MCPcopy Create free account

hub / github.com/Jiayi-Pan/TinyZero / functions

Functions1,293 in github.com/Jiayi-Pan/TinyZero

Methodget_tokenizer
(self)
verl/third_party/vllm/vllm_v_0_3_1/llm.py:139
Methodget_tokenizer
(self)
verl/third_party/vllm/vllm_v_0_6_3/llm.py:151
Methodget_tokenizer
(self)
verl/third_party/vllm/vllm_v_0_5_4/llm.py:157
Methodget_tokenizer
(self)
verl/third_party/vllm/vllm_v_0_4_2/llm.py:143
Methodget_tokenizer_for_seq
(sequence: Sequence)
verl/third_party/vllm/vllm_v_0_6_3/llm_engine_sp.py:200
Functionget_unpad_data
(attention_mask)
verl/utils/torch_functional.py:483
Methodget_vocab
(self)
tests/e2e/envs/digit_completion/tokenizer.py:89
Methodget_weight_buffer_meta
(self)
verl/utils/memory_buffer.py:156
Functiongpt2_dtensor_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_6_3/dtensor_weight_loaders.py:308
Functiongpt2_dtensor_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_5_4/dtensor_weight_loaders.py:270
Functiongpt2_dtensor_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_4_2/dtensor_weight_loaders.py:201
Functiongpt2_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_3_1/weight_loaders.py:40
Functiongpt2_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_6_3/megatron_weight_loaders.py:46
Functiongpt2_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_5_4/megatron_weight_loaders.py:45
Functiongpt2_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_4_2/megatron_weight_loaders.py:45
Functiongptbigcode_dtensor_load_weights
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_6_3/dtensor_weight_loaders.py:61
Functiongptbigcode_dtensor_load_weights
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_5_4/dtensor_weight_loaders.py:64
Functiongptbigcode_dtensor_load_weights
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_4_2/dtensor_weight_loaders.py:74
Functiongreedy_partition
(seqlen_list: List[int], k_partitions: int, equal_size: bool)
verl/utils/seqlen_balancing.py:133
Methodinit_cache_engine
(self, cache_config: CacheConfig)
verl/third_party/vllm/vllm_v_0_3_1/worker.py:168
Methodinit_cache_engine
(self)
verl/third_party/vllm/vllm_v_0_6_3/llm.py:145
Methodinit_cache_engine
(self)
verl/third_party/vllm/vllm_v_0_6_3/llm_engine_sp.py:344
Methodinit_cache_engine
(self)
verl/third_party/vllm/vllm_v_0_6_3/spmd_gpu_executor.py:156
Methodinit_cache_engine
(self)
verl/third_party/vllm/vllm_v_0_5_4/llm.py:151
Methodinit_cache_engine
(self)
verl/third_party/vllm/vllm_v_0_5_4/llm_engine_sp.py:270
Methodinit_cache_engine
(self)
verl/third_party/vllm/vllm_v_0_5_4/spmd_gpu_executor.py:150
Methodinit_cache_engine
(self)
verl/third_party/vllm/vllm_v_0_4_2/llm.py:137
Methodinit_cache_engine
(self)
verl/third_party/vllm/vllm_v_0_4_2/llm_engine_sp.py:236
Methodinit_cache_engine
(self)
verl/third_party/vllm/vllm_v_0_4_2/spmd_gpu_executor.py:140
Functioninit_fn
(x: torch.nn.Module)
verl/utils/fsdp_utils.py:29
Methodinit_megatron
(self, default_megatron_kwargs: Dict = None)
verl/single_controller/base/megatron/worker_group.py:28
Methodinit_model
(self)
verl/workers/megatron_workers.py:261
Methodinit_model
(self)
verl/workers/megatron_workers.py:514
Methodinit_model
(self)
verl/workers/megatron_workers.py:672
Methodinit_model
(self)
verl/workers/fsdp_workers.py:285
Methodinit_model
(self)
verl/workers/fsdp_workers.py:652
Methodinit_model
(self)
verl/workers/fsdp_workers.py:851
Methodinitialize_cache
Initialize the KV cache in all workers.
