Functions — github.com/Agent-RL/ReCall

hub / github.com/Agent-RL/ReCall / functions

Functions1,539 in github.com/Agent-RL/ReCall

⨍ Functions 1,539◇ Types & classes 237↳ Endpoints 28

↓ 2 callersMethodmake_minibatch_iterator

Make minibatch iterator for updating the actor Args: data (DataProto): a DataProto containing keys ``input_ids``

src/verl/workers/actor/megatron_actor.py:209

r"""Works like os.makedirs() but supports hdfs. Super-mkdir; create a leaf directory and all intermediate ones. Works like mkdir, except tha

src/verl/utils/hdfs_io.py:50

↓ 2 callersFunctionmath_equal

Exact match of math if and only if: 1. numerical equal: both can convert to float and are equal 2. symbolic equal: both can convert to sy

src/verl/utils/reward_score/prime_math/grader.py:174

↓ 2 callersFunctionmd5_encode

(path: str)

src/verl/utils/fs.py:36

↓ 2 callersMethodmerge

(self, other)

src/verl/utils/seqlen_balancing.py:37

↓ 2 callersFunctionmerge_dataset

(dataset_split: dict, split_symbol: list)

src/flashrag/dataset/utils.py:46

↓ 2 callersMethodmulti_gpu_encode

(self, query_list: Union[List[str], str], batch_size=64, is_query=True)

src/flashrag/retriever/encoder.py:79

↓ 2 callersFunctionnormalize

(answer, pi)

src/verl/utils/reward_score/prime_math/grader.py:119

↓ 2 callersFunctionnormalize_answer

(s)

src/verl/utils/reward_score/re_call.py:48

↓ 2 callersMethodoffload_params_to_cpu

(self)

src/verl/workers/reward_model/megatron/reward_model.py:254

↓ 2 callersFunctionpooling

(pooler_output, last_hidden_state, attention_mask=None, pooling_method="mean")

src/flashrag/retriever/utils.py:66

↓ 2 callersFunctionreduce_metrics

(metrics: Dict[str, List[Any]])

src/verl/trainer/ppo/metric_utils.py:26

↓ 2 callersFunctionremove_boxed

(s)

src/verl/utils/reward_score/re_call.py:6

↓ 2 callersFunctionremove_boxed

Remove the LaTeX boxed command from a string. Args: s: String with format "\\boxed{content}" Returns: The conten

src/verl/utils/reward_score/math_dapo.py:51

↓ 2 callersMethodremove_previous_save_local_path

(self, path)

src/verl/utils/checkpoint/checkpoint_manager.py:75

↓ 2 callersFunctionrepeat_kv

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

src/verl/models/qwen2/megatron/layers/parallel_attention.py:131

↓ 2 callersFunctionrepeat_kv

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

src/verl/models/llama/megatron/layers/parallel_attention.py:163

↓ 2 callersFunctionresolve_tokens

Resolve between max_tokens and max_new_tokens values based on priority. Returns the resolved token value or None if no valid value fo

src/flashrag/generator/utils.py:22

↓ 2 callersMethodrrf_merge

Perform Reciprocal Rank Fusion (RRF) on retrieval results. Args: results (list of list of dict): Retrieval results for m

src/flashrag/retriever/retriever.py:780

↓ 2 callersMethodrun

r"""Get refining result. Args: item: dataset item, contains question, retrieval result... Returns: str: refi

src/flashrag/refiner/refiner.py:19

↓ 2 callersFunctionset_ulysses_sequence_parallel_group

Set ulysses sequence parallel process group.

