MCPcopy Create free account

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

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

↓ 1 callersFunctionremove_articles
(text)
slime/examples/search-r1/qa_em_format.py:22
↓ 1 callersFunctionremove_articles
(text)
slime/slime/rollout/rm_hub/f1.py:8
↓ 1 callersFunctionremove_boxed
(s)
slime/slime/rollout/rm_hub/math_utils.py:407
↓ 1 callersMethodremove_codes
Remove added codes after adding to index. Args: config (RetroPreprocessingConfig): Retro preprocessing config.
Megatron-LM/megatron/core/datasets/retro/index/indexes/faiss_par_add.py:220
↓ 1 callersFunctionremove_embeddings
Remove embeddings after training. Args: config (RetroPreprocessingConfig): Retro preprocessing config.
Megatron-LM/megatron/core/datasets/retro/index/build.py:195
↓ 1 callersFunctionremove_iter_ckpts
(_iter_ckpts)
Megatron-LM/megatron/training/checkpointing.py:755
↓ 1 callersFunctionremove_punc
(text)
Megatron-LM/tools/retro/text_generation/evaluate.py:24
↓ 1 callersFunctionremove_punc
(text)
slime/examples/search-r1/qa_em_format.py:28
↓ 1 callersFunctionremove_punc
(text)
slime/slime/rollout/rm_hub/f1.py:14
↓ 1 callersFunctionremove_sharded_tensors
determine the appropriate sharding strategy and delegate removal to the sharded strategy
Megatron-LM/megatron/core/dist_checkpointing/serialization.py:304
↓ 1 callersMethodremove_sharded_tensors
Remove all tensors whose key starts with key_prefix
Megatron-LM/megatron/core/dist_checkpointing/strategies/base.py:173
↓ 1 callersFunctionremove_te_extra_state
(state_dict)
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/fully_shard.py:209
↓ 1 callersFunctionreplace_expert_index_in_key
Replace expert index in key with new index corresponding to the current rank
Megatron-LM/megatron/core/transformer/fsdp_dtensor_checkpoint.py:128
↓ 1 callersMethodreplace_value_with_repetition
Replace every occurrence of value V in the input array with R repetitions of W. Args: arr (Array): Input array to be mod
Megatron-LM/examples/multimodal/dataset_helpers.py:552
↓ 1 callersFunctionreport_current_memory_info
Report current memory usage.
Megatron-LM/examples/post_training/modelopt/validate.py:82
↓ 1 callersFunctionreport_draft_acceptance_length
Report MTBench acceptance length.
Megatron-LM/megatron/post_training/non_loss_data_func.py:11
↓ 1 callersFunctionreport_memory
Simple GPU memory report.
Megatron-LM/megatron/training/utils.py:278
↓ 1 callersMethodrequires_returning_mismatch_keys
Whether a given strict option results in extra return value from the `load` function.
Megatron-LM/megatron/core/dist_checkpointing/validation.py:105
↓ 1 callersMethodreroute_samples_to_hdp_ranks
Reroutes the sub-samples to the correct rank after scheduling. For each key in the batch dict, we perform an all-to-all communicatio
Megatron-LM/megatron/core/datasets/data_schedule.py:136
↓ 1 callersFunctionres_to_sample
Convert InteractionResult to Sample format for slime training. This function transforms the tau-bench interaction result into the format
slime/examples/tau-bench/generate_with_tau.py:39
↓ 1 callersMethodreset
( self, lease_id: str, task_meta: dict[str, Any], run_ctx: dict[str, Any],
terminal-rl/env_client.py:53
↓ 1 callersMethodreset
This method is called to reset the metrics state of the instance It is generally called from within elapsed() after extracting per rank metri
Megatron-LM/megatron/core/utils.py:1447
↓ 1 callersMethodreset
Forget all data.
Megatron-LM/megatron/core/rerun_state_machine.py:1137
↓ 1 callersMethodreset
Reset metadata in bucket group in preparation for the next iteration of training.
Megatron-LM/megatron/core/distributed/param_and_grad_buffer.py:204
↓ 1 callersMethodreset
Reset the internal RNG state.
Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/utils.py:244
↓ 1 callersMethodreset
Reset counter
Megatron-LM/megatron/core/inference/utils.py:27
↓ 1 callersMethodreset
Reset the metadata for the next batch.