verl/third_party/vllm/vllm_v_0_6_3/spmd_gpu_executor.py:134
Methodinitialize_cache
Initialize the KV cache in all workers.
verl/third_party/vllm/vllm_v_0_5_4/spmd_gpu_executor.py:127
Methodinitialize_cache
Initialize the KV cache in all workers.
verl/third_party/vllm/vllm_v_0_4_2/spmd_gpu_executor.py:117
Functioninitialize_parallel_state
( distributed_init_method: str = "env://", backend: str = "nccl", tensor_model_parallel_size: int
verl/third_party/vllm/vllm_v_0_6_3/parallel_state.py:38
Functioninitialize_parallel_state
( distributed_init_method: str = "env://", backend: str = "nccl", tensor_model_parallel_size: int
verl/third_party/vllm/vllm_v_0_5_4/parallel_state.py:37
Functioninitialize_parallel_state
( distributed_init_method: str = "env://", backend: str = "nccl", tensor_model_parallel_size: int
verl/third_party/vllm/vllm_v_0_4_2/parallel_state.py:35
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
verl/utils/memory_buffer.py:181
Functioninner
(*args, **kwargs)
verl/single_controller/base/decorator.py:401
Methodis_bf16
(precision)
verl/utils/torch_dtypes.py:59
Methodis_fast
(self)
verl/workers/rollout/tokenizer.py:161
Methodis_fp16
(precision)
verl/utils/torch_dtypes.py:51
Methodis_fp32
(precision)
verl/utils/torch_dtypes.py:55
Functionis_sequence_parallel_param
(param)
verl/utils/megatron/sequence_parallel.py:25
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
verl/trainer/ppo/core_algos.py:242
Methodlist_loras
(self)
verl/third_party/vllm/vllm_v_0_3_1/llm_engine_sp.py:724
Methodlist_loras
(self)
verl/third_party/vllm/vllm_v_0_3_1/worker.py:266
Methodlist_loras
(self)
verl/third_party/vllm/vllm_v_0_6_3/spmd_gpu_executor.py:178
Methodlist_loras
(self)
verl/third_party/vllm/vllm_v_0_5_4/spmd_gpu_executor.py:172
Methodlist_loras
(self)
verl/third_party/vllm/vllm_v_0_4_2/spmd_gpu_executor.py:162
Methodlist_prompt_adapters
(self)
verl/third_party/vllm/vllm_v_0_6_3/spmd_gpu_executor.py:193
Methodlist_prompt_adapters
(self)
verl/third_party/vllm/vllm_v_0_5_4/spmd_gpu_executor.py:188
Functionllama_dtensor_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_6_3/dtensor_weight_loaders.py:107
Functionllama_dtensor_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_5_4/dtensor_weight_loaders.py:110
Functionllama_dtensor_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_4_2/dtensor_weight_loaders.py:120
Functionllama_flash_attn_forward
adapt from transformers 4.47.1
verl/models/transformers/llama.py:27
Functionllama_megatron_core_te_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_6_3/megatron_weight_loaders.py:86
Functionllama_megatron_core_te_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_5_4/megatron_weight_loaders.py:85
Functionllama_megatron_core_te_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_4_2/megatron_weight_loaders.py:85
Functionllama_megatron_core_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_6_3/megatron_weight_loaders.py:117
Functionllama_megatron_core_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_5_4/megatron_weight_loaders.py:116
Functionllama_megatron_core_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_4_2/megatron_weight_loaders.py:116
Functionllama_megatron_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_6_3/megatron_weight_loaders.py:74
Functionllama_megatron_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_5_4/megatron_weight_loaders.py:73