src/verl/utils/ulysses.py:29

↓ 2 callersFunctionslice_input_tensor

(x: Tensor, dim: int, padding: bool = True, group: ProcessGroup = None)

src/verl/utils/ulysses.py:116

↓ 2 callersFunctionsplit_dataset

(dataset: Dataset, split_symbol: list)

src/flashrag/dataset/utils.py:34

↓ 2 callersFunctionsplit_tuple

Split the elements in a tuple/interval, while handling well-formatted commas in large numbers

src/verl/utils/reward_score/prime_math/__init__.py:254

↓ 2 callersFunctionstrip_string

(string)

src/verl/utils/reward_score/math.py:165

↓ 2 callersFunctionstripped_string_compare

(s1, s2)

src/verl/utils/reward_score/prime_code/testing_util.py:611

↓ 2 callersMethodto_dict

Convert all information within the data sample into a dict. Information generated during the inference will be saved into output field.

src/flashrag/dataset/dataset.py:51

↓ 2 callersMethodupdate_config

(self)

src/flashrag/generator/openai_generator.py:50

↓ 2 callersMethodupdate_config

(self)

src/flashrag/generator/generator.py:35

↓ 2 callersMethodupdate_config

(self)

src/flashrag/retriever/retriever.py:131

↓ 2 callersMethodupdate_options

(self, options: Dict)

src/verl/single_controller/ray/base.py:160

↓ 2 callersMethodupdate_params

(self, updated_params)

src/verl/workers/sharding_manager/fsdp_vllm.py:174

↓ 2 callersMethodupdate_policy

Update the policy with an iterator of DataProto Args: data (DataProto): an iterator over the DataProto that returns by

src/verl/workers/actor/base.py:54

↓ 2 callersMethodupdate_sampling_params

(self, **kwargs)

src/verl/workers/rollout/vllm_rollout/fire_vllm_rollout.py:86

↓ 2 callersMethodvalidation_step

(self, batch: TensorDict)

src/verl/trainer/fsdp_sft_trainer.py:427

↓ 2 callersMethodwrap_encoder

Wrap T5 encoder to obtain a Fusion-in-Decoder model.

src/flashrag/generator/fid.py:57

↓ 2 callersMethodwrap_encoder

Wrap T5 encoder to obtain a Fusion-in-Decoder model.

src/flashrag/generator/fid.py:163

↓ 1 callersMethod__chunk_context

(self, origin_text, chunk_end_tokens)

src/flashrag/refiner/llmlingua_compressor.py:2167

↓ 1 callersMethod__get_context_prob

( self, context_list: list, token_to_word="mean", force_tokens: List[str] = []

src/flashrag/refiner/llmlingua_compressor.py:2094

↓ 1 callersMethod__getattr__

(self, attr_name: str)

src/flashrag/dataset/dataset.py:37

↓ 1 callersMethod__init__

(self, model_path, device='cpu', max_input_len=4096)

src/flashrag/generator/multimodal_generator.py:60

↓ 1 callersMethod__init__

(self, config)

src/flashrag/judger/judger.py:42

↓ 1 callersMethod__init__

(self, config)

src/flashrag/retriever/reranker.py:112

↓ 1 callersMethod__init__

(self, config, prompt_template=None, retriever=None, generator=None)

src/flashrag/pipeline/branching_pipeline.py:65

↓ 1 callersMethod__init__

(self, numel: int, numel_padded: int, dtype: torch.dtype, source: Optional[torch.Tensor] = None)

src/verl/utils/memory_buffer.py:30

↓ 1 callersMethod__init__

(self, resource_pool: RayResourcePool, ray_cls_with_init: RayClassWithInitAr

src/verl/single_controller/ray/megatron.py:44

↓ 1 callersMethod__init__

A vLLM rollout. It requires the module is supported by the vllm. Args: module: module here follows huggingface APIs c

src/verl/workers/rollout/vllm_rollout/vllm_rollout_spmd.py:68

↓ 1 callersMethod__init__

(self, config: Qwen2Config, megatron_config: ModelParallelConfig, layer_idx: int)

src/verl/models/qwen2/megatron/layers/parallel_decoder.py:37

↓ 1 callersMethod__init__

(self, input_size, num_heads, num_key_value_heads,

src/verl/models/qwen2/megatron/layers/parallel_linear.py:23

↓ 1 callersMethod__init__

(self, config: LlamaConfig, megatron_config: ModelParallelConfig, layer_idx: int)

src/verl/models/llama/megatron/layers/parallel_decoder.py:37

↓ 1 callersMethod__str__

Return a string representation of the item with its main attributes.

