MCPcopy Create free account

hub / github.com/Gen-Verse/OpenClaw-RL / functions

Functions10,561 in github.com/Gen-Verse/OpenClaw-RL

↓ 1 callersMethod_get_current_node_ip_and_free_port
(start_port=10000, consecutive=1)
slime/slime/ray/ray_actor.py:6
↓ 1 callersMethod_get_enum_kwargs
Build kwargs for Enums. With these settings, the user must provide a valid enum value, e.g. 'flash', for `AttnBackend.flash`.
Megatron-LM/megatron/training/argument_utils.py:85
↓ 1 callersFunction_get_exec
(url: str, timeout: int = 30)
swe-rl/server/swe_env_pool_server.py:40
↓ 1 callersMethod_get_extra_state_dict
(self)
Megatron-LM/tests/unit_tests/dist_checkpointing/test_strict.py:67
↓ 1 callersFunction_get_extra_state_offsets
Turns ShardedTensor offsets into offsets suitable for ShardedObject.
Megatron-LM/megatron/core/transformer/utils.py:178
↓ 1 callersMethod_get_field_docstrings
Extract field-level docstrings from a dataclass by inspecting its AST. Recurses on parent classes of `src_cfg_class`. Args:
Megatron-LM/megatron/training/argument_utils.py:209
↓ 1 callersMethod_get_fp8_params_and_shard_fp32_from_fp8
Get lists of FP8 model params, corresponding shard main params, and the starting index of the shard main param in the FP8 param. Para
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:2319
↓ 1 callersFunction_get_fsdp_tensor_spec
Get the DeviceMesh for the parameter and modify the placement for Megatron-FSDP.
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/param_and_grad_buffer.py:3786
↓ 1 callersMethod_get_gbuf_name
(self, buffer_id: int)
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/param_and_grad_buffer.py:626
↓ 1 callersFunction_get_generation_prompt_suffix
Extract suffix after the last '<|im_start|>assistant\\n' in sample.prompt. sample.prompt is produced by tokenizer.apply_chat_template with the us
toolcall-rl/generate_with_retool.py:34
↓ 1 callersMethod_get_global_min_max_time_string
Report strings for max/minmax times across all ranks.
Megatron-LM/megatron/core/timers.py:327
↓ 1 callersFunction_get_harness_tools
(instance: dict)
swe-rl/generate_with_swe_trajectory.py:385
↓ 1 callersFunction_get_heterogenous_attention_spec
( attn_config: AttentionConfig, use_te: bool, qk_layernorm: bool, normalization: str )
Megatron-LM/megatron/core/models/gpt/heterogeneous/heterogeneous_layer_specs.py:95
↓ 1 callersFunction_get_heterogenous_mlp_spec
(mlp_config: MLPConfig, use_te: bool)
Megatron-LM/megatron/core/models/gpt/heterogeneous/heterogeneous_layer_specs.py:123
↓ 1 callersFunction_get_hsdp_tp_mesh
(outer_fsdp_dp_group, dp_cp_group, tp_group)
Megatron-LM/megatron/core/distributed/fsdp/mcore_fsdp_adapter.py:311
↓ 1 callersMethod_get_layer
(self, layer_number: int)
Megatron-LM/megatron/core/transformer/transformer_block.py:419
↓ 1 callersFunction_get_layer_type
(model_type, default_layer_type, retro_layer_numbers, layer_number)
Megatron-LM/megatron/legacy/model/transformer.py:1367
↓ 1 callersFunction_get_libreoffice_version
Function to get the LibreOffice version as a tuple of integers.
gui-rl/desktop_env/server/main.py:416
↓ 1 callersMethod_get_lr_for_group
Compute learning rate for a specific parameter group. Args: param_group (dict): parameter group from the optimizer. Retu
slime/slime/backends/fsdp_utils/lr_scheduler.py:82
↓ 1 callersMethod_get_ltor_masks_and_position_ids
Build masks and position id for left to right model for SFT
Megatron-LM/megatron/training/datasets/sft_dataset.py:124
↓ 1 callersFunction_get_machine_architecture
Get the machine architecture, e.g., x86_64, arm64, aarch64, i386, etc.
gui-rl/desktop_env/server/main.py:227
↓ 1 callersFunction_get_megatron_full_params
( megatron_local_param_infos: Sequence[ParamInfo], megatron_local_weights, )
slime/slime/backends/megatron_utils/update_weight/hf_weight_iterator_direct.py:44
↓ 1 callersFunction_get_megatron_local_param_info_buckets
Partition params into buckets ≤ update_weight_buffer_size (with TP replication).
