MCPcopy Create free account

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

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

↓ 1 callersMethod_start_capturing
Start capturing CUDA Graphs.
Megatron-LM/megatron/core/transformer/cuda_graphs.py:1856
↓ 1 callersMethod_start_container
Start the Docker container and return the container ID.
swe-rl/mini-swe-agent/src/minisweagent/environments/docker.py:55
↓ 1 callersFunction_str_to_int
(x: str)
slime/slime/rollout/rm_hub/math_utils.py:223
↓ 1 callersMethod_stream_response
(self, result: dict[str, Any])
openclaw-rl/openclaw_api_server.py:677
↓ 1 callersMethod_stream_response
(self, result: dict[str, Any])
openclaw-rl/oel/openclaw_oel_api_server.py:1603
↓ 1 callersMethod_stream_response
(self, result: dict[str, Any])
openclaw-opd/openclaw_opd_api_server.py:950
↓ 1 callersMethod_stream_response
(self, result: dict[str, Any])
openclaw-fireworks/fireworks_server.py:896
↓ 1 callersMethod_stream_response
(self, result: dict)
openclaw-tinker/api_server.py:415
↓ 1 callersFunction_strip_best_prefix
(keys: list[str], target_keys: set[str])
slime/tools/convert_fsdp_to_hf.py:94
↓ 1 callersFunction_strip_string
(string)
slime/slime/rollout/rm_hub/math_utils.py:29
↓ 1 callersMethod_submit_rl_sample
(self, turn_data, session_id, eval_score)
openclaw-tinker/api_server.py:1060
↓ 1 callersMethod_submit_rl_turn_sample
( self, turn_data: dict, session_id: str, eval_score: float, )
openclaw-fireworks/fireworks_server.py:624
↓ 1 callersMethod_submit_rl_turn_sample
( self, turn_data: dict, session_id: str, eval_score: float, )
openclaw-combine/openclaw_combine_api_server.py:109
↓ 1 callersMethod_submit_turn_sample
(self, turn_data: dict[str, Any], session_id: str, prm_result: dict | None)
openclaw-rl/openclaw_api_server.py:620
↓ 1 callersMethod_submit_turn_sample
(self, turn_data: dict[str, Any], session_id: str, teacher_result: dict[str, Any])
openclaw-rl/oel/openclaw_oel_api_server.py:1435
↓ 1 callersMethod_submit_turn_sample
(self, turn_data: dict[str, Any], session_id: str, opd_result: dict[str, Any])
openclaw-opd/openclaw_opd_api_server.py:867
↓ 1 callersMethod_submit_turn_sample
(self, turn_data, session_id, prm_result)
openclaw-tinker/api_server.py:610
↓ 1 callersMethod_submit_turn_sample
(self, turn_data, session_id, opd_result)
openclaw-tinker/api_server.py:808
↓ 1 callersMethod_submit_turn_sample_no_teacher
Submit a sample without teacher log-probs (when PRM is not available).
openclaw-rl/oel/openclaw_oel_api_server.py:1479
↓ 1 callersMethod_swap_book_keeping_tensors
Swaps all the relevent booking tensors with src idxs to dst idxs
Megatron-LM/megatron/core/inference/contexts/dynamic_context.py:1568
↓ 1 callersFunction_sympy_parse
Parses an expression with sympy.
slime/slime/rollout/rm_hub/math_utils.py:168
↓ 1 callersMethod_sync_sub_optimizers_state_to_hdo
Update HDO state attribute to sub-optimizers.
Megatron-LM/megatron/core/optimizer/cpu_offloading/hybrid_optimizer.py:302
↓ 1 callersFunction_te_rms_norm_kernel
(x: torch.Tensor, weight: torch.Tensor, eps: float)
Megatron-LM/megatron/core/tensor_parallel/inference_layers.py:30
↓ 1 callersMethod_temporarily_bypass_shape_validation
Temporarily set the size of tensors to their expected shapes to bypass DCP shape validation. This is used when validating the shapes
Megatron-LM/megatron/core/dist_checkpointing/strategies/torch.py:623
↓ 1 callersFunction_test_fused_apply_mla_rope_for_kv
(input_format)
Megatron-LM/tests/unit_tests/fusions/test_mla_yarn_rope_apply.py:119
↓ 1 callersFunction_test_fused_apply_mla_rope_for_q
(input_format)
Megatron-LM/tests/unit_tests/fusions/test_mla_yarn_rope_apply.py:40
↓ 1 callersFunction_test_parallel_attention_correctness
( transformer_config, transformer_layer_spec, tmp_path_dist_ckpt, atol, rtol, tp=1,
Megatron-LM/tests/unit_tests/transformer/test_attention.py:514
↓ 1 callersMethod_text_to_ids
Converts text to tokens ids.