Megatron-LM/megatron/core/inference/contexts/attention_context/mha_metadata.py:121
↓ 1 callersMethodreset
Reset the internal RNG state.
Megatron-LM/megatron/core/extensions/transformer_engine.py:2007
↓ 1 callersMethodreset
(self, name=None)
slime/slime/utils/timer.py:34
↓ 1 callersMethodreset_batch_size_offset
Reset batch size offset to 0. No-op for dynamic batching.
Megatron-LM/megatron/core/inference/contexts/base_context.py:40
↓ 1 callersMethodreset_cache_dir
(self, cache_dir: str)
gui-rl/desktop_env/controllers/setup.py:55
↓ 1 callersMethodreset_cudagraph_mapping
Resets only the CUDA graph mapping tensor.
Megatron-LM/megatron/core/inference/contexts/attention_context/mamba_metadata.py:67
↓ 1 callersMethodreset_eval_scores
(self)
openclaw-rl/openclaw_api_server.py:375
↓ 1 callersMethodreset_eval_scores
(self)
openclaw-rl/oel/openclaw_oel_api_server.py:1574
↓ 1 callersMethodreset_eval_scores
(self)
openclaw-opd/openclaw_opd_api_server.py:921
↓ 1 callersFunctionreset_hybrid_ep_buffer
Reset the HybridEP buffer
Megatron-LM/megatron/core/transformer/moe/fused_a2a.py:324
↓ 1 callersMethodreset_iteration_counters
Reset per-iteration counters.
Megatron-LM/megatron/rl/rl_utils.py:100
↓ 1 callersMethodreset_parameters
(self)
Megatron-LM/megatron/legacy/model/fused_layer_norm.py:69
↓ 1 callersMethodreset_parameters
(self)
Megatron-LM/megatron/core/fusions/fused_layer_norm.py:122
↓ 1 callersMethodreset_parameters
Reset the parameters.
Megatron-LM/megatron/core/ssm/gated_delta_net.py:224
↓ 1 callersFunctionreset_persistent_async_worker
()
Megatron-LM/megatron/training/async_utils.py:69
↓ 1 callersFunctionresolveConfig
(api: OpenClawPluginApi)
extensions/rl-training-headers/index.ts:9
↓ 1 callersFunctionresolve_prm_sglang_url
(args: Any)
terminal-rl/rollout_agent.py:109
↓ 1 callersFunctionresolve_rollout_non_think_mode
(args: Any)
terminal-rl/rollout_agent.py:61
↓ 1 callersMethodresolved_teacher_model
Return the teacher model name, defaulting to policy model if unset.
openclaw-tinker/config.py:66
↓ 1 callersMethodrestore
(self, tag: str)
slime/slime/utils/tensor_backper.py:38
↓ 1 callersMethodrestore_from_cpu
Restore optimizer state from CPU back to GPU for training.
Megatron-LM/megatron/core/optimizer/optimizer.py:1419
↓ 1 callersMethodresume
Resume engine by reallocating context's GPU state.
Megatron-LM/megatron/core/inference/engines/dynamic_engine.py:570
↓ 1 callersMethodresume_submission
(self)
openclaw-rl/openclaw_rollout.py:71
↓ 1 callersMethodresume_submission
(self)
openclaw-rl/oel/openclaw_oel_rollout.py:84
↓ 1 callersMethodresume_submission
(self)
openclaw-opd/openclaw_opd_rollout.py:75
↓ 1 callersMethodresume_submission
(self)
openclaw-combine/openclaw_combine_rollout.py:75
↓ 1 callersMethodresume_submission
(self)
openclaw-combine/openclaw_combine_select_rollout.py:91
↓ 1 callersMethodretrieve_from_text
Get tokens from text by looking up in radix tree or using tokenizer. Also fetches weight version from worker during this operation.
slime/slime/router/middleware_hub/radix_tree.py:579
↓ 1 callersMethodretrieve_write_results
Turn the latest dict including write results from `self.results_queue` into a single results lists. Includes error check.
Megatron-LM/megatron/core/dist_checkpointing/strategies/filesystem_async.py:424
↓ 1 callersMethodretro_decoder_cross_attention
Cross attention for Retro decoder. Notation: ns : Sequence length. bs : Batch size. d : Hidden size.