src/flashrag/dataset/dataset.py:70

↓ 1 callersMethod_balance_batch

Reorder the data on single controller such that each dp rank gets similar total tokens

src/verl/trainer/ppo/ray_trainer.py:782

↓ 1 callersMethod_batch_search_with_rerank

(self, *args, **kwargs)

src/flashrag/retriever/retriever.py:211

↓ 1 callersMethod_bind_worker_method

Bind the worker method to the WorkerGroup

src/verl/single_controller/base/worker_group.py:138

↓ 1 callersFunction_bind_workers_method_to_parent

Binds the methods of each worker to the WorkerDict. Note that we only bind public methods that are decorated by register

src/verl/single_controller/ray/base.py:420

↓ 1 callersMethod_block_until_all_workers_alive

(self)

src/verl/single_controller/base/worker_group.py:115

↓ 1 callersFunction_broadcast_tp_shard_tensor_gate_up

broadcast tensor in tp shards across mp_group

src/verl/models/qwen2/megatron/checkpoint_utils/qwen2_loader_depracated.py:227

↓ 1 callersFunction_broadcast_tp_shard_tensor_gate_up

broadcast tensor in tp shards across mp_group

src/verl/models/qwen2/megatron/checkpoint_utils/qwen2_saver.py:206

↓ 1 callersFunction_broadcast_tp_shard_tensor_gate_up

broadcast tensor in tp shards across mp_group

src/verl/models/mcore/saver.py:220

↓ 1 callersFunction_broadcast_tp_shard_tensor_gate_up

broadcast tensor in tp shards across mp_group

src/verl/models/mcore/loader.py:227

↓ 1 callersFunction_broadcast_tp_shard_tensor_gate_up

broadcast tensor in tp shards across mp_group

src/verl/models/llama/megatron/checkpoint_utils/llama_saver.py:206

↓ 1 callersFunction_broadcast_tp_shard_tensor_gate_up

broadcast tensor in tp shards across mp_group

src/verl/models/llama/megatron/checkpoint_utils/llama_loader_depracated.py:231

↓ 1 callersFunction_broadcast_tp_shard_tensor_qkv

broadcast tensor in tp shards across mp_group

src/verl/models/llama/megatron/checkpoint_utils/llama_saver.py:259

↓ 1 callersFunction_broadcast_tp_shard_tensor_qkv

broadcast tensor in tp shards across mp_group

src/verl/models/llama/megatron/checkpoint_utils/llama_loader_depracated.py:285

↓ 1 callersFunction_broadcast_tp_shard_tensor_vocab

broadcast tensor in tp shards across mp_group

src/verl/models/qwen2/megatron/checkpoint_utils/qwen2_loader_depracated.py:134

↓ 1 callersFunction_broadcast_tp_shard_tensor_vocab

broadcast tensor in tp shards across mp_group

src/verl/models/mcore/loader.py:134

↓ 1 callersFunction_broadcast_tp_shard_tensor_vocab

broadcast tensor in tp shards across mp_group

src/verl/models/llama/megatron/checkpoint_utils/llama_loader_depracated.py:138

↓ 1 callersMethod_build_critic_model_optimizer

(self, model_path, megatron_config

src/verl/workers/megatron_workers.py:529

↓ 1 callersMethod_build_critic_model_optimizer

(self, config)

src/verl/workers/fsdp_workers.py:678

↓ 1 callersMethod_build_dataloader

(self)

src/verl/trainer/fsdp_sft_trainer.py:122

↓ 1 callersMethod_build_model

(self, config)