Megatron-LM/megatron/core/tokenizers/text/libraries/sentencepiece_tokenizer.py:148
↓ 1 callersFunction_thd_to_bshd
( packed_values: torch.Tensor, cu_seqlens: torch.Tensor, *, batch_size: int, )
slime/slime_plugins/megatron_bridge/qwen35_vl.py:69
↓ 1 callersMethod_to_conversation
(cls, question, response)
Megatron-LM/examples/post_training/modelopt/finetune.py:309
↓ 1 callersFunction_to_cuda
Recursively move tensors to current CUDA device. Handles bare tensors, lists, dicts (including nested), and None. Non-tensor scalars (int, fl
slime/slime/backends/megatron_utils/data.py:25
↓ 1 callersFunction_to_dtensor
(wrapped_model, model_state_dict)
Megatron-LM/megatron/training/checkpointing.py:1854
↓ 1 callersFunction_to_local_gpu_id
(physical_gpu_id: int)
slime/slime/backends/sglang_utils/sglang_engine.py:36
↓ 1 callersMethod_tokenize_chinese_chars
Adds whitespace around any CJK character.
Megatron-LM/megatron/training/tokenizer/bert_tokenization.py:283
↓ 1 callersFunction_tokenize_prompts_and_batch
Given a set of prompts and number of tokens to generate: - tokenize prompts - set the sequence length to be the max of length of promp
Megatron-LM/tools/retro/text_generation/retro_api.py:53
↓ 1 callersFunction_tokenize_prompts_and_batch
Given a set of prompts and number of tokens to generate: - tokenize prompts - set the sequence length to be the max of length of prompts
Megatron-LM/megatron/core/inference/text_generation_server/tokenization.py:70
↓ 1 callersFunction_torch_native_attention
Torch native attention implementation This was not in the original implementation and slightly affect results; it prevents overflow in BF16/FP
Megatron-LM/tests/unit_tests/transformer/test_attention.py:755
↓ 1 callersFunction_train
Train the model.
Megatron-LM/tasks/finetune_utils.py:145
↓ 1 callersMethod_train
Train index (rank 0's method). Args: config (RetroPreprocessingConfig): Retro preprocessing config.
Megatron-LM/megatron/core/datasets/retro/index/indexes/faiss_base.py:48
↓ 1 callersMethod_train_core
(self, rollout_id: int, rollout_data)
slime/slime/backends/fsdp_utils/actor.py:530
↓ 1 callersFunction_train_index
Train index on DB chunks. Args: config (RetroPreprocessingConfig): Retro preprocessing config.
Megatron-LM/megatron/core/datasets/retro/index/build.py:209
↓ 1 callersMethod_train_on_batch
(self, batch: list[TrainingSample], step: int)
openclaw-tinker/trainer.py:131
↓ 1 callersMethod_train_step
(self, packed_batch, reported_accum, mbs_id, grad_accum)
slime/slime/backends/fsdp_utils/actor.py:586
↓ 1 callersFunction_transform
(img_h, img_w)
Megatron-LM/megatron/legacy/data/multimodal_dataset.py:19
↓ 1 callersMethod_truncate_input_ids
(self, input_ids: List[int])
terminal-rl/inference_client.py:187
↓ 1 callersFunction_try_acquire
(args)
slime/tests/ci/gpu_lock_exec.py:86
↓ 1 callersFunction_try_acquire_count
(count: int, total_gpus: int, path_pattern: str, timeout: int)
slime/tests/ci/gpu_lock_exec.py:119
↓ 1 callersFunction_try_acquire_specific
(devs: list[int], path_pattern: str, timeout: int)
slime/tests/ci/gpu_lock_exec.py:94
↓ 1 callersMethod_try_ci_fault_injection
Try to inject fault during generate (when health monitor is running).
slime/slime/ray/rollout.py:102
↓ 1 callersFunction_undo_attention_load_balancing
Undoes the context parallel attention load balancing For example, for cp_size=3, converts 162534 to 123456 for sequential processing by t
Megatron-LM/megatron/core/ssm/mamba_context_parallel.py:360
↓ 1 callersMethod_unscale_main_grads_and_check_for_nan
(self)
Megatron-LM/megatron/core/optimizer/optimizer.py:496
↓ 1 callersFunction_unwrap_pyt_sharded_tensor
Unwrap tensor from PyT ShardedTensor instance. If `prepend_axis_num` was non-zero (which is specific to MCore ShardedTensor) then the tensor
Megatron-LM/megatron/core/dist_checkpointing/strategies/torch.py:385
↓ 1 callersMethod_update_expert_weight_from_distributed
Expert: gather TP → rm pad → buffer. EP gather + HF deferred. Threshold × EP size.