slime/slime/backends/megatron_utils/update_weight/hf_weight_iterator_direct.py:108
↓ 1 callersFunction_get_megatron_local_param_infos
Build global param metadata: collect → exchange PP/EP → resolve duplicates (MTP virtual PP) by min src_rank → validate. Returns sorted ParamI
slime/slime/backends/megatron_utils/update_weight/hf_weight_iterator_direct.py:138
↓ 1 callersMethod_get_merges_file
Gets merge file from cache or downloads it. Args: tokenizer_name (str): pretrained model name. merges_file (
Megatron-LM/megatron/core/tokenizers/text/libraries/megatron_hf_tokenizer.py:122
↓ 1 callersFunction_get_moe_activation_scale
Calc moe activation scale factor relative to dense activation. For more detail please refer to https://arxiv.org/abs/2410.07524.
Megatron-LM/megatron/core/transformer/moe/upcycling_utils.py:144
↓ 1 callersFunction_get_mtp_block_submodules
Retrieve or construct MultiTokenPredictionBlockSubmodules based on the provided specification. Args: config (TransformerConfig): Con
Megatron-LM/megatron/core/transformer/multi_token_prediction.py:1033
↓ 1 callersFunction_get_non_persistent_iteration
(non_persistent_global_dir, args, checkpointing_context=None)
Megatron-LM/megatron/training/checkpointing.py:982
↓ 1 callersMethod_get_num_epochs
Calculate the number of epochs Args: num_tokens_per_epoch (int): The number of tokens in a single epoch Returns:
Megatron-LM/megatron/core/datasets/gpt_dataset.py:541
↓ 1 callersFunction_get_num_layers
Compute the number of transformer layers resident on the current rank.
Megatron-LM/megatron/legacy/model/transformer.py:1340
↓ 1 callersMethod_get_num_tokens_per_epoch
Calculate the number of tokens in a single epoch Returns: int: The number of tokens in a single epoch
Megatron-LM/megatron/core/datasets/gpt_dataset.py:533
↓ 1 callersFunction_get_parameter_groups
Get the parameter group for the given module and parameters. Args: module (torch.nn.Module): The module whose parameters are to be g
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/param_and_grad_buffer.py:1258
↓ 1 callersMethod_get_pos_embeddings
Get RADIO absolute positional embeddings
Megatron-LM/megatron/core/models/vision/radio.py:281
↓ 1 callersMethod_get_pp_layer_offset_for_inference
Return the pipeline parallel layer offset for inference.
Megatron-LM/megatron/core/transformer/attention.py:312
↓ 1 callersFunction_get_prm_semaphore
(args)
toolcall-rl/generate_with_retool.py:377
↓ 1 callersMethod_get_prm_semaphore
(self, args: Any)
gui-rl/agents/qwen3vl_reward_agent.py:151
↓ 1 callersMethod_get_prm_semaphore
(self, args: Any)
swe-rl/swe_prm.py:142
↓ 1 callersFunction_get_prm_tokenizer
(args)
toolcall-rl/generate_with_retool.py:387
↓ 1 callersFunction_get_qk_layernorm
(use_te: bool, normalization: str)
Megatron-LM/megatron/core/models/gpt/heterogeneous/heterogeneous_layer_specs.py:77
↓ 1 callersMethod_get_rank_subset
Helper method to get the subset of prompts for a given rank. Args: prompts: List of all prompts num_prompts: Total nu
Megatron-LM/megatron/rl/agent/reward_only_agent.py:62
↓ 1 callersMethod_get_remove_vocab_padding
(self, layer_name, model_state_dict, tokenizer_vocab_size)
Megatron-LM/megatron/core/export/trtllm/trtllm_weights_converter/distributed_trtllm_model_weights_converter.py:209
↓ 1 callersFunction_get_rng_state_dict
()
Megatron-LM/megatron/core/distributed/fsdp/mcore_fsdp_adapter.py:419
↓ 1 callersMethod_get_rollout_data
(self, rollout_id)
slime/slime/ray/rollout.py:234
↓ 1 callersFunction_get_rollout_dump_paths
(load_debug_rollout_data: str, categories: list[str])
slime/slime/utils/debug_utils/display_debug_rollout_data.py:58
↓ 1 callersMethod_get_sample_arguments
Generate sample arguments and keyword arguments for CUDA Graph capturing with memory-optimized buffer reuse. This method cre
