MCPcopy Create free account

hub / github.com/deepspeedai/DeepSpeed / functions

Functions10,922 in github.com/deepspeedai/DeepSpeed

⨍ Functions 10,922◇ Types & classes 1,850↳ Endpoints 34
↓ 3 callersFunction_quantize_module
(model, policies)
deepspeed/module_inject/module_quantize.py:63
↓ 3 callersMethod_raise_if_autoep_zero3_consolidated_export
(self, operation)
deepspeed/runtime/engine.py:5278
↓ 3 callersMethod_reassign_or_swap_out_partitioned_parameters_async
Asynchronously update partitioned parameters with optimized values.
deepspeed/runtime/superoffload/superoffload_stage3.py:129
↓ 3 callersMethod_recursive_getattr
Allow getting an attribute like "linear.weight
deepspeed/runtime/pipe/module.py:447
↓ 3 callersMethod_release_tensor
(self, tensor_id)
deepspeed/runtime/zero/contiguous_memory_allocator.py:147
↓ 3 callersMethod_replace_with_config
Replace layer using the new configurable AutoTP API. This method uses TPLayerSpec to determine how to partition the layer.
deepspeed/module_inject/auto_tp.py:411
↓ 3 callersMethod_residual_add_bias
(hidden_state, residual, attention_output, attention_bias, final_bias, mp_size, mlp
op_builder/npu/inference.py:255
↓ 3 callersFunction_router_params
(engine)
tests/unit/v1/moe/test_autoep_checkpoint.py:135
↓ 3 callersFunction_run_folded_checkpoint_same_topology_resume
(checkpoint_dir, *, ep_size=2, mixed_precision=True)
tests/unit/v1/moe/test_autoep_autotp_checkpoint.py:256
↓ 3 callersFunction_run_multi_backward
(config_dict, hidden_dim, num_chunks, num_steps=1, seed=42)
tests/unit/v1/zero/test_zero2_offload_multi_backward.py:64
↓ 3 callersMethod_run_test
(self, tmpdir, dtype, ds_config, load_optim, use_torch_adam, world_size)
tests/unit/checkpoint/test_universal_checkpoint.py:216
↓ 3 callersMethod_set_client_model
(self, model)
deepspeed/runtime/engine.py:1620
↓ 3 callersFunction_set_module_by_name
Replace the submodule at *dotted_name* with *new_module* in-place.
deepspeed/sequence/auto_sp.py:134
↓ 3 callersMethod_setup_baseline_linears
(self, model)
tests/unit/model_parallelism/test_tp_plan_e2e.py:50
↓ 3 callersMethod_softmax_context
(query_key_value, attn_mask, rotary_dim, rotate_half, rotate_every_two, heads, num_kv,
op_builder/npu/inference.py:114
↓ 3 callersFunction_source_param_shape
(param: torch.Tensor | nn.Parameter)
deepspeed/module_inject/auto_ep.py:85
↓ 3 callersMethod_stop_timers
(self, timer_names)
deepspeed/runtime/engine.py:3364
↓ 3 callersMethod_sycl_env_paths
Find the SYCL include and lib directories from the Python environment. When using PyTorch XPU wheels, libsycl.so and SYCL headers are
op_builder/xpu/builder.py:51
↓ 3 callersMethod_sync_cpu_optimizer_step
(self, param_group_id: int, sub_group_id: in
deepspeed/runtime/superoffload/superoffload_stage3.py:314
↓ 3 callersMethod_test_set_z3_leaf_modules
(self, cls, requires_grad)
tests/unit/runtime/zero/test_zero_leaf_module.py:327
↓ 3 callersMethod_unflatten_partitioned_parameters
(self, sub_group_id)
deepspeed/runtime/zero/stage3.py:2449
↓ 3 callersMethod_update_hp_grad
(self, lp, group_idx, param_idx, clear_lp_grads)
deepspeed/runtime/bf16_optimizer.py:356
↓ 3 callersMethod_update_param_state_info
(self, swap_info, parameter)
deepspeed/runtime/swap_tensor/optimizer_utils.py:486