slime/slime/backends/megatron_utils/update_weight/update_weight_from_distributed.py:158
↓ 1 callersMethod_update_fp32_params_by_new_state
(self)
Megatron-LM/megatron/core/optimizer/cpu_offloading/hybrid_optimizer.py:369
↓ 1 callersMethod_update_headers
Update just the title with current state and spinner if needed.
swe-rl/mini-swe-agent/src/minisweagent/agents/interactive_textual.py:386
↓ 1 callersMethod_update_legacy_world_tensors
Reshard buckets (where each bucket is a tensor) to new target numels, where the total numel remains the same.
Megatron-LM/megatron/core/optimizer/distrib_optimizer.py:1816
↓ 1 callersFunction_update_multimodal_state
( sample: Sample, current_image_data, obs_image_data, obs_multimodal_inputs, obs_multimoda
slime/examples/geo3k_vlm_multi_turn/rollout.py:250
↓ 1 callersFunction_update_offsets_and_cumulative_shape
( mesh_dim: int, offsets: List[int], cumulative_shape: List[int] )
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/uneven_dtensor.py:44
↓ 1 callersFunction_update_router_expert_bias
Update the expert bias of the router for a global batch. This requires all-reduce of local_tokens_per_expert across TPxCPxDP ranks
Megatron-LM/megatron/core/distributed/finalize_model_grads.py:293
↓ 1 callersFunction_update_timeouts
(selected_sections, calc_out_of_section)
Megatron-LM/megatron/training/ft_integration.py:226
↓ 1 callersFunction_update_vm
Update the VMX file with the new VM name and other parameters, so that the VM can be started successfully without conflict with the original VM.
gui-rl/desktop_env/providers/vmware/manager.py:62
↓ 1 callersMethod_update_weight_from_distributed
Non-expert: gather TP → rm pad → HF → buffer (flush if full). All gather, PP source buffers. Returns updated bytes on source, None on
slime/slime/backends/megatron_utils/update_weight/update_weight_from_distributed.py:134
↓ 1 callersFunction_use_text_only_qwen35_bridge
()
slime/slime_plugins/megatron_bridge/__init__.py:22
↓ 1 callersFunction_use_text_only_qwen35_bridge
()
slime/slime/utils/megatron_bridge_utils.py:72
↓ 1 callersMethod_use_url
Select worker URL with minimal active requests.
slime/slime/router/router.py:225
↓ 1 callersFunction_validate_common_state_dict
Validate consistancy across ranks for the common state dict We save the common state dict only on rank 0. We validate to make sure that the commo
Megatron-LM/megatron/core/dist_checkpointing/validation.py:375
↓ 1 callersFunction_validate_objects_for_key
Ensure uniqueness of saved objects.
Megatron-LM/megatron/core/dist_checkpointing/validation.py:529
↓ 1 callersFunction_validate_sharding_for_key
(rank_sharding: List[Tuple[int, ShardedTensor]])
Megatron-LM/megatron/core/dist_checkpointing/validation.py:447
↓ 1 callersMethod_validate_subtree_weight_versions
Validate that all nodes in subtree have weight_version <= parent weight_version. Args: node: Root node of subtree to vali
slime/slime/router/middleware_hub/radix_tree.py:503
↓ 1 callersFunction_version_no_greater_than
(version, version_limit)
Megatron-LM/megatron/core/ssm/triton_cache_manager.py:20
↓ 1 callersFunction_vocab_parallel_global_lse
Global LSE over the full vocab, TP-aware (autograd-connected). Returns ``[R, 1]`` in float32. Use to convert raw logits to global log-probabi
openclaw-combine/hint_opd_loss.py:208
↓ 1 callersFunction_vocab_parallel_log_probs_at_indices
No-grad: GLOBAL log-probs at given GLOBAL vocab indices, TP-sharded. Args: logits: ``[seq, V_local]``. indices: ``[seq, K]`` glo
slime/slime/backends/megatron_utils/loss.py:383
↓ 1 callersFunction_w_opd
()
openclaw-combine/openclaw_topk_select_loss.py:106
↓ 1 callersFunction_w_rl
()
openclaw-combine/openclaw_topk_select_loss.py:102
↓ 1 callersMethod_wait_for_vm_ready
Wait for VM to be ready by checking screenshot endpoint.
