Functions — github.com/Alibaba-NLP/VRAG

CodeHub MCPcopy Create free account

hub / github.com/Alibaba-NLP/VRAG / functions

Functions1,037 in github.com/Alibaba-NLP/VRAG

⨍ Functions 1,037◇ Types & classes 150↳ Endpoints 26

Methodexecute_func_rank_zero

(self, func, *args, **kwargs)

VRAG-RL/verl/single_controller/base/worker.py:215

Methodexecute_rank_zero

(self, method_name: str, *args, **kwargs)

VRAG-RL/verl/single_controller/ray/base.py:346

Methodexecute_rank_zero_sync

(self, method_name: str, *args, **kwargs)

VRAG-RL/verl/single_controller/ray/base.py:339

Methodexecute_with_func_generator

(self, func, *args, **kwargs)

VRAG-RL/verl/single_controller/base/worker.py:210

r"""Works like os.path.exists() but supports hdfs. Test whether a path exists. Returns False for broken symbolic links. Args: path (

VRAG-RL/verl/utils/hdfs_io.py:27

Functionextract_pg_from_exist

(resource_pools: Dict[str, RayResourcePool], src_role_names: List[str], resource_poo

VRAG-RL/verl/single_controller/ray/base.py:111

Functionextract_step

VRAG-RL/verl/trainer/fsdp_sft_trainer.py:59

Methodextract_tags

VRAG-RL/vrag_agent/generation.py:99

VRAG-RL/verl/utils/logger/aggregate_logger.py:37

Functionfold_batch_dim

Fold a batch dim from [bsz, xxx] into [new_bsz, bsz // new_bsz, xxx]

VRAG-RL/verl/protocol.py:115

( ctx: Any, group: dist.ProcessGroup, local_input: Tensor, scatter_dim: int,

VRAG-RL/verl/utils/ulysses.py:167

(ctx: Any, group: dist.ProcessGroup, local_tensor: Tensor, gat

VRAG-RL/verl/utils/ulysses.py:200

(self, *args, **kwargs)

VRAG-RL/verl/utils/model.py:34

(ctx, vocab_parallel_logits: torch.Tensor)

VRAG-RL/verl/utils/megatron/tensor_parallel.py:103

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

VRAG-RL/verl/models/qwen2/megatron/modeling_qwen2_megatron.py:177

Args: input_ids: input ids. shape (1, totol_nnz) position_ids: position ids. shape (batch_size, seq_length)

VRAG-RL/verl/models/qwen2/megatron/modeling_qwen2_megatron.py:244

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

VRAG-RL/verl/models/qwen2/megatron/modeling_qwen2_megatron.py:312

( self, input_ids: torch.LongTensor = None, attention_mask: Optional[torch.Tensor] = N

VRAG-RL/verl/models/qwen2/megatron/modeling_qwen2_megatron.py:388

Args: input_ids: input ids. shape (1, totol_nnz) position_ids: position ids. shape (batch_size, seq_length)

VRAG-RL/verl/models/qwen2/megatron/modeling_qwen2_megatron.py:472

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

VRAG-RL/verl/models/qwen2/megatron/modeling_qwen2_megatron.py:622

( self, *, input_ids: torch.LongTensor = None, attention_mask: Optional[torch.

VRAG-RL/verl/models/qwen2/megatron/modeling_qwen2_megatron.py:703

Args: hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)` attention_mask (

VRAG-RL/verl/models/qwen2/megatron/layers/parallel_decoder.py:47

( self, hidden_states: torch.Tensor, position_ids: Optional[torch.LongTensor] = None,

VRAG-RL/verl/models/qwen2/megatron/layers/parallel_decoder.py:115

VRAG-RL/verl/models/qwen2/megatron/layers/parallel_mlp.py:71

(self, x, seq_len=None)

VRAG-RL/verl/models/qwen2/megatron/layers/parallel_attention.py:61

( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None,

VRAG-RL/verl/models/qwen2/megatron/layers/parallel_attention.py:218

(self, hidden_states: torch.Tensor, position_ids: Optional[torch.LongTensor] =

VRAG-RL/verl/models/qwen2/megatron/layers/parallel_attention.py:324

(self, hidden_states)

VRAG-RL/verl/models/qwen2/megatron/layers/parallel_rmsnorm.py:41

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

VRAG-RL/verl/models/llama/megatron/modeling_llama_megatron.py:178

Args: input_ids: input ids. shape (1, totol_nnz) position_ids: position ids. shape (batch_size, seq_length)

