MCPcopy Create free account

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

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

↓ 1 callersMethod_backward_output_proj
Update weights for output projection layer
Megatron-LM/megatron/core/transformer/attention.py:1274
↓ 1 callersMethod_backward_output_proj
Computes weight gradients of output projection layer
Megatron-LM/megatron/core/transformer/multi_latent_attention.py:1000
↓ 1 callersMethod_backward_q_proj
Computes weight gradients of Q projection layers
Megatron-LM/megatron/core/transformer/multi_latent_attention.py:992
↓ 1 callersMethod_backward_qkv_proj
Update weights for QKV projection layer
Megatron-LM/megatron/core/transformer/attention.py:1270
↓ 1 callersMethod_batch_search
(self, query_list: list[str], num: int, return_score: bool)
slime/examples/search-r1/local_dense_retriever/retrieval_server.py:135
↓ 1 callersFunction_bias_dropout_add_func_layer_scaling
(ls, x_with_bias, residual, prob, training)
Megatron-LM/examples/multimodal/layer_scaling.py:8
↓ 1 callersFunction_broadcast_num_samples_this_group
(num_samples_this_group)
Megatron-LM/megatron/core/pipeline_parallel/hybrid_cp_schedule.py:520
↓ 1 callersFunction_bshd_to_thd
( unpacked_values: torch.Tensor, cu_seqlens: torch.Tensor, *, batch_size: int, )
slime/slime_plugins/megatron_bridge/qwen35_vl.py:84
↓ 1 callersMethod_bucket_group_gradient_reduce
Mark the bucket ready for reduce-scatter/all-reduce, if all bucket in the bucket group are ready, then do the reduce-scatter/all-reduce.
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/param_and_grad_buffer.py:3083
↓ 1 callersMethod_buffer_record
(self, session_id: str, turn_num: int, messages: list, prompt_text: str, response_text:
openclaw-rl/openclaw_api_server.py:337
↓ 1 callersMethod_buffer_record
(self, session_id, turn_num, messages, prompt_text, response_text, tool_calls)
openclaw-rl/oel/openclaw_oel_api_server.py:1513
↓ 1 callersMethod_buffer_record
( self, session_id: str, turn_num: int, messages: list[dict[str, Any]],
openclaw-opd/openclaw_opd_api_server.py:374
↓ 1 callersMethod_buffer_record
(self, session_id, turn_num, messages, prompt_text, response_text, tool_calls)
openclaw-fireworks/fireworks_server.py:834
↓ 1 callersMethod_build_app
(self)
openclaw-rl/openclaw_api_server.py:266
↓ 1 callersMethod_build_app
(self)
openclaw-rl/oel/openclaw_oel_api_server.py:542
↓ 1 callersMethod_build_app
(self)
openclaw-opd/openclaw_opd_api_server.py:326
↓ 1 callersMethod_build_app
(self)
openclaw-fireworks/fireworks_server.py:323
↓ 1 callersMethod_build_app
(self)
openclaw-tinker/api_server.py:185
↓ 1 callersMethod_build_attention_mask_and_position_ids
Builds the full attention mask and position ids for the input tokens Args: prompts_tokens (torch.Tensor): A tensor of shape [batc
Megatron-LM/megatron/core/inference/model_inference_wrappers/gpt/gpt_inference_wrapper.py:61
↓ 1 callersFunction_build_attention_mask_from_cu_seqlens
( cu_seqlens: torch.Tensor, *, max_seq_len: int, batch_size: int, device: torch.device, )
slime/slime_plugins/megatron_bridge/qwen35_vl.py:54
↓ 1 callersMethod_build_blended_dataset_splits
Build all dataset splits according to the provided blend(s) See the BlendedMegatronDatasetBuilder.build alias for more information.
Megatron-LM/megatron/core/datasets/blended_megatron_dataset_builder.py:137
↓ 1 callersMethod_build_bridge_comms
Construct and store BridgeCommunicator objects that describe the outgoing communication relationships for all of the modules.