gui-rl/desktop_env/providers/docker/provider.py:65
↓ 1 callersMethod_wait_if_paused
(self)
terminal-rl/agent/camel_agent.py:388
↓ 1 callersFunction_wait_until_server_ready
Wait until the server is ready
gui-rl/desktop_env/providers/aliyun/manager.py:252
↓ 1 callersFunction_want_megatron_teacher
()
openclaw-combine/prm_teacher_postprocess.py:39
↓ 1 callersFunction_warmup_jit_function
Compilie JIT functions before the main training steps
Megatron-LM/megatron/training/initialize.py:463
↓ 1 callersMethod_weight_name_mapping_mtp
(self, name: str, num_layers: int)
slime/slime_plugins/mbridge/glm4moe.py:35
↓ 1 callersFunction_worker_unreachable
( *, worker_idx: int, worker_url: str, path: str, exc: BaseException, lease_id: str |
terminal-rl/router_server.py:176
↓ 1 callersFunction_wrap_te_linear_for_padding
Wrap a TE linear module to automatically pad sequences for FP8 inference. Modifies the module's forward method to: 1. Pad input seque
Megatron-LM/megatron/core/fp8_utils.py:672
↓ 1 callersMethod_wrap_tool_response
(self, parts: List[Dict])
gui-rl/agents/qwen35_agent.py:123
↓ 1 callersFunction_write_checkpoint_metadata
(path: Path, metadata: dict[str, Any])
slime/slime/backends/fsdp_utils/checkpoint.py:72
↓ 1 callersFunction_write_items_closure
(uneven_chunk_meta)
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/uneven_dtensor.py:107
↓ 1 callersFunction_yarn_find_correction_range
( low_rot: float, high_rot: float, dim: int, rotary_base: float = 10000, max_position_embe
Megatron-LM/megatron/core/models/common/embeddings/yarn_rotary_pos_embedding.py:228
↓ 1 callersFunction_yarn_linear_ramp_mask
(min: float, max: float, dim: int, device: torch.device)
Megatron-LM/megatron/core/models/common/embeddings/yarn_rotary_pos_embedding.py:244
↓ 1 callersFunctionaccimage_loader
(path: str)
Megatron-LM/megatron/legacy/data/image_folder.py:245
↓ 1 callersMethodadd
Add to index. Args: config (RetroPreprocessingConfig): Retro preprocessing config. text_dataset (GPTToTextDataset): T
Megatron-LM/megatron/core/datasets/retro/index/indexes/faiss_base.py:159
↓ 1 callersMethodadd_additional_special_tokens
(self, tokens_list)
Megatron-LM/megatron/training/tokenizer/tokenizer.py:231
↓ 1 callersFunctionadd_algo_arguments
(parser)
slime/slime/utils/arguments.py:754
↓ 1 callersMethodadd_attributes
Utility to add more attributes to inference params Use this method to pass in a custom dictionary to add more configs to the instance created
Megatron-LM/megatron/core/inference/model_inference_wrappers/inference_wrapper_config.py:53
↓ 1 callersMethodadd_block_data
Add data for set of blocks :param row_id: 1D array of unique int ids for the blocks :param block_embeds: 2D array of embeddin
Megatron-LM/megatron/legacy/data/realm_index.py:62
↓ 1 callersFunctionadd_ci_arguments
(parser)
slime/slime/utils/arguments.py:1610
↓ 1 callersFunctionadd_cluster_arguments
(parser)
slime/slime/utils/arguments.py:38
↓ 1 callersMethodadd_codes
Read codes from disk, and add them to the index. Args: config (RetroPreprocessingConfig): Retro preprocessing config.
Megatron-LM/megatron/core/datasets/retro/index/indexes/faiss_par_add.py:158
↓ 1 callersFunctionadd_config_args
(parser, config_cls)
Megatron-LM/tools/retro/config_utils.py:526
↓ 1 callersFunctionadd_custom_megatron_plugins_arguments
Add custom Megatron plugins arguments. This is a placeholder for any additional arguments that might be needed.
slime/slime/utils/arguments.py:1565
↓ 1 callersFunctionadd_data_arguments
(parser)
slime/slime/utils/arguments.py:533
↓ 1 callersFunctionadd_debug_arguments
(parser)
slime/slime/utils/arguments.py:1199
← previousnext →3,301–3,400 of 10,561, ranked by callers