Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/Gen-Verse/OpenClaw-RL
/ functions
Functions
10,561 in github.com/Gen-Verse/OpenClaw-RL
⨍
Functions
10,561
◇
Types & classes
1,405
↳
Endpoints
144
↓ 1 callers
Function
gather_split_1d_tensor
Opposite of split_tensor_into_1d_equal_chunks. Gather values from tensor model parallel ranks. Returns a new Tensor with the gathered data.
Megatron-LM/megatron/core/tensor_parallel/utils.py:79
↓ 1 callers
Method
gating
Forward pass of the router gate. Args: input (torch.Tensor): Input tensor. Returns: torch.Tensor: Logits ten
Megatron-LM/megatron/core/transformer/moe/router.py:78
↓ 1 callers
Function
gelu_impl
OpenAI's gelu implementation.
Megatron-LM/megatron/legacy/model/utils.py:47
↓ 1 callers
Function
gelu_impl
OpenAI's gelu implementation.
Megatron-LM/megatron/core/transformer/utils.py:55
↓ 1 callers
Method
gen_mask
(self, image_size, mask_size, mask_type, patch_size)
Megatron-LM/megatron/legacy/data/vit_dataset.py:111
↓ 1 callers
Method
gen_multi_turn_loss_mask_qwen
( self, messages: list[dict], tools: list[dict] = None )
slime/slime/utils/mask_utils.py:48
↓ 1 callers
Function
generate
( inference_engine: Union[StaticInferenceEngine, DynamicInferenceEngine], sampling_params: SamplingPar
Megatron-LM/tools/run_inference_performance_test.py:150
↓ 1 callers
Function
generate
( inference_engine: StaticInferenceEngine, sampling_params: SamplingParams, prompts: List[str] )
Megatron-LM/examples/inference/gpt/gpt_static_inference.py:110
↓ 1 callers
Function
generate
Generate using traditional SGLang router with token-based workflow
slime/slime/rollout/sglang_rollout.py:102
↓ 1 callers
Method
generate
The abstract backend's generate function. To define a new backend, implement this and return the outputs as a dictionary. Returns:
Megatron-LM/megatron/core/inference/engines/abstract_engine.py:9
↓ 1 callers
Function
generate_and_rm_group
( args: Namespace, group: list[Sample], sampling_params: dict[str, Any], evaluation: bool = False )
slime/slime/rollout/sglang_rollout.py:261
↓ 1 callers
Function
generate_and_write_samples
(model)
Megatron-LM/tools/run_vlm_text_generation.py:160
↓ 1 callers
Function
generate_and_write_samples
Generate text and write to an output file.
Megatron-LM/examples/multimodal/run_text_generation.py:439
↓ 1 callers
Function
generate_and_write_samples_conditional
(model)
Megatron-LM/tools/retro/text_generation/retro_text_generation.py:226
↓ 1 callers
Function
generate_and_write_samples_conditional
(model)
Megatron-LM/examples/academic_paper_scripts/detxoify_lm/generate_samples_gpt.py:218
↓ 1 callers
Function
generate_and_write_samples_unconditional
(model)
Megatron-LM/examples/academic_paper_scripts/detxoify_lm/generate_samples_gpt.py:209
↓ 1 callers
Method
generate_cuda_graph_batch_dimensions_list
Generate CUDA graph batch dimensions. This function constructs CUDA graph batch dimensions for different token counts and re
Megatron-LM/megatron/core/inference/batch_dimensions_utils.py:190
↓ 1 callers
Function
generate_dynamic
( args: argparse.Namespace, inference_requests: List[InferenceRequest], inference_engine: DynamicI
Megatron-LM/tools/run_inference_performance_test.py:224
↓ 1 callers
Method
generate_initial_solution
Generates the first solution attempt.
slime/examples/multi_agent/agent_system.py:103
↓ 1 callers
Function
generate_masked_orthogonal_rank_groups
r"""Generate orthogonal parallel groups based on the parallel size and mask. Arguments: world_size (int): world size parallel_si
Megatron-LM/megatron/core/parallel_state.py:246
↓ 1 callers
Function
generate_new_vm_name
(vms_dir, os_type)
gui-rl/desktop_env/providers/virtualbox/manager.py:36
↓ 1 callers
Function
generate_new_vm_name
(vms_dir, os_type)
gui-rl/desktop_env/providers/vmware/manager.py:50
↓ 1 callers
Method
generate_output_tokens_dynamic_batch
Synchronous wrapper for `self.async_generate_output_tokens_dynamic_batch.