VRAG-RL/verl/models/llama/megatron/modeling_llama_megatron.py:245

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

VRAG-RL/verl/models/llama/megatron/modeling_llama_megatron.py:313

( self, input_ids: torch.LongTensor = None, attention_mask: Optional[torch.Tensor] = N

VRAG-RL/verl/models/llama/megatron/modeling_llama_megatron.py:389

Args: input_ids: input ids. shape (1, totol_nnz) position_ids: position ids. shape (batch_size, seq_length)

VRAG-RL/verl/models/llama/megatron/modeling_llama_megatron.py:473

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

VRAG-RL/verl/models/llama/megatron/modeling_llama_megatron.py:569

( self, *, input_ids: torch.LongTensor = None, attention_mask: Optional[torch.

VRAG-RL/verl/models/llama/megatron/modeling_llama_megatron.py:651

Args: hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)` attention_mask (

VRAG-RL/verl/models/llama/megatron/layers/parallel_decoder.py:47

( self, hidden_states: torch.Tensor, position_ids: Optional[torch.LongTensor] = None,

VRAG-RL/verl/models/llama/megatron/layers/parallel_decoder.py:115

VRAG-RL/verl/models/llama/megatron/layers/parallel_mlp.py:71

(self, x, seq_len=None)

VRAG-RL/verl/models/llama/megatron/layers/parallel_attention.py:61

( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None,

VRAG-RL/verl/models/llama/megatron/layers/parallel_attention.py:234

(self, hidden_states: torch.Tensor, position_ids: Optional[torch.LongTensor] =

VRAG-RL/verl/models/llama/megatron/layers/parallel_attention.py:340

(self, hidden_states)

VRAG-RL/verl/models/llama/megatron/layers/parallel_rmsnorm.py:41

( self, query: Dict[str, Tensor] = None, doc: Dict[str, Tensor] = None, temper

search_engine/models/GVE/models.py:157

( self, query: Dict[str, Tensor] = None, pos: Dict[str, Tensor] = None, # renamed fro

search_engine/models/GVE/models.py:211

(self, reps_q, reps_d, temperature=1.0, use_all_pairs=True, only_self_neg=False, check_false_neg=True)

search_engine/models/GVE/model_utils.py:17

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

search_engine/models/GVE/qwen25vl/qwen25vl_model.py:55

(self, input_ids: torch.LongTensor = None, attention_mask: Optional[torch.Tens

search_engine/models/Qwen3_VL_Embedding/qwen3_vl_embedding.py:85

Methodforward_step

(batch_iter, model)

VRAG-RL/verl/workers/critic/megatron_critic.py:171

Methodforward_step

(batch_iter, model)

VRAG-RL/verl/workers/actor/megatron_actor.py:318

Methodforward_step

(batch_iter, model)

VRAG-RL/verl/workers/reward_model/megatron/reward_model.py:217

Methodgenerate_sequences

(self, prompts: DataProto)

VRAG-RL/verl/workers/megatron_workers.py:376

Methodgenerate_sequences

(self, prompts: DataProto)

VRAG-RL/verl/workers/fsdp_workers.py:485

Methodgenerate_sequences

(self, prompts: DataProto)

VRAG-RL/verl/workers/rollout/hf_rollout.py:42

Methodgenerate_sequences

(self, prompts: DataProto, **kwargs)

VRAG-RL/verl/workers/rollout/sglang_rollout/sglang_rollout.py:200

Methodgenerate_sequences

Generate sequences

VRAG-RL/verl/workers/rollout/naive/naive_rollout.py:52

Methodgenerate_sequences

(self, prompts: DataProto, **kwargs)

VRAG-RL/verl/workers/rollout/vllm_rollout/fire_vllm_rollout.py:113

Methodgenerate_sequences

(self, prompts: DataProto, **kwargs)

VRAG-RL/verl/workers/rollout/vllm_rollout/vllm_rollout.py:157

Methodgenerate_sequences

(self, prompts: DataProto, **kwargs)

VRAG-RL/verl/workers/rollout/vllm_rollout/vllm_rollout_spmd.py:159

Methodget_added_vocab

Returns the added tokens in the vocabulary as a dictionary of token to index. Results might be different from the fast call because f