Megatron-LM/megatron/core/transformer/cuda_graphs.py:1513
↓ 1 callersMethod_get_samples_from_buffer
(self, num_samples: int)
slime/slime/rollout/data_source.py:180
↓ 1 callersMethod_get_save_and_finalize_callbacks
(self, writer, save_state_dict_ret)
Megatron-LM/megatron/core/dist_checkpointing/strategies/torch.py:839
↓ 1 callersFunction_get_shadow_model_dir
(model_path: str)
slime/slime/backends/sglang_utils/qwen3_5.py:329
↓ 1 callersFunction_get_sharded_state_dict_keys_map
Generate a mapping of sharded state dictionary keys. Mapping in case of not using Transformer Engine with regular attention and mlp. Args
Megatron-LM/megatron/core/models/gpt/heterogeneous/heterogeneous_layer_specs.py:144
↓ 1 callersMethod_get_states_from_cache
Initializes or retrieves the SSM state tensors from the cache. At the start of any inference (at the prefill step), if there is no cache or i
Megatron-LM/megatron/core/ssm/mamba_mixer.py:1017
↓ 1 callersMethod_get_sub_optimizer_param_groups
(self, offload_fraction: float)
Megatron-LM/megatron/core/optimizer/cpu_offloading/hybrid_optimizer.py:251
↓ 1 callersMethod_get_submodules_under_cudagraphs
Get the submodules that are covered by cudagraphs. Return a list that only contains the module itself if the whole layer is covered b
Megatron-LM/megatron/core/transformer/module.py:235
↓ 1 callersFunction_get_swe_save_dir
()
swe-rl/generate_with_swe_trajectory.py:110
↓ 1 callersFunction_get_swe_semaphore
()
swe-rl/generate_with_swe_trajectory.py:91
↓ 1 callersFunction_get_sweagent_config
()
swe-rl/generate_with_swe_trajectory.py:96
↓ 1 callersFunction_get_task_clear_lock
()
gui-rl/generate_with_gui.py:51
↓ 1 callersMethod_get_te_cuda_graph_replay_args
Helper function to get tensor arguments for TE CUDA graph.
Megatron-LM/megatron/core/transformer/module.py:270
↓ 1 callersFunction_get_teacher_model_path
(args: Namespace)
openclaw-combine/prm_teacher_postprocess.py:53
↓ 1 callersFunction_get_text_config
Unwrap text_config from VLM config if present.
slime/slime_plugins/megatron_bridge/qwen3_5_text.py:24
↓ 1 callersFunction_get_tile_tags
Tile tags are used in NVLM to surround image tiles with text tags.
Megatron-LM/examples/multimodal/model.py:236
↓ 1 callersMethod_get_token_mask
(self, numpy_random_state: numpy.random.RandomState)
Megatron-LM/megatron/core/datasets/masked_dataset.py:422
↓ 1 callersFunction_get_train_data_iterator
()
Megatron-LM/tests/unit_tests/export/trtllm/test_single_device_fp8.py:53
↓ 1 callersFunction_get_train_data_iterator
()
Megatron-LM/tests/unit_tests/export/trtllm/test_distributed_fp8.py:56
↓ 1 callersFunction_get_transformer_layer_spec
Get transformer layer specification based on configuration. Args: use_te (bool): Whether to use Transformer Engine args: Trai
Megatron-LM/gpt_builders.py:106
↓ 1 callersMethod_get_trtllm_pretrained_config_and_model_weights_in_distributed_setting
Get the TRTLLM Pretrained config and model weights list in a distributed setting This function assumes the model state dict is distributed a
Megatron-LM/megatron/core/export/trtllm/trtllm_helper.py:377
↓ 1 callersMethod_get_trtllm_pretrained_config_and_model_weights_list_on_single_device
Get the TRTLLM Pretrained config and model weights list (one per gpu rank) on single device (CPU/GPU) This function assumes the entire model
Megatron-LM/megatron/core/export/trtllm/trtllm_helper.py:451
↓ 1 callersMethod_get_turn_context
(self)
terminal-rl/agent_runner.py:48
↓ 1 callersMethod_get_used_ports
Get all currently used ports (both system and Docker).
gui-rl/desktop_env/providers/docker/provider.py:39
↓ 1 callersMethod_get_validation_call_info
Internal method to get the context about the caller to validate_result().