Megatron-LM/megatron/core/inference/text_generation_controllers/text_generation_controller.py:950
↓ 1 callers
Function
generate_prompt
Generating few-shot prompts.
Megatron-LM/examples/post_training/modelopt/mmlu.py:112
↓ 1 callers
Method
generate_reminders
Generate all reminders.
Megatron-LM/tests/test_utils/python_scripts/auto_reminder_github.py:248
↓ 1 callers
Function
generate_response
(args, prompt, key)
slime/examples/multi_agent/agent_system.py:14
↓ 1 callers
Function
generate_rewriter_template
Dynamically generate rewrite templates based on the number of solutions.
slime/examples/multi_agent/prompts.py:7
↓ 1 callers
Function
generate_rollout
An example to implement the generate_rollout function for an rule based rm rollout generation. Args: args: the whole args rollout
slime/slime/rollout/sglang_rollout.py:542
↓ 1 callers
Function
generate_rollout_async
(args, rollout_id: int, data_buffer, evaluation: bool = False)
slime/slime_plugins/rollout_buffer/rollout_buffer_example.py:207
↓ 1 callers
Function
generate_rollout_async
An example to implement the generate_rollout function for an rule based rm rollout generation. Args: args: the whole args rollout
slime/slime/rollout/sglang_rollout.py:330
↓ 1 callers
Function
generate_samples
Text generation using a trained vision language model. This is an example for the COCO dataset.
Megatron-LM/tools/run_vlm_text_generation.py:87
↓ 1 callers
Function
generate_samples
Text generation using a trained vision language model.
Megatron-LM/examples/multimodal/run_text_generation.py:156
↓ 1 callers
Function
generate_samples_conditional
(model)
Megatron-LM/tools/retro/text_generation/retro_text_generation.py:134
↓ 1 callers
Function
generate_samples_conditional
(model)
Megatron-LM/examples/academic_paper_scripts/detxoify_lm/generate_samples_gpt.py:156
↓ 1 callers
Function
generate_samples_unconditional
(model)
Megatron-LM/examples/academic_paper_scripts/detxoify_lm/generate_samples_gpt.py:119
↓ 1 callers
Function
generate_select_template
Dynamically generate select templates based on the number of solutions.
slime/examples/multi_agent/prompts.py:36
↓ 1 callers
Function
generate_student_message
Have the student LLM decide what to say next.
openclaw-test/student_chat.py:150
↓ 1 callers
Function
generate_ta_message
Have the TA LLM decide what to say next.
openclaw-test/TA_chat.py:180
↓ 1 callers
Function
generate_teacher_message
Have the teacher LLM decide what to say next.
openclaw-test/teacher_chat.py:185
↓ 1 callers
Function
generate_trajectory
Entry point with total-rollout timeout watchdog. Algorithm wrappers (generate_with_swe_{kiro,rllm,prorl}) delegate here.
swe-rl/generate_with_swe_trajectory.py:1116
↓ 1 callers
Method
generate_using_dynamic_engine
Generate using dynamic engine Generate using dynamic engine. Args: prompts (List[str]): All the prompts as a list of str
Megatron-LM/megatron/core/inference/engines/static_engine.py:199
↓ 1 callers
Method
generate_using_legacy_static_engine
The megatron core inference backend generate function This backend returns the output generations as a list. Args: prompt
Megatron-LM/megatron/core/inference/engines/static_engine.py:240
↓ 1 callers
Method
generator
Creates an AsyncGenerator over the stream queue
Megatron-LM/megatron/core/inference/async_stream.py:55
↓ 1 callers
Method
get_active_sequence_lengths
Total sequence length (query + key) for active requests.
Megatron-LM/megatron/core/inference/contexts/dynamic_context.py:761
↓ 1 callers
Function
get_actual_sequence_lengths
Get actual sequence lengths for pre-padded sequences. Args: sequences: Tensor of shape [batch_size, seq_len] with pre-padded sequences
Megatron-LM/megatron/rl/rl_utils.py:1129
↓ 1 callers
Function
get_added_code_paths
Get paths to all saved encodings. Args: config (RetroPreprocessingConfig): Retro preprocessing config. Returns: Paths of all
Megatron-LM/megatron/core/datasets/retro/index/utils.py:117
↓ 1 callers
Function
get_advantages_and_returns_batch
Batched GAE with CP support. Input: total_lengths: list[int], each sample's total_len response_lengths: list[int], each
slime/slime/utils/ppo_utils.py:374
↓ 1 callers
Function
get_all_subjects
Return all MMLU subjects.