↓ 3 callersMethod_valid_micro_batch
(self, micro_batch_id)
deepspeed/runtime/pipe/schedule.py:80
↓ 3 callersMethod_validate_autoep_zero3_partitioned_metadata
(autoep_layers, model=None, require_partitioned=True)
deepspeed/runtime/engine.py:4184
↓ 3 callersMethod_vector_matmul
(input, weight, async_op, q_scale, q_int8, transposed_mode)
op_builder/npu/inference.py:196
↓ 3 callersFunction_verify_continuous_decrease
(values)
tests/unit/runtime/test_lr_schedulers.py:22
↓ 3 callersMethod_wildcard_to_regex
(pattern: str)
deepspeed/module_inject/tp_plan_converter.py:56
↓ 3 callersMethod_writeback_swap_state
(self, sub_group_id, write_opt_state, write_gradients)
deepspeed/runtime/zero/stage3.py:2444
↓ 3 callersMethod_zero3_partition_group_metadata
(self)
deepspeed/runtime/zero/stage3.py:3087
↓ 3 callersFunction_zero_optimizer_stub
(*, partition_gradients, overlap_comm)
tests/unit/v1/moe/test_autoep_autotp_zero1_overlap_folding.py:15
↓ 3 callersMethodabsolute_name
Returns absolute build path for cases where the op is pre-installed, e.g., deepspeed.ops.adam.cpu_adam will be installed as something
op_builder/builder.py:132
↓ 3 callersFunctionadam_update
(grad, buf1, buf2, step, betas, eps)
deepspeed/runtime/zero/muon/original_muon.py:267
↓ 3 callersFunctionadd_end_backward
(graph: Graph, graph_id: int, release_reduce_buckets: bool = True)
deepspeed/compile/fx.py:28
↓ 3 callersMethodadd_pointer_offset
Adds a pointer offset in units of Element
csrc/deepspeed4science/evoformer_attn/iterators/predicated_tile_iterator_atomic.h:330
↓ 3 callersMethodall_gather_coalesced
(params: Iterable[Parameter], safe_mode: bool = False,
deepspeed/runtime/zero/partition_parameters.py:1446
↓ 3 callersMethodall_gather_coalesced
(self, output_tensors, input_tensors, group=None, async_op=False)
deepspeed/comm/torch.py:287
↓ 3 callersFunctionall_to_all_quant_reduce
(tensors: List[Tensor], groups: {})
deepspeed/runtime/comm/coalesced_collectives.py:31
↓ 3 callersFunctionallclose
(x, y)
tests/unit/ops/spatial/test_nhwc_bias_add.py:17
↓ 3 callersMethodallocate_tensor
(self, swap_path, numel, aligned_numel)
deepspeed/runtime/swap_tensor/utils.py:56
↓ 3 callersMethodallreduce_and_copy_with_multiple_ranks
(self, small_bucket,
deepspeed/runtime/zero/stage_1_and_2.py:1217
↓ 3 callersMethodallreduce_gradients
(self, bucket_size=MEMORY_OPT_ALLREDUCE_SIZE)
deepspeed/runtime/engine.py:2756
↓ 3 callersFunctionapply_autotp_with_partition_config
(model, tp_size, partition_config)
tests/unit/model_parallelism/test_autotp_custom_patterns.py:156
↓ 3 callersMethodapply_rotary_pos_emb
(mixed_query, key_layer, rotary_dim, offset, num_heads, rotate_half, rope_theta)
op_builder/supa/inference.py:299
↓ 3 callersFunctionassert_ints_same_as_other_ranks
NOTE: creates both communication and synchronization overhead so should be used sparingly takes a list of ints from each rank and ensure
deepspeed/runtime/zero/utils.py:106
↓ 3 callersFunctionassert_load_balance_coeff_rejection_message
(exc: BaseException, value: object)
tests/unit/v1/moe/autoep_test_utils.py:173
↓ 3 callersMethodattention_o_mp
(self, mp_replace, reversed_dim=False)
deepspeed/module_inject/containers/base.py:249
↓ 3 callersMethodautotuning_enabled