Megatron-LM/megatron/core/rerun_state_machine.py:874
↓ 1 callersMethod_get_varlen_generation_state
Constructs the variable length generation state for non-decode dynamic inference. The returned state includes the following: `seq
Megatron-LM/megatron/core/ssm/mamba_mixer.py:971
↓ 1 callersMethod_get_vocab_file
Gets vocabulary file from cache or downloads it. Args: tokenizer_name (str): pretrained model name. vocab_fi
Megatron-LM/megatron/core/tokenizers/text/libraries/megatron_hf_tokenizer.py:102
↓ 1 callersFunction_get_wrapper_tokens
(tokenizer)
swe-rl/generate_with_swe_trajectory.py:523
↓ 1 callersFunction_get_write_results_queue
()
Megatron-LM/megatron/core/dist_checkpointing/strategies/filesystem_async.py:51
↓ 1 callersMethod_gid_to_src_rank
(self, gid: int, offsets: List[int])
Megatron-LM/megatron/core/datasets/data_schedule.py:126
↓ 1 callersFunction_gpt_builder
(args, pre_process, post_process, vp_stage=None, config=None)
Megatron-LM/train_rl.py:25
↓ 1 callersFunction_gpt_te_layer_spec_with_hetro_pgs
(attn_pg_collection, mlp_pg_collection)
Megatron-LM/tests/unit_tests/transformer/test_transformer_block_custom_pgs.py:171
↓ 1 callersFunction_gui_eval_rollout
(args)
gui-rl/generate_with_gui.py:798
↓ 1 callersMethod_handle_request
( self, body: dict[str, Any], session_id: str, turn_type: str, session
openclaw-rl/openclaw_api_server.py:472
↓ 1 callersMethod_handle_request
( self, body: dict[str, Any], session_id: str, turn_type: str, session
openclaw-rl/oel/openclaw_oel_api_server.py:600
↓ 1 callersMethod_handle_request
( self, body: dict[str, Any], session_id: str, turn_type: str, session
openclaw-opd/openclaw_opd_api_server.py:702
↓ 1 callersMethod_handle_request
( self, body: dict[str, Any], session_id: str, turn_type: str, session
openclaw-fireworks/fireworks_server.py:716
↓ 1 callersMethod_handle_request
(self, body: dict, session_id: str, turn_type: str, session_done: bool)
openclaw-tinker/api_server.py:350
↓ 1 callersFunction_has_active_terminal
A quick check whether the terminal window is open and active.
gui-rl/desktop_env/server/main.py:336
↓ 1 callersMethod_health_check_loop
Background loop to monitor worker health and adjust routing pool.
slime/slime/router/router.py:89
↓ 1 callersFunction_hf_device
()
openclaw-combine/prm_teacher_postprocess.py:44
↓ 1 callersMethod_import_class_from_path
Helper function to import classes from string paths.
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/megatron_fsdp.py:395
↓ 1 callersMethod_indices_to_multihot
Converts a tensor of indices to a multihot vector. Args: indices (torch.Tensor): [num_tokens, topk] token indices, where
Megatron-LM/megatron/core/transformer/moe/token_dispatcher.py:1201
↓ 1 callersMethod_indices_to_multihot
(self, indices, probs)
Megatron-LM/tests/unit_tests/transformer/moe/test_multihot_indices_converter.py:17
↓ 1 callersFunction_init_autoresume
Set autoresume start time.
Megatron-LM/megatron/training/initialize.py:396
↓ 1 callersMethod_init_dist_index
Initialize the distributed index for the module.
Megatron-LM/megatron/core/distributed/fsdp/mcore_fsdp_adapter.py:193
↓ 1 callersMethod_init_distributed_params
Register model training and high-precision parameters as optimizer named parameters and DTensor(s). Specifically, we utilize the high
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/param_and_grad_buffer.py:2402
↓ 1 callersMethod_init_each_parameter_group_buffers
Initialize the buffers for each parameter group.
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/param_and_grad_buffer.py:1768
↓ 1 callersMethod_init_external
(self, expect_server_args, external_engine_need_check_fields)
slime/slime/backends/sglang_utils/sglang_engine.py:179
↓ 1 callersMethod_init_fsdp_param_and_grad_buffer
(self)
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/megatron_fsdp.py:311
↓ 1 callersMethod_init_llava_model
(self, cp_size, tp_size, sequence_parallel)
Megatron-LM/tests/unit_tests/models/test_llava_model.py:542
↓ 1 callersFunction_init_moe_expert_cache
Initialize the cache of MoE layers once
Megatron-LM/megatron/core/inference/utils.py:48
↓ 1 callersMethod_init_normal
(self, server_args_dict)
slime/slime/backends/sglang_utils/sglang_engine.py:203
↓ 1 callersMethod_init_optimizer_named_parameters
(self)
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/param_and_grad_buffer.py:2473
↓ 1 callersMethod_init_optimizer_states_with_dummy_values
(self)
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:1128
↓ 1 callersFunction_init_ray_distributed_post
Initialize one or more Ray async actors per node for HTTP POST. Uses NodeAffinitySchedulingStrategy to place actors on distinct nodes. Contro
slime/slime/utils/http_utils.py:268
← previousnext →2,901–3,000 of 10,561, ranked by callers