Megatron-LM/megatron/legacy/model/transformer.py:1061
↓ 1 callersMethodretro_encoder_cross_attention
Cross attention for Retro encoder. Notation: ns : Sequence length. bs : Batch size. d : Hidden size.
Megatron-LM/megatron/legacy/model/transformer.py:994
↓ 1 callersFunctionretro_generate
Given prompts and input parameters, run inference and return: tokens: prompts plus the generated tokens. lengths: length of the prompt +
Megatron-LM/tools/retro/text_generation/retro_api.py:142
↓ 1 callersFunctionretro_generate_tokens_probs_and_return_on_first_stage
Main token generation function. Args: model: no interleaving is supported. tokens: prompt tokens extended to be of size [b, max-s
Megatron-LM/tools/retro/text_generation/retro_generation.py:20
↓ 1 callersMethodreturn_layer_name_and_number
Helper function to return layer name and number Given an input layer e.g decoder.layers.2.self_attention.linear_qkv.weight, this funct
Megatron-LM/megatron/core/export/trtllm/trtllm_layers.py:56
↓ 1 callersMethodrevert_to_snapshot
Method to revert the VM to a given snapshot.
gui-rl/desktop_env/providers/base.py:33
↓ 1 callersFunctionreward_func
The reward function for retrieval-based question answering. Args: args: the arguments sample: the sample to evaluate
slime/examples/search-r1/generate_with_search.py:250
↓ 1 callersMethodrewrite
Generates the rewrited solution.
slime/examples/multi_agent/agent_system.py:115
↓ 1 callersFunctionrewrite_worker
(args, previous_solutions, problem_statement, worker_id)
slime/examples/multi_agent/agent_system.py:165
↓ 1 callersMethodrollout
(self, request: RolloutRequest)
Megatron-LM/megatron/rl/agent/api.py:120
↓ 1 callersMethodrollout
(self, request: RolloutRequest)
Megatron-LM/megatron/rl/agent/api.py:154
↓ 1 callersFunctionround_to_quantized_type_dtype
( tensor, dtype, cast_to_original_dtype=False, )
slime/tools/convert_hf_to_int4_direct.py:76
↓ 1 callersFunctionround_to_quantized_type_dtype
( tensor, dtype, cast_to_original_dtype=False, )
slime/slime/backends/megatron_utils/megatron_to_hf/processors/quantizer_compressed_tensors.py:181
↓ 1 callersMethodrouting
Top-k routing function Args: logits (torch.Tensor): Logits tensor after gating. Returns: probs (torch.Tensor
Megatron-LM/megatron/core/transformer/moe/router.py:485
↓ 1 callersFunctionrun
(agent_cls: type[Agent], cls_args: dict, port: int)
Megatron-LM/megatron/rl/server/agent/fastapi_env_server.py:173
↓ 1 callersMethodrun
(self, task: str, **kwargs)
swe-rl/mini-swe-agent/src/minisweagent/agents/interactive_textual.py:279
↓ 1 callersFunctionrun_all_tests
Run all tests.
slime/tests/test_fused_experts_backward.py:536
↓ 1 callersMethodrun_engine_with_coordinator
Continually steps the engine asynchronously.
Megatron-LM/megatron/core/inference/engines/dynamic_engine.py:1433
↓ 1 callersMethodrun_episode
(self, user_msg: Any)
terminal-rl/agent_runner.py:78
↓ 1 callersFunctionrun_eval
(config, iteration=None)
Megatron-LM/examples/multimodal/run_text_generation.py:690
↓ 1 callersMethodrun_evaluation
(self, request: EvaluationRequest)
Megatron-LM/megatron/rl/agent/api.py:225
↓ 1 callersMethodrun_gpu_forward
(self, recompute_granularity, use_transformer_engine)
Megatron-LM/tests/unit_tests/transformer/test_retro_attention.py:113
↓ 1 callersFunctionrun_hf_vision
Run HF vision model.
Megatron-LM/examples/multimodal/model_converter/vision_model_tester.py:74
↓ 1 callersFunctionrun_inference
Add requests to engine and generate tokens. Args: requests (List[Request]): Requests that are to be added and processed. engine (
Megatron-LM/examples/inference/gpt/gpt_dynamic_inference.py:230
↓ 1 callersFunctionrun_instances_with_retry
Cross-process throttle + exponential backoff on 429 for RunInstances.
gui-rl/desktop_env/providers/volcengine/manager.py:113
↓ 1 callersFunctionrun_mcore_vision
Run mcore vision model.