src/verl/workers/fsdp_workers.py:960

↓ 1 callersMethod_build_model_optimizer

(self)

src/verl/trainer/fsdp_sft_trainer.py:185

↓ 1 callersMethod_build_param_buffer

Build the parameter buffer in each pp rank

src/verl/workers/sharding_manager/megatron_vllm.py:90

↓ 1 callersMethod_build_rm_model

(self, model_path, megatron_config: ModelParallelConfig, override_model_config)

src/verl/workers/megatron_workers.py:731

↓ 1 callersMethod_build_rollout

(self, trust_remote_code=False)

src/verl/workers/megatron_workers.py:258

↓ 1 callersMethod_build_rollout

(self, trust_remote_code=False)

src/verl/workers/fsdp_workers.py:303

↓ 1 callersMethod_build_yaml_loader

(self)

src/flashrag/config/config.py:30

↓ 1 callersMethod_calculate_lexical_unit

(self, tokens, self_info)

src/flashrag/refiner/selective_context_compressor.py:184

↓ 1 callersFunction_check_and_sort_partitions

(partitions)

src/verl/utils/seqlen_balancing.py:170

↓ 1 callersMethod_check_contain_doc

r"""Check if the index contains document content

src/flashrag/retriever/retriever.py:267

↓ 1 callersMethod_check_dir

r"""Check if the dir path exists and if there is content.

src/flashrag/retriever/index_builder.py:111

↓ 1 callersFunction_check_dispatch_mode

(dispatch_mode)

src/verl/single_controller/base/decorator.py:371

↓ 1 callersFunction_check_execute_mode

(execute_mode)

src/verl/single_controller/base/decorator.py:380

↓ 1 callersMethod_check_final_config

(self)

src/flashrag/config/config.py:94

↓ 1 callersMethod_check_resource_available

Check if the resource pool can be satisfied in this ray cluster.

src/verl/trainer/ppo/ray_trainer.py:109

↓ 1 callersMethod_collect_metrics

Collect all classes based on ```BaseMetric```.

src/flashrag/evaluator/evaluator.py:26

↓ 1 callersFunction_compute_mlflow_params_from_objects

(params)

src/verl/utils/tracking.py:146

↓ 1 callersMethod_configure_before_init

(self, register_center_name: str, rank: int)

src/verl/single_controller/base/worker.py:103

↓ 1 callersMethod_configure_with_meta

This function should only be called inside by WorkerGroup

src/verl/single_controller/base/worker.py:180

↓ 1 callersFunction_copy

(from_path: str, to_path: str, timeout: int = None)

src/verl/utils/hdfs_io.py:113

↓ 1 callersMethod_create_dataloader

(self)

src/verl/trainer/ppo/ray_trainer.py:410

↓ 1 callersMethod_download

(self)

src/verl/utils/dataset/rm_dataset.py:70

↓ 1 callersMethod_download

(self)

src/verl/utils/dataset/sft_dataset.py:72

↓ 1 callersMethod_download

(self)

src/verl/utils/dataset/multiturn_sft_dataset.py:57

↓ 1 callersFunction_exists

hdfs capable to check whether a file_path is exists

src/verl/utils/hdfs_io.py:43

↓ 1 callersFunction_expand_mask

Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.

src/verl/utils/torch_functional.py:505

↓ 1 callersFunction_expand_mask

Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.

src/verl/models/qwen2/megatron/modeling_qwen2_megatron.py:61

↓ 1 callersFunction_expand_mask

Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.

src/verl/models/llama/megatron/modeling_llama_megatron.py:61

↓ 1 callersMethod_expand_to_token_level

(self, data: DataProto, scores: torch.Tensor)

src/verl/workers/fsdp_workers.py:1080

↓ 1 callersFunction_fetch_tp_shard_tensor_gate_up

fetch gate_up tensor in tp shards

src/verl/models/qwen2/megatron/checkpoint_utils/qwen2_loader.py:138

← previous next →301–400 of 1,539, ranked by callers