MCPcopy Create free account

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

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

↓ 2 callersMethodlog
(self, metrics, iteration)
Megatron-LM/tests/unit_tests/transformer/test_multi_token_prediction.py:659
↓ 2 callersFunctionlog_rollout_data
Summarize rollout fields and log reduced metrics on PP last stage, TP rank 0. - Tensor-valued lists are concatenated and averaged. For token
slime/slime/backends/megatron_utils/data.py:498
↓ 2 callersFunctionlogger_stack
Context manager for managing logger and name stack. Temporarily pushes a logger and/or name onto their respective stacks, allowing hierarchical
Megatron-LM/megatron/core/dist_checkpointing/utils.py:255
↓ 2 callersFunctionloss_func
Loss function (with KD Loss support). Args: loss_mask (Tensor): Used to mask out some portions of the loss output_tensor (Tensor)
Megatron-LM/megatron/post_training/loss_func.py:39
↓ 2 callersFunctionmain
Run on a single SWE-Bench instance.
swe-rl/mini-swe-agent/src/minisweagent/run/extra/swebench_single.py:28
↓ 2 callersFunctionmake_attention_mask
Returns a 2-dimensional (2-D) attention mask :param source_block: 1-D array :param target_block: 1-D array
Megatron-LM/megatron/legacy/data/ict_dataset.py:13
↓ 2 callersMethodmake_prefix
(self, target, nums)
Megatron-LM/examples/rl/environments/countdown/countdown_agent.py:15
↓ 2 callersFunctionmake_sharded_object_for_checkpoint
Helper for instantiating a non-sharded ShardedObject (replicated across TP and DP group). Args: obj (object): any object to be sharded
Megatron-LM/megatron/core/transformer/utils.py:151
↓ 2 callersFunctionmake_sharded_optimizer_tensor
Build a ShardedTensor or ShardedTensorFactory for optimizer param based on model param Args: model_param (Union[ShardedTensor, ShardedTen
Megatron-LM/megatron/core/dist_checkpointing/optimizer.py:81
↓ 2 callersFunctionmake_sharded_tensor_for_checkpoint
Helper for instantiating a non-sharded ShardedTensor (replicated across TP and DP group). Optionally, can provide offsets which prepend new dimen
Megatron-LM/megatron/core/utils.py:953
↓ 2 callersMethodmark_proxy_success
Mark proxy as successful
gui-rl/desktop_env/providers/aws/proxy_pool.py:107
↓ 2 callersFunctionmask
( weights: torch.Tensor, loss_mask: torch.Tensor, metrics: dict[str, list[torch.Tensor]], lowe
slime/examples/train_infer_mismatch_helper/mis.py:131
↓ 2 callersFunctionmathd_normalize_answer
(answer: str | None)
slime/slime/rollout/rm_hub/math_utils.py:15
↓ 2 callersFunctionmathvista_eval
Run MathVista evaluation.
Megatron-LM/examples/multimodal/evaluation/evaluate_mathvista.py:108
↓ 2 callersFunctionmaybe_cat
Concatenates `a` and `b` along `dim` if `a` and `b` exist.
Megatron-LM/megatron/core/utils.py:2257
↓ 2 callersFunctionmaybe_raise_signal
Check if the MoE layer should early return for CUDA graph capture. If so, raise a MoECudaGraphPartialCaptureSignal.
