MCPcopy Create free account

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

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

↓ 1 callersMethodupdate_sampling_params
(self, **kwargs)
verl/workers/rollout/vllm_rollout/vllm_rollout.py:126
↓ 1 callersFunctionvalidate_equation
Validate that equation only uses available numbers and each number once.
verl/utils/reward_score/countdown.py:28
↓ 1 callersMethodvalidation_step
(self, batch: TensorDict)
verl/trainer/fsdp_sft_trainer.py:288
↓ 1 callersFunctionvocab_parallel_compute_entropy_loss
Compute Categorical entropy loss Args: logits: `(torch.Tensor)` shape: (bs, response_length, vocab_size) eos_mask: `(
verl/utils/megatron/tensor_parallel.py:168
↓ 1 callersFunctionvocab_parallel_entropy
Compute entropy when the logits are sharded in tp ranks Args: vocab_parallel_logits: (total_nnz, vocab_size // tp_size) Returns:
verl/utils/megatron/tensor_parallel.py:124
↓ 1 callersMethodwarm_up_model
(self)
verl/third_party/vllm/vllm_v_0_3_1/worker.py:181
Method__call__
(self, *args: Any, **kwds: Any)
verl/utils/megatron_utils.py:222
Method__call__
(self, placement_group, placement_group_bundle_idx, use_gpu
verl/single_controller/ray/base.py:142
Method__call__
(self)
verl/single_controller/base/worker_group.py:42
Method__call__
(self)
verl/single_controller/base/worker_group.py:77
Method__call__
We will expand this function gradually based on the available datasets
verl/trainer/main_ppo.py:45
Method__call__
(self, *inputs, **kwargs)
verl/workers/sharding_manager/megatron_vllm.py:146
Method__call__
We will expand this function gradually based on the available datasets
examples/split_placement/main_ppo_split.py:39
Method__enter__
(self)
verl/workers/sharding_manager/fsdp_vllm.py:69
Method__enter__
(self)
verl/workers/sharding_manager/base.py:23
Method__enter__
(self)
verl/workers/sharding_manager/fsdp_ulysses.py:43
Method__enter__
(self)
verl/workers/sharding_manager/megatron_vllm.py:345
Method__exit__
(self, exc_type, exc_value, traceback)
verl/workers/sharding_manager/fsdp_vllm.py:93
Method__exit__
(self, exc_type, exc_value, traceback)
verl/workers/sharding_manager/base.py:26
Method__exit__
(self, exc_type, exc_value, traceback)
verl/workers/sharding_manager/fsdp_ulysses.py:51
Method__exit__
(self, exc_type, exc_value, traceback)
verl/workers/sharding_manager/megatron_vllm.py:360
Method__getitem__
(self, item)
verl/protocol.py:189
Method__getitem__
(self, item)
verl/utils/dataset/rm_dataset.py:114
Method__getitem__
(self, item)
verl/utils/dataset/sft_dataset.py:110
Method__getitem__
Note that we also return the raw_input_ids so that it can be combined with other chat template
verl/utils/dataset/rl_dataset.py:120
Method__getstate__
(self)
verl/protocol.py:194
Method__init__
(self)
tests/ray/test_data_transfer.py:39
Method__init__
(self, rank, world_size, group_name)
tests/ray/test_rvdz.py:21
Method__init__
(self, cuda_visible_devices=None)
tests/ray/test_high_level_scheduling_api.py:26
Method__init__
(self, size)
tests/ray/test_worker_group_torch.py:52
Method__init__
(self, config)
tests/ray/test_colocated_workers.py:39
Method__init__
(self)
tests/ray/test_driverfunc_to_worker.py:32
Method__init__
(self)
tests/ray/test_driverfunc_to_worker.py:38
Method__init__
(self)
tests/ray/test_ray_local_envs.py:28
Method__init__
(self, x)
tests/ray/test_worker_group_basics.py:44
Method__init__
(self)
tests/ray/check_worker_alive/main.py:29
Method__init__
(self)
tests/ray/detached_worker/server.py:51
Method__init__