Megatron-LM/megatron/core/pipeline_parallel/multimodule_communicator.py:89
↓ 1 callersMethod_build_callable_nodes
Builds the callable nodes for the transformer/mtp layer: attn, post_attn, mlp, moe_dispatch and moe_combine, and mtp_post_process
Megatron-LM/megatron/core/models/common/model_chunk_schedule_plan.py:117
↓ 1 callersFunction_build_document_index
Build an array with length = num epochs * num documents Args: documents (numpy.ndarray): the subset of exposed document indices
Megatron-LM/megatron/core/datasets/gpt_dataset.py:564
↓ 1 callersMethod_build_document_sample_shuffle_indices
Build the document index, the sample index, and the shuffle index The document index: -- 1-D -- An ordered array of d
Megatron-LM/megatron/core/datasets/gpt_dataset.py:318
↓ 1 callersFunction_build_env
Instantiate the interaction environment using the provided module.
slime/examples/geo3k_vlm_multi_turn/rollout.py:42
↓ 1 callersMethod_build_exec_env
(self)
swe-rl/mini-swe-agent/src/minisweagent/environments/docker.py:125
↓ 1 callersMethod_build_gbuf_range_map
Build mapping between params and their grad buffers. These mappings are partitioned according to data type. Iterate through
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:218
↓ 1 callersFunction_build_hf_model
(config: AutoConfig)
slime/tools/convert_fsdp_to_hf.py:108
↓ 1 callersFunction_build_hint_judge_messages
( response_text: str, next_state_text: str, next_state_role: str = "user", )
openclaw-fireworks/fireworks_server.py:118
↓ 1 callersFunction_build_import_path
(subdomains: list, imp)
Megatron-LM/tests/unit_tests/test_imports.py:35
↓ 1 callersMethod_build_indices
Build and optionally cache the dataset index and the dataset sample index The dataset index is a 1-D mapping which determines the dataset to
Megatron-LM/megatron/core/datasets/blended_dataset.py:96
↓ 1 callersMethod_build_inference_context
The inference context manages the KV cache and other inference state.
Megatron-LM/tests/unit_tests/inference/engines/test_dynamic_engine.py:208
↓ 1 callersFunction_build_infinite_size_dataloader
Build a looped dataloader with infinite size.
Megatron-LM/tasks/finetune_utils.py:95
↓ 1 callersMethod_build_initial_messages
Build initial conversation messages. Args: obs: Initial observation from environment Returns: List
slime/examples/tau-bench/trainable_agents.py:145
↓ 1 callersFunction_build_input_example
(metadata: JsonDict)
slime/slime/rollout/rm_hub/ifbench.py:131
↓ 1 callersFunction_build_key_size_numel_dictionaries
Build the size on rank 0 and broadcast.
Megatron-LM/megatron/core/tensor_parallel/data.py:20
↓ 1 callersMethod_build_layers
(self)
Megatron-LM/megatron/core/transformer/transformer_block.py:331
↓ 1 callersMethod_build_layers
(self, pg_collection)
Megatron-LM/megatron/core/transformer/multi_token_prediction.py:1109
↓ 1 callersMethod_build_load_plan
( self, sharded_state_dict: ShardedStateDict )
Megatron-LM/megatron/core/dist_checkpointing/strategies/two_stage.py:170
↓ 1 callersMethod_build_matchers
(matchers_dict: Dict)
Megatron-LM/megatron/core/quantization/quant_config.py:137
↓ 1 callersFunction_build_messages
(data: dict, prompt_key: str, as_conversation: bool, multimodal_keys: dict = None)
slime/slime/utils/data.py:113
↓ 1 callersMethod_build_messages
(self, turn_idx: int)
terminal-rl/agent/prm_agent.py:186
↓ 1 callersMethod_build_model_and_main_param_groups
Create main parameter groups needed for the optimizer step. These groups encompass both: 1) groups used by this class, for r
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:304
↓ 1 callersMethod_build_model_gbuf_param_range_map
Build mapping from param reference to grad buffer shard ranges. This method builds a mapping from parameter references to grad
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:109
↓ 1 callersMethod_build_model_gbuf_range
Build mapping between params and their grad buffers. This method does the initial setup for the method above. This setup inc
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:171
↓ 1 callersMethod_build_model_param_gbuf_map
Create a reverse of the gbuf_ranges, for referencing in opposite direction.
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:238
↓ 1 callersMethod_build_model_param_to_state_dict_param_map
Create a map from model params to tensors in state_dict based on their names.