Megatron-LM/megatron/core/transformer/moe/moe_utils.py:1199
↓ 2 callersFunctionmaybe_report_missing_and_unexpected_keys
Raises or logs an error in case missing or unexpected keys are non-empty. Args: missing_keys (Set[str]): missing keys in the state dict
Megatron-LM/megatron/core/dist_checkpointing/validation.py:331
↓ 2 callersFunctionmcore_to_pyt_state_dict
Convert state dict with ShardedTensors and ShardedObjects to state dict compatible with PyT Dist format. Operates in-place and returns the or
Megatron-LM/megatron/core/dist_checkpointing/strategies/torch.py:295
↓ 2 callersFunctionmeasure_saturation
Measure the average saturation of an image
gui-rl/desktop_env/evaluators/metrics/gimp.py:173
↓ 2 callersFunctionmegatron_rl_inference_mode
Manage the model inference context when collecting rollouts. Args: model: model to prepare. optimizer: optimizer used to train th
Megatron-LM/megatron/rl/rl_utils.py:2230
↓ 2 callersMethodmerge
(self, other)
slime/slime/utils/seqlen_balancing.py:32
↓ 2 callersFunctionmmmu_main_eval
(output_dict, task_cfg)
Megatron-LM/examples/multimodal/evaluation/mmmu_utils.py:405
↓ 2 callersFunctionmodel_provider
Builds the model. If you set the use_legacy_models to True, it will return the legacy GPT model and if not the mcore GPT model. Args:
Megatron-LM/model_provider.py:24
↓ 2 callersFunctionmodel_provider
Builds the model. Args: pre_process (bool): Include the embedding layer in the gpt decoder (used with pipeline parallelism). Defaults to
Megatron-LM/examples/multimodal/model.py:18
↓ 2 callersFunctionmost_recent_pipeline
(target_branch: str)
Megatron-LM/tests/test_utils/python_scripts/check_status_of_main.py:25
↓ 2 callersFunctionmotionbench_eval
(input_path)
Megatron-LM/examples/multimodal/evaluation/evaluate_video_motionbench.py:33
↓ 2 callersFunctionmove_torch_optimizer
ref: https://github.com/volcengine/verl/blob/main/verl/utils/fsdp_utils.py
slime/slime/backends/fsdp_utils/actor.py:964
↓ 2 callersFunctionmvbench_eval
(input_path)
Megatron-LM/examples/multimodal/evaluation/evaluate_video_mvbench.py:98
↓ 2 callersFunctionnested_items_iter
Returns iterator over (nested) tuples (container, key, value) of a given dict or list.
Megatron-LM/megatron/core/dist_checkpointing/dict_utils.py:176
↓ 2 callersFunctionnew_get_model_chunk_id
Helper method to get the model chunk ID given the iteration number.
Megatron-LM/tests/unit_tests/pipeline_parallel/test_helpers.py:71
↓ 2 callersFunctionnormalize_answer
Lower text and remove punctuation, articles and extra whitespace.
Megatron-LM/tools/retro/text_generation/metrics.py:19
↓ 2 callersFunctionnormalize_answer
(s)
slime/slime/rollout/rm_hub/f1.py:6
↓ 2 callersFunctionnormalize_brightness
Normalize the brightness of an image to a target brightness in [0, 1]
gui-rl/desktop_env/evaluators/metrics/gimp.py:161
↓ 2 callersFunctionnormalize_outputs
Strip leading/trailing whitespace and normalize internal whitespace
swe-rl/mini-swe-agent/tests/run/test_github_issue.py:10
↓ 2 callersFunctionnormalize_url
(url)
gui-rl/desktop_env/evaluators/getters/chrome.py:1569
↓ 2 callersFunctionnum_floating_point_operations
(args, batch_size)
Megatron-LM/megatron/training/training.py:168
↓ 2 callersFunctionnvtx_decorated_function
()
Megatron-LM/tests/unit_tests/test_utils.py:164
↓ 2 callersFunctionnvtx_decorated_function_with_message
()
Megatron-LM/tests/unit_tests/test_utils.py:168
↓ 2 callersMethodoccupy_vm
(self, vm_path, pid, region)
gui-rl/desktop_env/providers/azure/manager.py:41
↓ 2 callersFunctionocrbench_eval
Run OCRBench evaluation.
Megatron-LM/examples/multimodal/evaluation/evaluate_ocrbench.py:123
↓ 2 callersFunctionoffset_slice
(s, offset)
Megatron-LM/megatron/core/transformer/fsdp_dtensor_checkpoint.py:209
↓ 2 callersMethodoffsets
Convert embedding ids to text offsets Args: ids (list[int]): The ids to convert text (str): The text to convert
Megatron-LM/megatron/core/datasets/megatron_tokenizer.py:68
↓ 2 callersMethodoffsets
(self, ids: list[int], text: str)
Megatron-LM/megatron/training/tokenizer/tokenizer.py:831
↓ 2 callersFunctionoffsets_to_substrs
(offsets, string)
Megatron-LM/tests/unit_tests/test_tokenizer.py:22
↓ 2 callersMethodon_init_end
(self)
slime/slime/utils/profile_utils.py:26
↓ 2 callersMethodon_input_submitted
Handle single-line input submission.
swe-rl/mini-swe-agent/src/minisweagent/agents/interactive_textual.py:217
↓ 2 callersMethodon_uncaught_exception
(self, instance_id: str, exception: Exception)
swe-rl/mini-swe-agent/src/minisweagent/run/extra/utils/batch_progress.py:158
↓ 2 callersMethodonload
(self, tags: list[str] | None = None)
slime/slime/ray/rollout.py:185
↓ 2 callersFunctionpad_and_convert_to_numpy
Pad sequences and convert them to numpy.