Args: max_number: the maximum number allowed in the arithmetic sequence max_diff: the maximum diff. The actual commo
tests/e2e/envs/digit_completion/task.py:35
Method__init__
Character tokenizer for Hugging Face transformers. Args: characters (Sequence[str]): List of desired characters. Any character wh
tests/e2e/envs/digit_completion/tokenizer.py:31
Method__init__
(self)
verl/utils/megatron_utils.py:218
Method__init__
(self, config: PretrainedConfig)
verl/utils/flops_counter.py:61
Method__init__
(self, project_name, experiment_name, default_backend: Union[str, List[str]] = 'console', config=None)
verl/utils/tracking.py:27
Method__init__
(self)
verl/utils/seqlen_balancing.py:29
Method__init__
(self, items: List[Tuple[int, int]], k: int)
verl/utils/seqlen_balancing.py:51
Method__init__
(self, dictionary, **kwargs)
verl/utils/py_functional.py:50
Method__init__
(self, module: nn.Module)
verl/utils/memory_buffer.py:146
Method__init__
(self, transform_memory_param_fn)
verl/utils/memory_buffer.py:175
Method__init__
(self, fn)
verl/utils/model.py:30
Method__init__
(self, numel, numel_padded, dtype)
verl/utils/megatron/memory.py:20
Method__init__
(self, remote_logger=None, enable_wandb=False, print_to_console=False)
verl/utils/logger/aggregate_logger.py:32
Method__init__
(self, parquet_files: Union[str, List[str]], tokenizer, pro
verl/utils/dataset/rm_dataset.py:42
Method__init__
(self, parquet_files: Union[str, List[str]], tokenizer, pro
verl/utils/dataset/sft_dataset.py:39
Method__init__
(self, parquet_files: Union[str, List[str]], tokenizer: PreTrainedTokenizer,
verl/utils/dataset/rl_dataset.py:63
Method__init__
(self, hdfs_dir, verbose)
verl/utils/debug/trajectory_tracker.py:52
Method__init__
(self, nccl_id)
verl/utils/rendezvous/ray_backend.py:27
Method__init__
(self, process_on_nodes: List[int] = None, use_gpu: bool = True,
verl/single_controller/ray/base.py:51
Method__init__
(self, cls, *args, **kwargs)
verl/single_controller/ray/base.py:130
Method__init__
(self, resource_pool: RayResourcePool = None, ray_cls_with_init: RayClassWit
verl/single_controller/ray/base.py:178
Method__init__
(self, resource_pool: RayResourcePool, ray_cls_with_init: RayClassWithInitArgs, **kwargs)
verl/single_controller/ray/megatron.py:31
Method__init__
(self, process_on_nodes=None, max_collocate_count: int = 10, n_gpus_per_node=8)
verl/single_controller/base/worker_group.py:28
Method__init__
(self, cls, *args, **kwargs)
verl/single_controller/base/worker_group.py:66
Method__init__
(self, resource_pool: ResourcePool, **kwargs)
verl/single_controller/base/worker_group.py:93
Method__init__
(self, store)
verl/single_controller/base/worker.py:75
Method__init__
(self, cuda_visible_devices=None)
verl/single_controller/base/worker.py:119
Method__init__
(self, resource_pool: ResourcePool, **kwargs)
verl/single_controller/base/megatron/worker_group.py:23
Method__init__
(self, cuda_visible_devices=None)
verl/single_controller/base/megatron/worker.py:22
Method__init__
(self, rank_zero_info)
verl/single_controller/base/register_center/ray.py:21
Method__init__
(self, tokenizer, num_examine)
verl/trainer/main_ppo.py:41
Method__init__
(self, config, device_mesh: DeviceMesh)
verl/trainer/fsdp_sft_trainer.py:60
Method__init__
(self, config, tokenizer, role_worker_mapping: dict[Role, W
verl/trainer/ppo/ray_trainer.py:298
Method__init__
(self, init_kl_coef, target_kl, horizon)
verl/trainer/ppo/core_algos.py:34
Method__init__
(self, kl_coef)
verl/trainer/ppo/core_algos.py:49
Method__init__
(self, config: DictConfig, role: str)
