercore.dist_checkpointing.dict_utilscore.dist_checkpointing.validationcore.dist_checkpointing.tensor_aware_state_dictcore.dist_checkpointing.gpt_checkpoint_interopcore.dist_checkpointing.serializationcore.dist_checkpointing.utilscore.dist_checkpointing.corecore.dist_checkpointing.exchange_utilscore.dist_checkpointing.mappingcore.dist_checkpointing.state_dict_utils
core.post_trainingcore.allocatorcore.transformercore.transformer.moecore.transformer.moe.opscore.transformer.moe.token_dispatcher_inferencecore.transformer.moe.router_diagnosticscore.transformer.moe.inference_routing_mask_kernelcore.transformer.moe.moe_utilscore.transformer.moe.token_dispatchercore.transformer.moe.upcycling_utilscore.transformer.moe.moe_layer_configcore.transformer.moe.fused_a2acore.transformer.moe.router_tracecore.transformer.moe.routercore.transformer.moe.shared_expertscore.transformer.moe.moe_layercore.transformer.moe.paged_stashcore.transformer.moe.expertscore.transformer.moe.moe_loggingcore.transformer.moe.router_replaycore.transformer.moe.batch_invariant
core.transformer.custom_layerscore.transformer.transformer_blockcore.transformer.modulecore.transformer.multi_token_predictioncore.transformer.transformer_layercore.transformer.hyper_connectioncore.transformer.pipeline_parallel_layer_layoutcore.transformer.spec_utilscore.transformer.mla_qk_norm_configcore.transformer.torch_layer_normcore.transformer.transformer_configcore.transformer.mlpcore.transformer.utilscore.transformer.multi_latent_attentioncore.transformer.attention_layer_configcore.transformer.cuda_graph_configcore.transformer.enumscore.transformer.identity_opcore.transformer.fsdp_dtensor_checkpointcore.transformer.torch_normcore.transformer.mla_layer_configcore.transformer.cuda_graphscore.transformer.attentioncore.transformer.dot_product_attention
core.datasetscore.datasets.gpt_datasetcore.datasets.bert_datasetcore.datasets.multimodal_datasetcore.datasets.t5_datasetcore.datasets.object_storage_utilscore.datasets.blended_datasetcore.datasets.utils_s3core.datasets.helperscore.datasets.data_schedulecore.datasets.utilscore.datasets.data_schedule_utilscore.datasets.megatron_datasetcore.datasets.indexed_datasetcore.datasets.blended_megatron_dataset_buildercore.datasets.masked_datasetcore.datasets.blended_megatron_dataset_config
core.context_parallelcore.distributedcore.distributed.fsdpcore.distributed.fsdp.srccore.distributed.fsdp.src.megatron_fsdpcore.distributed.fsdp.src.megatron_fsdp.experimentalcore.distributed.fsdp.src.megatron_fsdp.experimental.dbuffercore.distributed.fsdp.src.megatron_fsdp.experimental.uneven_dtensorcore.distributed.fsdp.src.megatron_fsdp.experimental.optimizercore.distributed.fsdp.src.megatron_fsdp.experimental.modulecore.distributed.fsdp.src.megatron_fsdp.experimental.countdowncore.distributed.fsdp.src.megatron_fsdp.experimental.quantized_dbuffercore.distributed.fsdp.src.megatron_fsdp.experimental.owner_planningcore.distributed.fsdp.src.megatron_fsdp.experimental.parameter_groupcore.distributed.fsdp.src.megatron_fsdp.experimental.schedulecore.distributed.fsdp.src.megatron_fsdp.experimental.module_utilscore.distributed.fsdp.src.megatron_fsdp.experimental.placementcore.distributed.fsdp.src.megatron_fsdp.experimental.checkpointcore.distributed.fsdp.src.megatron_fsdp.experimental.layoutcore.distributed.fsdp.src.megatron_fsdp.experimental.fully_shardcore.distributed.fsdp.src.megatron_fsdp.experimental.indexed_order