Megatron-LM/examples/post_training/modelopt/mmlu.py:37
↓ 1 callers
Method
get_all_timers_string
Returns the output string with logged timer values according to configured options. Args: names (List[str]): Names of the timers
Megatron-LM/megatron/core/timers.py:367
↓ 1 callers
Function
get_analytics_per_pipeline
(pipeline_id: int)
Megatron-LM/tests/test_utils/python_scripts/dashboard.py:128
↓ 1 callers
Function
get_arg_value
(name)
slime/slime/utils/arguments.py:2034
↓ 1 callers
Function
get_args
()
Megatron-LM/tools/preprocess_mmdata.py:73
↓ 1 callers
Function
get_args
()
Megatron-LM/tools/preprocess_data_nmt.py:33
↓ 1 callers
Function
get_args
()
Megatron-LM/tools/preprocess_data.py:198
↓ 1 callers
Function
get_args
()
Megatron-LM/tools/merge_datasets.py:18
↓ 1 callers
Function
get_args
()
slime/tools/convert_hf_to_torch_dist.py:109
↓ 1 callers
Function
get_async_loop
()
slime/slime/utils/async_utils.py:27
↓ 1 callers
Function
get_audio_submodules_spec
Get the submodule spec for the audio modality.
Megatron-LM/tests/unit_tests/models/test_mimo_model.py:83
↓ 1 callers
Function
get_base_gpu_id
(args, rank)
slime/slime/backends/sglang_utils/sglang_engine.py:23
↓ 1 callers
Function
get_batch
Build the batch.
Megatron-LM/pretrain_bert.py:73
↓ 1 callers
Function
get_batch
Build the batch.
Megatron-LM/pretrain_t5.py:144
↓ 1 callers
Function
get_batch
Generate a batch.
Megatron-LM/pretrain_gpt.py:43
↓ 1 callers
Function
get_batch
Generate a batch. Args: data_iterator: Iterable dataset. Returns: sample: A data sample with images, tokens, etc.
Megatron-LM/pretrain_vlm.py:280
↓ 1 callers
Function
get_batch
Generate a batch.
Megatron-LM/pretrain_mamba.py:38
↓ 1 callers
Function
get_batch
Build the batch.
Megatron-LM/pretrain_vision_dino.py:23
↓ 1 callers
Function
get_batch
Build the batch.
Megatron-LM/pretrain_vision_inpaint.py:36
↓ 1 callers
Function
get_batch
Build the batch.
Megatron-LM/pretrain_vision_classify.py:40
↓ 1 callers
Function
get_batch
Generate a batch for MIMO model training. Args: data_iterator: Iterator over the dataset Returns: tuple: Batch data for mode
Megatron-LM/examples/mimo/train.py:83
↓ 1 callers
Function
get_batch
Generate a batch. For OfflineDataset, the aux_hidden_states and final hidden_states from the base model are loaded for offline speculative mo
Megatron-LM/examples/post_training/modelopt/finetune.py:387
↓ 1 callers
Function
get_batch
Generate a batch
Megatron-LM/examples/academic_paper_scripts/detxoify_lm/finetune_gpt.py:41
↓ 1 callers
Function
get_batch
Generate a batch Note: attn_mask_type in layer_specs.py sets the attention mask. Attention mask is None here.
Megatron-LM/examples/multimodal/train.py:33
↓ 1 callers
Method
get_batch
(self, seq_length, micro_batch_size)
Megatron-LM/tests/unit_tests/test_fp8_param.py:163
↓ 1 callers
Method
get_batch
(self, seq_length, micro_batch_size, cp_size)
Megatron-LM/tests/unit_tests/transformer/test_cuda_graphs.py:1149
↓ 1 callers
Function
get_batch_on_this_hybrid_cp_rank
Slice batch input along sequence dimension into multiple chunks, which are parallelized across GPUs in a context parallel group.