verl/workers/megatron_workers.py:69
Method__init__
(self, config)
verl/workers/megatron_workers.py:581
Method__init__
(self, config: DictConfig, role: str)
verl/workers/fsdp_workers.py:53
Method__init__
(self, config)
verl/workers/fsdp_workers.py:770
Method__init__
(self, module: FSDP, inference_engine: LLM, model_config,
verl/workers/sharding_manager/fsdp_vllm.py:36
Method__init__
(self, device_mesh: DeviceMesh)
verl/workers/sharding_manager/fsdp_ulysses.py:38
Method__init__
(self, model_provider)
verl/workers/sharding_manager/megatron_vllm.py:37
Method__init__
(self, module: AllGatherPPModel, inference_engine: LLM, model_config, layer_name_mapping)
verl/workers/sharding_manager/megatron_vllm.py:240
Method__init__
(self, config)
verl/workers/critic/base.py:28
Method__init__
(self, config, model_config, megatron_config, critic_module: nn.ModuleList, critic_optimizer:
verl/workers/critic/megatron_critic.py:43
Method__init__
(self, config, critic_module: nn.Module, critic_optimizer: optim.Optimizer)
verl/workers/critic/dp_critic.py:41
Method__init__
Args: dataloader: an Iterable of TensorDict that consistently generates prompts. Note that the dataloader should han
verl/workers/rollout/base.py:25
Method__init__
(self, module: nn.Module, config)
verl/workers/rollout/hf_rollout.py:37
Method__init__
A naive rollout. It requires the module to be compatible with huggingface APIs. That is: The module should define __call__ to receive input_id
verl/workers/rollout/naive/naive_rollout.py:38
Method__init__
A vLLM rollout. It requires the module is supported by the vllm. Args: module: module here follows huggingface APIs c
verl/workers/rollout/vllm_rollout/vllm_rollout.py:59
Method__init__
MeagtronPPOActor class. This class implements the simple PPO logics when the model is built with Megatron. Args: config (OmegaCon
verl/workers/actor/megatron_actor.py:50
Method__init__
The base class for PPO actor Args: config (DictConfig): a config passed to the PPOActor. We expect the type to be
verl/workers/actor/base.py:28
Method__init__
When optimizer is None, it is Reference Policy
verl/workers/actor/dp_actor.py:41
Method__init__
(self, config)
verl/workers/reward_model/base.py:25
Method__init__
(self, config, model_config, reward_model_module: torch.nn.
verl/workers/reward_model/megatron/reward_model.py:39
Method__init__
(self, config: LlamaConfig, megatron_config: ModelParallelConfig)
verl/models/llama/megatron/modeling_llama_megatron.py:157
Method__init__
(self, config: LlamaConfig, megatron_config: ModelParallelConfig)
verl/models/llama/megatron/modeling_llama_megatron.py:223
Method__init__
(self, config: LlamaConfig, megatron_config: ModelParallelConfig)
verl/models/llama/megatron/modeling_llama_megatron.py:281
Method__init__
(self, config: LlamaConfig, megatron_config: ModelParallelConfig, pre_process, post_process)
verl/models/llama/megatron/modeling_llama_megatron.py:410
Method__init__
(self, config: LlamaConfig, megatron_config: ModelParallelConfig, pre_process, post_process)
verl/models/llama/megatron/modeling_llama_megatron.py:516
Method__init__
(self, config: LlamaConfig, megatron_config: ModelParallelConfig)
verl/models/llama/megatron/layers/parallel_decoder.py:101
Method__init__
(self, config, megatron_config: ModelParallelConfig = None)
verl/models/llama/megatron/layers/parallel_mlp.py:33
Method__init__
(self, dim, max_position_embeddings=2048, base=10000, device=None, scaling_factor=1.0)
verl/models/llama/megatron/layers/parallel_attention.py:75
← previousnext →601–700 of 1,293, ranked by callers