Megatron-LM/tools/retro/sft/dataset_conv.py:420
↓ 2 callersFunctionpad_routing_map
Pad the routing map to ensure each expert has a multiple of pad_multiple tokens. This function ensures that each expert has a number of tokens th
Megatron-LM/megatron/core/transformer/moe/moe_utils.py:495
↓ 2 callersFunctionpad_tokens
(tokens, pad)
slime/slime/backends/megatron_utils/cp_utils.py:187
↓ 2 callersFunctionpad_vocab_size
Pad vocab size based on inference size
Megatron-LM/megatron/core/export/trtllm/trtllm_weights_converter/single_device_trtllm_model_weights_converter.py:26
↓ 2 callersFunctionparallel_self_attention
(tensor_model_parallel_size, num_att_heads_per_partition, hidden_size_per_att_head
Megatron-LM/megatron/legacy/mpu/tests/test_layers.py:315
↓ 2 callersFunctionparallel_transformer
(tensor_model_parallel_size, num_att_heads_per_partition, hidden_size_per_att_head, b
Megatron-LM/megatron/legacy/mpu/tests/test_layers.py:400
↓ 2 callersFunctionparam_is_not_shared
(param)
Megatron-LM/megatron/legacy/model/module.py:18
↓ 2 callersFunctionparam_is_not_shared
(param)
Megatron-LM/megatron/core/transformer/module.py:24
↓ 2 callersFunctionparam_is_not_tensor_parallel_duplicate
Returns true if the passed-in parameter is not a duplicate parameter on another TP rank.
Megatron-LM/megatron/core/tensor_parallel/layers.py:89
↓ 2 callersFunctionparse_judge_result
Extract score and hint from hint-judge output.
openclaw-tinker/scorers.py:132
↓ 2 callersFunctionparse_open_response
Parse the prediction from the generated response. Return a list of predicted strings or numbers.
Megatron-LM/examples/multimodal/evaluation/mmmu_utils.py:266
↓ 2 callersMethodparse_response
( self, response: str, original_width: int, original_height: int, proc
gui-rl/agents/qwen35_agent.py:466
↓ 2 callersFunctionparse_snippet
(snippet: str)
slime/examples/search-r1/google_search_server.py:11
↓ 2 callersMethodparse_str_to_list
Parse a layout string to a list of lists. Example: "Ettt|(tt|)*29,m|L" will be parsed to [["E","t","t","t"]]+[["t","t"]]*29+[["m"],["L
Megatron-LM/megatron/core/transformer/pipeline_parallel_layer_layout.py:283
↓ 2 callersFunctionparse_strict_flag
Parse user passed strict flag from a string to StrictHandling instance. Args: strict (str, StrictHandling): strict flag to parse. If alre
Megatron-LM/megatron/core/dist_checkpointing/validation.py:110
↓ 2 callersFunctionpartition_buckets
Automatically regroup the buckets of input buffers and return a list of bucket groups. In some scenarios, we need to put buckets from differ
Megatron-LM/megatron/core/distributed/param_and_grad_buffer.py:934
↓ 2 callersMethodpartition_input
(self, input)
Megatron-LM/tests/unit_tests/transformer/moe/test_aux_loss.py:42
↓ 2 callersFunctionpatch_rotary_embedding
(cls)
slime/slime/backends/megatron_utils/__init__.py:29
↓ 2 callersFunctionpatch_sglang_qwen35
()
slime/slime/backends/sglang_utils/qwen3_5.py:53
↓ 2 callersMethodpause
Pause energy monitor (must resume afterward).
Megatron-LM/megatron/core/energy_monitor.py:48
↓ 2 callersFunctionphys_game_bench_eval
(input_path)
Megatron-LM/examples/multimodal/evaluation/evaluate_video_phys_game_bench.py:83
↓ 2 callersFunctionpil_loader
(path: str)
Megatron-LM/megatron/legacy/data/image_folder.py:237
↓ 2 callersMethodpop_layer
Pops the transformer layer in FILO order.
Megatron-LM/megatron/core/models/common/model_chunk_schedule_plan.py:411
↓ 2 callersMethodpost_init
(self, root_tmp_dir, tp, pp, async_save, algo, repl_groups)
Megatron-LM/tests/unit_tests/dist_checkpointing/test_replication.py:93
↓ 2 callersFunctionpostprocess_numpy_array
Turn numpy array to torch tensor.
Megatron-LM/megatron/core/dist_checkpointing/strategies/zarr.py:263
↓ 2 callersMethodpre_conv_ssm
Method to be applied before the convolution and SSM
Megatron-LM/megatron/core/ssm/mamba_context_parallel.py:119
↓ 2 callersMethodprepare_gradient_calculation_operands
Prepare gradient calculation operands.