(self)
deepspeed/runtime/engine.py:1042
↓ 3 callersFunctionbaseline_implementation
(hidden_states: torch.Tensor, seq_lens: List[int])
tests/unit/inference/v2/kernels/ragged_ops/test_logits_gather.py:17
↓ 3 callersMethodbf16_optimizer_states
(self)
deepspeed/runtime/engine.py:1255
↓ 3 callersMethodbuild_atoms
Build the atoms for this module. This is not a strict requirement for the class, so this method is a no-op by default rather than abs
deepspeed/inference/v2/modules/interfaces/attention_base.py:51
↓ 3 callersFunctioncan_send_recv
()
deepspeed/runtime/pipe/p2p.py:21
↓ 3 callersFunctioncheck_deepspeed_config
(config)
deepspeed/compression/compress.py:20
↓ 3 callersMethodcheck_using_norm
(self, norm_group, reduce_overflow=True)
deepspeed/runtime/utils.py:241
↓ 3 callersMethodcheckpoint_event_epilogue
(self)
deepspeed/runtime/zero/stage3.py:3502
↓ 3 callersMethodcheckpoint_event_prologue
(self)
deepspeed/runtime/zero/stage3.py:3499
↓ 3 callersMethodclear_grad_attribute
(self, param)
deepspeed/runtime/zero/stage_1_and_2.py:1078
↓ 3 callersMethodclear_lp_grads
(self, set_to_none=False)
deepspeed/runtime/bf16_optimizer.py:473
↓ 3 callersMethodclear_selected_mv
(self)
deepspeed/ops/adam/zenflow_torch_adam.py:83
↓ 3 callersFunctionclone_tensors_for_torch_save
Returns a copy of ``item`` with all enclosed tensors replaced by clones on a specified device. Works on individual tensors, and tensors conta
deepspeed/checkpoint/utils.py:41
↓ 3 callersFunctioncombine_from_routed
Scatter-add expert outputs back to original token positions.
deepspeed/module_inject/auto_ep_layer.py:331
↓ 3 callersMethodcommit
(self, info: CheckpointCommitInfo)
deepspeed/runtime/checkpoint_engine/checkpoint_engine.py:43
↓ 3 callersMethodcompile_autosp
Determines if AutoSP is set in deepcompile's passes attributes.
deepspeed/runtime/engine.py:1160
↓ 3 callersFunctionconcat_to_compat_param
(quantized_weight: Tensor, quant_scale: Tensor, quant_mi
deepspeed/inference/quantization/utils.py:186
↓ 3 callersMethodconfig_requires_grad
(self, weight)
deepspeed/module_inject/layers.py:314
↓ 3 callersFunctionconvert_size
(size_bytes)
deepspeed/utils/comms_logging.py:22
↓ 3 callersMethodconvert_to_zero_parameters
(param_list)
deepspeed/runtime/zero/partition_parameters.py:1566
↓ 3 callersFunctioncount_used_parameters_in_backward
Count the number of parameters that participate in the currently running backward graph. This helper is designed to be invoked from within a
deepspeed/runtime/utils.py:1444
↓ 3 callersMethodcurrent_device
(self)
tests/unit/v1/compile/test_graph_profile.py:30
↓ 3 callersMethodcxx_args
Returns optional list of compiler flags to forward to the build
op_builder/builder.py:301
↓ 3 callersFunctiondebug_module2name
(module)
deepspeed/utils/debug.py:32
↓ 3 callersFunctiondebug_param2name_id_shape
(param)
deepspeed/utils/debug.py:72
↓ 3 callersMethoddequantize
(self, input_q, fp_out=None, q_bits=8, q_mantisa_bits=3, scale=None)
deepspeed/ops/fp_quantizer/quantize.py:39
↓ 3 callersMethoddequantize
(self, input_q, fp_out=None, q_bits=8, q_mantisa_bits=3, scale=None)
deepspeed/ops/fp_quantizer/quantize.py:110
↓ 3 callersMethoddequantized
Return a tensor containing the dequantized weights of this parameter.