VRAG-RL/verl/workers/rollout/tokenizer.py:135

Methodget_cuda_visible_devices

VRAG-RL/verl/single_controller/base/worker.py:196

VRAG-RL/verl/protocol.py:595

Methodget_decoder

search_engine/models/Qwen3_VL_Embedding/qwen3_vl_embedding.py:61

Functionget_default_kwargs_for_column_parallel_linear

VRAG-RL/verl/utils/megatron/tensor_parallel.py:58

Functionget_default_kwargs_for_parallel_embedding

VRAG-RL/verl/utils/megatron/tensor_parallel.py:77

Functionget_default_kwargs_for_row_parallel_linear

VRAG-RL/verl/utils/megatron/tensor_parallel.py:72

Methodget_image_features

(self, pixel_values: torch.FloatTensor, image_grid_thw: Optional[torch.LongTensor]

search_engine/models/Qwen3_VL_Embedding/qwen3_vl_embedding.py:70

Methodget_input_embeddings

search_engine/models/Qwen3_VL_Embedding/qwen3_vl_embedding.py:52

Functionget_kl_controller

VRAG-RL/verl/trainer/ppo/core_algos.py:56

Methodget_master_addr_port

VRAG-RL/verl/single_controller/base/worker.py:193

Methodget_megatron_global_info

VRAG-RL/verl/single_controller/base/megatron/worker_group.py:50

Methodget_megatron_rank_info

(self, rank: int)

VRAG-RL/verl/single_controller/base/megatron/worker_group.py:31

Methodget_memory_buffers

VRAG-RL/verl/utils/memory_buffer.py:156

Methodget_node_ip_by_sdk

VRAG-RL/verl/single_controller/base/worker.py:41

Methodget_rank_zero_info

VRAG-RL/verl/single_controller/base/register_center/ray.py:24

Functionget_tensor_parallel_partition_dim

VRAG-RL/verl/utils/megatron/tensor_parallel.py:90

Functionget_tensor_parallel_partition_stride

VRAG-RL/verl/utils/megatron/tensor_parallel.py:95

Functionget_unpad_data

(attention_mask)

VRAG-RL/verl/utils/torch_functional.py:516

Methodget_video_features

(self, pixel_values_videos: torch.FloatTensor, video_grid_thw: Optional[torch.LongT

search_engine/models/Qwen3_VL_Embedding/qwen3_vl_embedding.py:65

Methodget_weight_buffer_meta

VRAG-RL/verl/utils/memory_buffer.py:159

Functiongreedy_partition

(seqlen_list: List[int], k_partitions: int, equal_size: bool)

VRAG-RL/verl/utils/seqlen_balancing.py:133

Functionhandler

(signum, frame)

VRAG-RL/verl/utils/reward_score/prime_math/__init__.py:44

Functioninit_fn

(x: torch.nn.Module)

VRAG-RL/verl/utils/fsdp_utils.py:32

Methodinit_megatron

(self, default_megatron_kwargs: Dict = None)

VRAG-RL/verl/single_controller/base/megatron/worker_group.py:28

Methodinit_model

VRAG-RL/verl/workers/megatron_workers.py:284

Methodinit_model

VRAG-RL/verl/workers/megatron_workers.py:595

Methodinit_model

VRAG-RL/verl/workers/megatron_workers.py:762

Methodinit_model

VRAG-RL/verl/workers/fsdp_workers.py:368

Methodinit_model

VRAG-RL/verl/workers/fsdp_workers.py:998

Methodinitialize_weight_buffer

Initialize the weight buffer. The weight buffer is obtained according to the actor. We will construct a large buffer for each dtype i

VRAG-RL/verl/utils/memory_buffer.py:184

(*args, **kwargs)

VRAG-RL/verl/single_controller/base/decorator.py:401

VRAG-RL/verl/utils/torch_dtypes.py:59

VRAG-RL/verl/workers/rollout/tokenizer.py:161

VRAG-RL/verl/utils/torch_dtypes.py:51

VRAG-RL/verl/utils/torch_dtypes.py:55

Functionis_sequence_parallel_param

VRAG-RL/verl/utils/megatron/sequence_parallel.py:25

Functionis_sglang_available

VRAG-RL/verl/utils/import_utils.py:42

Functionkl_penalty

Compute KL divergence given logprob and ref_logprob. Copied from https://github.com/huggingface/trl/blob/main/trl/trainer/ppo_trainer.py#L1104

VRAG-RL/verl/trainer/ppo/core_algos.py:351

Functionlambda_policy_fn

VRAG-RL/verl/utils/fsdp_utils.py:81

Methodlanguage_model

search_engine/models/Qwen3_VL_Embedding/qwen3_vl_embedding.py:76

Functionllama_attn_forward

Adapted from transformers 4.49.0 to support Ulysses sequence parallelism for transformers >= 4.48.0. NOTE: This function has been tested onl

VRAG-RL/verl/models/transformers/llama.py:154

← previous next →801–900 of 1,037, ranked by callers