Megatron-LM/examples/multimodal/model_converter/vision_model_tester.py:24
↓ 1 callersFunctionrun_model_ref_with_capture
Runs the model in reference mode and captures outputs and gradients. Args: model: The transformer model to run. input_tensor
Megatron-LM/tests/unit_tests/transformer/test_submodule_callables.py:22
↓ 1 callersFunctionrun_model_submodules_with_capture
Runs the model with all-to-all overlap optimization and captures outputs and gradients. Args: model: The transformer model to run.
Megatron-LM/tests/unit_tests/transformer/test_submodule_callables.py:48
↓ 1 callersFunctionrun_mtp_layer_a2a_overlap_with_capture
Runs the model with all-to-all overlap optimization and captures outputs and gradients. Args: model: The transformer model to run.
Megatron-LM/tests/unit_tests/a2a_overlap/test_schedule_layer_1f1b.py:160
↓ 1 callersFunctionrun_mtp_layer_ref_with_capture
Runs the model in reference mode and captures outputs and gradients. Args: model: The transformer model to run. input_tensor
Megatron-LM/tests/unit_tests/a2a_overlap/test_schedule_layer_1f1b.py:113
↓ 1 callersFunctionrun_one_commenting
Run a single commenting session. Returns True if completed.
openclaw-test/teacher_chat.py:213
↓ 1 callersFunctionrun_one_grading
Run a single grading session. Returns True if completed.
openclaw-test/TA_chat.py:208
↓ 1 callersFunctionrun_one_problem
Run a single homework problem session. Returns True if completed.
openclaw-test/student_chat.py:178
↓ 1 callersMethodrun_realtime_tests
Performs a consistency check. This function makes sure that tensors across devices are the same during an experiment. This is often n
Megatron-LM/megatron/core/transformer/attention.py:1113
↓ 1 callersFunctionrun_router
Run the Slime router with the specified configuration.
slime/slime/router/router.py:17
↓ 1 callersFunctionrun_server
()
Megatron-LM/megatron/rl/server/agent/fastapi_env_server.py:176
↓ 1 callersFunctionrun_student_training_session
Run one student training session: GPT-4.1 plays lazy student, policy model solves homework. All turns sent as MAIN to trigger training.
openclaw-rl/oel/eval/gsm8k_personal_agent.py:330
↓ 1 callersFunctionrun_teacher_training_session
Run one teacher training session: GPT-4.1 plays strict teacher, policy model grades homework. All turns sent as MAIN to trigger training.
openclaw-rl/oel/eval/gsm8k_personal_agent.py:418
↓ 1 callersFunctionrun_tests
Run all tests, iterating over `num_cuda_graphs` and `NUM_TRIALS`.
Megatron-LM/tests/functional_tests/test_cases/gpt/gpt_dynamic_inference_tp1_pp1_583m_cuda_graphs_validation/cuda_graphs.py:35
↓ 1 callersFunctionrun_training_loop
Main training loop: drain -> train -> weight-sync -> repeat. Args: training_client: FiretitanTrainingClient instance. weight_sync
openclaw-fireworks/fireworks_training_loop.py:144
↓ 1 callersFunctionrun_training_round
Run one round of training sessions (student or teacher). Returns list of session results.
openclaw-rl/oel/eval/gsm8k_personal_agent.py:508
↓ 1 callersFunctionsafely_get_direct_li_playwright
(selector)
gui-rl/desktop_env/evaluators/getters/chrome.py:1647
↓ 1 callersFunctionsafely_get_direct_text_nodes_playwright
Extract all direct text node contents under the specified selector element (excluding text inside child div, span, etc.). Ret
gui-rl/desktop_env/evaluators/getters/chrome.py:1621
↓ 1 callersFunctionsafely_get_only_child_text_content
(selector)
gui-rl/desktop_env/evaluators/getters/chrome.py:1658
↓ 1 callersFunctionsample_to_datum
Convert one sample + scalar advantage into a Tinker Datum (RL / OPD). For OPD samples with teacher_logprobs, the advantage is augmented with
openclaw-tinker/data_formatter.py:106
← previousnext →4,901–5,000 of 10,561, ranked by callers