deepspeed/linear/quantization.py:67
↓ 3 callersFunctiondimension_problem
csrc/spatial/csrc/pt_binding.cpp:12
↓ 3 callersMethoddisable
(self)
deepspeed/linear/optimized_linear.py:120
↓ 3 callersFunctiondispatch_counters
(ctx: RestoreContext)
deepspeed/moe/ep_tp_dispatch.py:451
↓ 3 callersFunctionds_shape
(param)
deepspeed/utils/debug.py:61
↓ 3 callersFunctiondummy_init_engine
(config)
tests/unit/model_parallelism/test_autotp_training.py:582
↓ 3 callersMethoddynamic_loss_scale_args
(self)
deepspeed/runtime/engine.py:1361
↓ 3 callersFunctioneinsum
(rule, a, b)
deepspeed/moe/sharded_moe.py:119
↓ 3 callersMethodelapsed
(self)
tests/unit/modeling.py:110
↓ 3 callersMethodelasticity_enabled
(self)
deepspeed/runtime/engine.py:900
↓ 3 callersFunctionenable
Decorator factory to enable compiling of a function if the minimum PyTorch version requirement is met. Args: min_version (str, optio
deepspeed/runtime/compiler.py:42
↓ 3 callersMethodend_profile
Ends profiling. The added attributes and handles are removed recursively on all the modules.
deepspeed/profiling/flops_profiler/profiler.py:203
↓ 3 callersMethodeval
r
deepspeed/runtime/engine.py:2552
↓ 3 callersMethodextra_ldflags
(self)
op_builder/builder.py:313
↓ 3 callersFunctionfill_autoep_config_from_hf
Back-fill AutoEPConfig fields from HF model config when user hasn't set them. HF field names (e.g. n_group, topk_group, routed_scaling_factor) di
deepspeed/module_inject/auto_ep_config.py:330
↓ 3 callersFunctionfill_with_head_ids
Fills a tensor with the associated head ids. All columns should have the same value.
tests/unit/inference/v2/model_implementations/sharding/test_attn_out_sharding.py:16
↓ 3 callersFunctionfilter_args
(node: Node)
deepspeed/compile/list_schedule.py:48
↓ 3 callersFunctionfind_fit_int_dtype
(min_value, max_value)
deepspeed/runtime/data_pipeline/data_sampling/utils.py:12
↓ 3 callersMethodfind_matching_spec
Find the first matching spec for a parameter.
deepspeed/module_inject/autotp_config.py:256
↓ 3 callersFunctionfind_node_by_tag
(gm: GraphModule, tag: str)
deepspeed/compile/fx.py:178
↓ 3 callersMethodfork_rng
(self, devices)
tests/unit/v1/compile/test_graph_profile.py:15
↓ 3 callersMethodforward
(self, input)
tests/unit/runtime/zero/test_zero_context_return.py:88
↓ 3 callersMethodfp16_enabled
(self)
deepspeed/autotuning/autotuner.py:265
↓ 3 callersMethodfree
csrc/aio/py_lib/deepspeed_pin_tensor.cpp:42
↓ 3 callersMethodfrom_dict
Constructs a `BertConfig` from a Python dictionary of parameters.
tests/unit/modeling.py:238
↓ 3 callersMethodgK_strideM
csrc/deepspeed4science/evoformer_attn/kernel_backward.h:270
← previousnext →901–1,000 of 10,922, ranked by callers