Megatron-LM/megatron/core/utils.py:2053
↓ 1 callers
Method
get_batch_per_block
(sq, sk, b, np)
Megatron-LM/megatron/legacy/model/fused_softmax.py:228
↓ 1 callers
Method
get_batch_per_block
Return CUDA kernel's batch-per-block parameter for masked softmax. Args: sq (int): Query sequence length. sk (int): K
Megatron-LM/megatron/core/fusions/fused_softmax.py:345
↓ 1 callers
Function
get_bert_embedders
(config)
Megatron-LM/tools/retro/preprocess_data.py:83
↓ 1 callers
Function
get_bert_layer_with_transformer_engine_spec
Use this spec to use lower-level Transformer Engine modules (required for fp8 training). Returns: ModuleSpec: Module specification with T
Megatron-LM/megatron/core/models/bert/bert_layer_specs.py:42
↓ 1 callers
Function
get_bert_tokenizer
Bert (Wordpiece) tokenizer.
Megatron-LM/tools/retro/preprocess_data.py:177
↓ 1 callers
Function
get_bias_dropout_add
(training)
Megatron-LM/megatron/legacy/model/transformer.py:853
↓ 1 callers
Function
get_block_nload
Compute number of blocks to load. This is computed by multiplying the total number of available blocks with the fraction of blocks to load.
Megatron-LM/megatron/core/datasets/retro/index/build.py:65
↓ 1 callers
Function
get_block_samples_mapping
Get samples mapping for a dataset over fixed size blocks. This function also requires a dataset of the titles for the source documents since their
Megatron-LM/megatron/legacy/data/realm_dataset_utils.py:112
↓ 1 callers
Function
get_blocks
Divide range [0, num_samples) to sequence of block ranges. This is a core method within the concept of block processing. The idea is to divid
Megatron-LM/megatron/core/datasets/retro/utils.py:129
↓ 1 callers
Function
get_build_analytics
(pipeline_id: int)
Megatron-LM/tests/test_utils/python_scripts/dashboard.py:25
↓ 1 callers
Method
get_cached_cos_sin
Get cached cos and sin values.
Megatron-LM/megatron/core/models/common/embeddings/yarn_rotary_pos_embedding.py:203
↓ 1 callers
Function
get_calib_dataloader
Return a dataloader for calibration.
Megatron-LM/examples/post_training/modelopt/prune.py:140
↓ 1 callers
Function
get_calib_dataloader
Return a dataloader for calibration.
Megatron-LM/examples/post_training/modelopt/quantize.py:136
↓ 1 callers
Function
get_calibration_dataset
(tokenizer, num_samples, seq_len, local_data_path)
slime/tools/convert_hf_to_int4.py:27
↓ 1 callers
Function
get_causal_conv1d_version_str
()
Megatron-LM/megatron/core/utils.py:443
↓ 1 callers
Function
get_cli_requests
( args: Namespace, tokenizer: Any, sampling_params: Optional[SamplingParams] = None )
Megatron-LM/examples/inference/gpt/utils.py:249
↓ 1 callers
Function
get_client
Get shared client for connection pooling (like SLIME).
slime/examples/strands_sglang/generate_with_strands.py:28
↓ 1 callers
Function
get_comp_stream
Get the stream for computation
Megatron-LM/megatron/core/pipeline_parallel/utils.py:338
↓ 1 callers
Function
get_compare_function
()
gui-rl/desktop_env/evaluators/metrics/chrome.py:331
↓ 1 callers
Method
get_completed_groups
(self)
openclaw-rl/openclaw_rollout.py:76
↓ 1 callers
Method
get_completed_groups
(self)
openclaw-rl/oel/openclaw_oel_rollout.py:89
↓ 1 callers
Method
get_completed_groups
(self)
openclaw-opd/openclaw_opd_rollout.py:80
↓ 1 callers
Method
get_completed_groups
(self)
openclaw-combine/openclaw_combine_rollout.py:80
↓ 1 callers
Method
get_completed_groups
(self)
openclaw-combine/openclaw_combine_select_rollout.py:96
↓ 1 callers
Method
get_config_for_layer
Get the config for the given layer number. Based on the layer number, the corresponding block config is returned, overriding
Megatron-LM/megatron/core/transformer/heterogeneous/heterogeneous_config.py:229
↓ 1 callers
Function
get_config_leaf_field_names
(config_cls)
Megatron-LM/tools/retro/config_utils.py:569
← previous
next →
4,001–4,100 of 10,561, ranked by callers