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:2489
↓ 1 callersMethod_build_multi_meta
( section: Section, before: str, desc: str )
Megatron-LM/tools/retro/config_utils.py:344
↓ 1 callersFunction_build_num_microbatches_calculator
Build number of microbatches calculator. Internal helper method. Args: rank (int): Rank of the GPU, only rank 0 will log the
Megatron-LM/megatron/core/num_microbatches_calculator.py:191
↓ 1 callersMethod_build_optimizer_group_ranges
Create optimizer groups. Given the set of parameter shard ranges that are owned by the current data-parallel (DP) rank, gath
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:257
↓ 1 callersFunction_build_prm_eval_prompt
Build PRM evaluation prompt for monitoring.
openclaw-rl/oel/openclaw_oel_api_server.py:178
↓ 1 callersFunction_build_prm_eval_prompt
( response_text: str, next_state_text: str, next_state_role: str = "user", )
openclaw-fireworks/fireworks_server.py:151
↓ 1 callersFunction_build_prm_judge_prompt
(response_text: str, next_state_text: str, next_state_role: str = "user")
openclaw-rl/openclaw_api_server.py:94
↓ 1 callersFunction_build_prm_step_messages
( *, problem: str, history: str, action: str, observation: str, step_index: int, )
toolcall-rl/generate_with_retool.py:444
↓ 1 callersMethod_build_rank_module_info_map
For each module in the current rank, initialize the P2P communicator and build the bridge communicator info for the module. Each rank
Megatron-LM/megatron/core/pipeline_parallel/multimodule_communicator.py:195
↓ 1 callersMethod_build_requests
(cls, test_config: CoordinatorTestConfig)
Megatron-LM/tests/unit_tests/inference/test_data_parallel_inference_coordinator.py:163
↓ 1 callersMethod_build_requests
(cls, test_config: DynamicEngineTestConfig)
Megatron-LM/tests/unit_tests/inference/engines/test_dynamic_engine.py:150
↓ 1 callersMethod_build_reward_messages
( self, *, problem_statement: str, step_debug: list[dict], policy_resp
swe-rl/swe_prm.py:176
↓ 1 callersMethod_build_reward_messages_aligned
( self, *, instruction: str, actions_history: list[str], policy_respon
gui-rl/agents/qwen3vl_reward_agent.py:161
↓ 1 callersMethod_build_sandbox
(self)
swe-rl/mini-swe-agent/src/minisweagent/environments/singularity.py:39
↓ 1 callersMethod_build_single_meta
(section: Section, desc: str)
Megatron-LM/tools/retro/config_utils.py:323
↓ 1 callersFunction_build_task_dir
(args: Any, domain: str, example_id: str)
gui-rl/generate_with_gui.py:166
↓ 1 callersMethod_build_tool_feedback
Provide concise feedback for the model to continue reasoning.
slime/examples/geo3k_vlm_multi_turn/env_geo3k.py:150
↓ 1 callersFunction_build_train_valid_dataloaders
Traing and validation dataloaders.
Megatron-LM/tasks/finetune_utils.py:106
↓ 1 callersFunction_build_train_valid_test_datasets
(data_prefix, splits_string, train_valid_test_num_samples,
Megatron-LM/megatron/legacy/data/dataset_utils.py:467
↓ 1 callersFunction_build_trajectory_sample
Flatten per-message token_ids into a single Sample. Because the rollout produced byte-perfect token_ids for every message, this function does
swe-rl/generate_with_swe_trajectory.py:820
↓ 1 callersFunction_build_transform
(input_size, vision_model_type)
Megatron-LM/examples/multimodal/image_processing.py:131
↓ 1 callersMethod_calculate_seq_length
(self, audio_tensor)
Megatron-LM/tests/unit_tests/models/test_mimo_audio_submodules.py:222
↓ 1 callersMethod_call_llm
Make an LLM call tracking. Args: url: SGLang server URL payload: Request payload containing text and samplin
slime/examples/tau-bench/trainable_agents.py:90
↓ 1 callersFunction_changed_files_from_patch
(patch_text: str)
swe-rl/generate_with_swe_trajectory.py:210
↓ 1 callersFunction_check_and_sort_partitions
(partitions)
slime/slime/utils/seqlen_balancing.py:164
↓ 1 callersFunction_check_arg_is_not_none
(args, arg)
Megatron-LM/megatron/training/arguments.py:1333
↓ 1 callersFunction_check_arg_is_not_none
(args, arg)
Megatron-LM/megatron/training/yaml_arguments.py:400
↓ 1 callersMethod_check_auth
(self, authorization: str | None)
openclaw-rl/openclaw_api_server.py:305
↓ 1 callersMethod_check_auth
(self, authorization: str | None)
openclaw-rl/oel/openclaw_oel_api_server.py:587
↓ 1 callersMethod_check_auth
(self, authorization: str | None)
openclaw-opd/openclaw_opd_api_server.py:365
↓ 1 callersMethod_check_auth
(self, authorization: str | None)
openclaw-fireworks/fireworks_server.py:362
↓ 1 callersMethod_check_auth
(self, authorization: Optional[str])
openclaw-tinker/api_server.py:223
↓ 1 callersMethod_check_code_safety
Check code safety by scanning for dangerous patterns
toolcall-rl/tool_sandbox.py:112
↓ 1 callersMethod_check_contain_doc
(self)
slime/examples/search-r1/local_dense_retriever/retrieval_server.py:156
↓ 1 callersFunction_check_data_types
Check that all the keys have the same target data type.