Functionllama_megatron_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_4_2/megatron_weight_loaders.py:73
Functionllama_weight_loader
(actor_weights: Dict, vllm_model: nn.Module)
verl/third_party/vllm/vllm_v_0_3_1/weight_loaders.py:68
Methodload_checkpoint
(self, checkpoint_path)
verl/workers/megatron_workers.py:395
Methodload_checkpoint
(self, checkpoint_path)
verl/workers/megatron_workers.py:568
Functionload_fsdp_grad
(module, device_id)
verl/utils/fsdp_utils.py:86
Methodload_model
(self)
verl/third_party/vllm/vllm_v_0_3_1/model_runner.py:90
Methodload_model
( self, *, model_config: ModelConfig, device_config: DeviceConfig, lor
verl/third_party/vllm/vllm_v_0_6_3/model_loader.py:116
Methodload_model
( self, actor_model: Union[PreTrainedModel, Dict], model_config: ModelConfig,
verl/third_party/vllm/vllm_v_0_6_3/model_loader.py:156
Methodload_model
( self, actor_model: Union[PreTrainedModel, Dict], model_config: ModelConfig,
verl/third_party/vllm/vllm_v_0_6_3/model_loader.py:259
Methodload_model
(self)
verl/third_party/vllm/vllm_v_0_6_3/model_runner.py:101
Methodload_model
(self, *, model_config: ModelConfig, device_config: DeviceConfig, lora_config: Optional[LoRAConfig],
verl/third_party/vllm/vllm_v_0_5_4/model_loader.py:112
Methodload_model
(self, actor_model: Union[PreTrainedModel, Dict], model_config: ModelConfig, device_config:
verl/third_party/vllm/vllm_v_0_5_4/model_loader.py:143
Methodload_model
(self, actor_model: Union[PreTrainedModel, Dict], model_config: ModelConfig, device_config:
verl/third_party/vllm/vllm_v_0_5_4/model_loader.py:230
Methodload_model
(self)
verl/third_party/vllm/vllm_v_0_5_4/model_runner.py:89
Methodload_model
(self, *, model_config: ModelConfig, device_config: DeviceConfig, lora_config: Optional[LoRAConfig],
verl/third_party/vllm/vllm_v_0_4_2/model_loader.py:103
Methodload_model
(self, actor_model: Union[PreTrainedModel, Dict], model_config: Mo
verl/third_party/vllm/vllm_v_0_4_2/model_loader.py:133
Methodload_model
(self, actor_model: Union[PreTrainedModel, Dict], model_config: Mo
verl/third_party/vllm/vllm_v_0_4_2/model_loader.py:218
Methodload_model
(self)
verl/third_party/vllm/vllm_v_0_4_2/model_runner.py:105
Methodload_pretrained_model
(self, checkpoint_path)
verl/workers/megatron_workers.py:399
Functionload_state_dict_to_megatron_llama
Load merged state_dict to sharded Megatron module in training.
verl/models/llama/megatron/checkpoint_utils/llama_loader.py:51
Methodlocal_rank_list
(self)
verl/single_controller/base/worker_group.py:55
Methodlocal_world_size_list
(self)
verl/single_controller/base/worker_group.py:49
Methodlog
(self, data, step)
verl/utils/tracking.py:67
Functionlogitsprocessor_init
Args: scale: A scaling factor to apply to the logits.
verl/third_party/vllm/vllm_v_0_6_3/model_loader.py:313
Functionlogitsprocessor_init
Args: scale: A scaling factor to apply to the logits.
verl/third_party/vllm/vllm_v_0_5_4/model_loader.py:279
Functionlogprobs_of_labels_v2
A memory efficient implementation of logprobs_from_logits
verl/utils/torch_functional.py:76
Methodloss_func
(output, data, meta_info)
verl/workers/critic/megatron_critic.py:135
Methodloss_func
(output, data, meta_info)
verl/workers/actor/megatron_actor.py:249
Methodloss_func
(output)
verl/workers/reward_model/megatron/reward_model.py:219
Functionlr_lambda
(current_step)
verl/utils/torch_functional.py:412
← previousnext →1,001–1,100 of 1,293, ranked by callers