core.distributed.fsdp.src.megatron_fsdp.distributed_data_parallel_configcore.distributed.fsdp.src.megatron_fsdp.uneven_dtensorcore.distributed.fsdp.src.megatron_fsdp.utilscore.distributed.fsdp.src.megatron_fsdp.mixed_precisioncore.distributed.fsdp.src.megatron_fsdp.param_and_grad_buffercore.distributed.fsdp.src.megatron_fsdp.megatron_fsdpcore.distributed.fsdp.src.megatron_fsdp.fully_shardcore.distributed.fsdp.src.megatron_fsdp.package_info
core.distributed.fsdp.mcore_fsdp_adapter
core.distributed.distributed_data_parallel_configcore.distributed.finalize_model_gradscore.distributed.torch_fully_sharded_data_parallel_configcore.distributed.reduce_scatter_with_fp32_accumulationcore.distributed.param_and_grad_buffercore.distributed.data_parallel_basecore.distributed.torch_fully_sharded_data_parallelcore.distributed.distributed_data_parallel
core.quantizationcore.extensionscore.optimizercore.optimizer.cpu_offloadingcore.optimizer.distrib_optimizercore.optimizer.optimizercore.optimizer.optimizer_configcore.optimizer.optimizer_cuda_graphcore.optimizer.fully_sharded_optimizercore.optimizer.muoncore.optimizer.qk_clipcore.optimizer.clip_gradscore.optimizer.emerging_optimizerscore.optimizer.param_layoutcore.optimizer.layer_wise_optimizercore.optimizer.grad_scaler
core.telemetrycore.modelscore.models.T5core.models.visioncore.models.mambacore.models.multimodalcore.models.huggingfacecore.models.mimocore.models.hybridcore.models.gptcore.models.bertcore.models.commoncore.models.audiocore.models.audio.audio_processorcore.models.audio.nemo_audio_checkpointcore.models.audio.nemo_transformer_encodercore.models.audio.audio_feature_configcore.models.audio.nemo_audio_preprocessingcore.models.audio.audio_projectorcore.models.audio.packed_audiocore.models.audio.nemo_transformer_audio_model
core.models.backends
core.tensor_parallelcore.tensor_parallel.gtp_apicore.tensor_parallel.layerscore.tensor_parallel.datacore.tensor_parallel.generalized_tensor_parallelismcore.tensor_parallel.inference_layerscore.tensor_parallel.cross_entropycore.tensor_parallel.utilscore.tensor_parallel.gtp_symmetric_memorycore.tensor_parallel.mappingscore.tensor_parallel.gtp_cuda_graphscore.tensor_parallel.random
core.reshardingcore.resharding.nvshmem_copy_servicecore.resharding.nvshmem_copy_service.planningcore.resharding.nvshmem_copy_service.corecore.resharding.nvshmem_copy_service.memorycore.resharding.nvshmem_copy_service.nvshmem_typescore.resharding.nvshmem_copy_service.compatcore.resharding.nvshmem_copy_service.loggercore.resharding.nvshmem_copy_service.servicecore.resharding.nvshmem_copy_service.validation
core.resharding.copy_servicescore.resharding.executioncore.resharding.plannercore.resharding.utilscore.resharding.refitcore.resharding.shard_plannercore.resharding.transforms
core.fusionscore.fusions.fused_weighted_squared_relucore.fusions.fused_mhc_kernelscore.fusions.fused_bias_gelucore.fusions.fused_bias_dropoutcore.fusions.fused_bias_swiglucore.fusions.fused_layer_normcore.fusions.fused_bias_geglucore.fusions.fused_cross_entropycore.fusions.fused_indices_convertercore.fusions.fused_softmaxcore.fusions.fused_pad_routing_mapcore.fusions.fused_mla_yarn_rope_apply
core.tensor_observationcore.config_loggercore._slurm_utilscore.parallel_statecore.msc_utilscore.timerscore.configcore.typed_torchcore.process_groups_configcore.num_microbatches_calculatorcore.nccl_allocatorcore.optimizer_param_schedulercore.fp4_utilscore.parameter_namescore.fp8_utilscore._rank_utilscore.model_parallel_configcore.utilscore.energy_monitorcore.full_cuda_graphcore.enumscore.packed_seq_paramscore.hyper_comm_gridcore.activationscore.inference_paramscore.safe_globalscore.recomputecore.fault_injectorcore.rerun_state_machinecore.package_infocore.jit