Megatron-LM/megatron/core/tensor_parallel/data.py:10
↓ 1 callersMethod_check_engine_health
(self, rollout_engine_id, engine)
slime/slime/utils/health_monitor.py:146
↓ 1 callersMethod_check_module_parameter_types
Check if the module parameters include special parameters such as Megatron-Core Expert Parallel (EP/EXPT) parameters.
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/megatron_fsdp.py:297
↓ 1 callersMethod_check_return_values
( self, missing_keys, unexpected_keys, should_contain_missing, should_contain_unexpected )
Megatron-LM/tests/unit_tests/dist_checkpointing/test_strict.py:168
↓ 1 callersMethod_check_return_values_for_strict_handling
( self, strict, missing_keys, unexpected_keys, missing, unexpected )
Megatron-LM/tests/unit_tests/dist_checkpointing/test_strict.py:180
↓ 1 callersMethod_check_toggle
Helper method to check if a request to toggle the collection state was made It checks iof collection state toggle req was made via the server
Megatron-LM/megatron/core/utils.py:1644
↓ 1 callersMethod_check_worker_health
Encapsulated health check logic for better maintainability.
slime/slime/router/router.py:78
↓ 1 callersMethod_checkpointed_attention_forward
Forward method with activation checkpointing.
Megatron-LM/megatron/legacy/model/transformer.py:625
↓ 1 callersMethod_checkpointed_forward
Forward method with activation checkpointing.
Megatron-LM/megatron/legacy/model/transformer.py:1596
↓ 1 callersMethod_checkpointed_forward
Forward method with activation checkpointing.
Megatron-LM/megatron/core/transformer/transformer_block.py:422
↓ 1 callersMethod_checkpointed_forward
Wrap `forward_func` with activation checkpointing while only passing tensors. Non-tensor arguments (e.g., configuration objects, None) are ca
Megatron-LM/megatron/core/transformer/multi_token_prediction.py:831
↓ 1 callersFunction_chunk_by_size
(objects: Iterable[Any], compute_size: Callable[[Any], int], chunk_size: int)
slime/slime/utils/misc.py:131
↓ 1 callersFunction_chunk_list_closure
(chunk_meta)
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/uneven_dtensor.py:104
↓ 1 callersMethod_clean_text
Performs invalid character removal and whitespace cleanup on text.
Megatron-LM/megatron/training/tokenizer/bert_tokenization.py:318
↓ 1 callersMethod_cleanup_session
Clean up all state for a session.
openclaw-rl/oel/openclaw_oel_api_server.py:1538
↓ 1 callersFunction_clear_cache_control
(entry: dict)
swe-rl/mini-swe-agent/src/minisweagent/models/utils/cache_control.py:8
↓ 1 callersFunction_clip_eval_output
Clip eval output to bound payload size when needed. Set SWE_EVAL_OUTPUT_MAX_CHARS<=0 to disable clipping.
swe-rl/server/swe_exec_server.py:73
↓ 1 callersMethod_close_run_slot
( self, task_key: str, run_lease_id: str, run_slot: RunSlot, *, reason: str )
terminal-rl/remote/pool_server.py:115
↓ 1 callersFunction_coerce_kwargs_list
Convert stored kwargs into the list structure expected by IFBench.
slime/slime/rollout/rm_hub/ifbench.py:100
← previousnext →2,601–2,700 of 10,561, ranked by callers