Megatron-LM/megatron/core/tensor_parallel/cross_entropy.py:85
↓ 2 callersFunctionprepare_input_tensors_for_wgrad_compute
Ensure grad_output is stored in a contiguous buffer.
Megatron-LM/megatron/core/utils.py:1044
↓ 2 callersFunctionprepare_trajectories
Pad trajectories and extract the generation masks. Args: rollouts: Rollouts to extract trajectories from. tokenizer: Tokenizer to
Megatron-LM/megatron/rl/rl_utils.py:922
↓ 2 callersFunctionpreprocess
(dataset_path: str, config: JsonQADatasetConfig)
Megatron-LM/tools/retro/sft/dataset_conv.py:181
↓ 2 callersMethodpreprocess
Preprocess token routing for dispatch. This method preprocesses the hidden states and routing probabilities for the token dispatcher.
Megatron-LM/megatron/core/transformer/moe/moe_layer.py:191
↓ 2 callersMethodpreprocess_for_fine_grained_offloading
Preprocess for fine-grained activation offloading.
Megatron-LM/megatron/core/models/gpt/gpt_model.py:417
↓ 2 callersFunctionpreprocess_fsdp_dtensor_state_dict
(args, raw_state_dict, model)
Megatron-LM/megatron/training/checkpointing.py:890
↓ 2 callersFunctionprint_mimo_structure
Print a clean summary of MIMO model structure showing components and their types.
Megatron-LM/examples/mimo/utils/logging.py:9
↓ 2 callersMethodprint_processing_stats
(self, count, proc_start, total_bytes_processed)
Megatron-LM/tools/preprocess_data.py:122
↓ 2 callersMethodprint_stats
Return a string describing the data distribution.
Megatron-LM/megatron/core/rerun_state_machine.py:1145
↓ 2 callersFunctionprocess_batch
Process batch and produce inputs for the model.
Megatron-LM/tasks/finetune_utils.py:26
↓ 2 callersFunctionprocess_epub
(filename: str)
gui-rl/desktop_env/evaluators/metrics/others.py:18
↓ 2 callersFunctionprocess_futures
(futures: dict[concurrent.futures.Future, str])
swe-rl/mini-swe-agent/src/minisweagent/run/extra/swebench.py:235
↓ 2 callersFunctionprocess_image
Resize + re-encode screenshot and return base64 PNG.
gui-rl/agents/qwen35_agent.py:26
↓ 2 callersFunctionprocess_image
Process an image for Qwen VL models (thinking variant). Uses a tighter resize cap consistent with the thinking DUN agent.
gui-rl/agents/qwen3vl_agent.py:24
↓ 2 callersFunctionprocess_rollout_data
(args, rollout_data_ref, dp_rank, dp_size)
slime/slime/utils/data.py:270
↓ 2 callersFunctionprocess_vision_info
(prompt, processor)
slime/slime/utils/processing_utils.py:33
↓ 2 callersMethodput_nowait
Put an item into the queue without blocking
Megatron-LM/megatron/core/inference/utils.py:196
↓ 2 callersMethodquery
(self, messages: list[dict[str, str]], **kwargs)
swe-rl/mini-swe-agent/src/minisweagent/__init__.py:48
↓ 2 callersMethodquery
(self, messages: list[dict[str, str]], **kwargs)
swe-rl/mini-swe-agent/src/minisweagent/models/openrouter_model.py:92
↓ 2 callersFunctionquick_geglu_back
Backward helper for Quick-GEGLU. Args: g (torch.Tensor): Upstream gradient tensor. y (torch.Tensor): Input tensor used in the for
Megatron-LM/megatron/core/fusions/fused_bias_geglu.py:221
↓ 2 callersFunctionrank0_echo
(message)
Megatron-LM/tools/checkpoint/checkpoint_inspector.py:51
↓ 2 callersFunctionread_file
(path)
slime/slime/utils/data.py:25
↓ 2 callersFunctionread_thunderbird_folder_file
(path: str)
gui-rl/desktop_env/evaluators/metrics/thunderbird.py:157
↓ 2 callersFunctionrealworldqa_eval
Run RealWorldQA evaluation.
Megatron-LM/examples/multimodal/evaluation/evaluate_realworldqa.py:32
↓ 2 callersMethodrecord
(self, top_indices)
slime/slime/utils/routing_replay.py:22
← previousnext →2,201–2,300 of 10,561, ranked by callers