MCPcopy Create free account

hub / github.com/Unakar/Logic-RL / functions

Functions858 in github.com/Unakar/Logic-RL

⨍ Functions 858◇ Types & classes 113↳ Endpoints 30
Methodtest
(self)
tests/ray/test_rvdz.py:31
Functiontest_all_gather_torch
In this test, we instantiate 4 GPUs in a group and test the all_gather
tests/ray/test_worker_group_torch.py:69
Functiontest_all_gather_torch_v2
In this test, we instantiate 4 GPUs in a group and test the all_gather
tests/ray/test_worker_group_torch.py:93
Functiontest_chunk_concat
()
tests/utility/test_tensor_dict_utilities.py:106
Functiontest_colocated_workers
()
tests/ray/test_colocated_workers.py:49
Functiontest_data_transfer
()
tests/ray/test_data_transfer.py:52
Functiontest_dataproto_fold_unfold
()
tests/utility/test_tensor_dict_utilities.py:206
Functiontest_dataproto_pad_unpad
()
tests/utility/test_tensor_dict_utilities.py:168
Functiontest_flash_attn_cross_entropy
()
tests/gpu_utility/test_ops.py:16
Functiontest_import
()
tests/sanity/test_import.py:16
Functiontest_len
()
tests/utility/test_tensor_dict_utilities.py:245
Functiontest_log_probs_from_logits_response_rmpad
()
tests/gpu_utility/test_torch_functional.py:20
Functiontest_lr_scheduler
()
tests/gpu_utility/test_torch_functional.py:52
Functiontest_pop
()
tests/utility/test_tensor_dict_utilities.py:130
Functiontest_reorder
()
tests/utility/test_tensor_dict_utilities.py:95
Functiontest_repeat
()
tests/utility/test_tensor_dict_utilities.py:143
Functiontest_rl_dataset
()
tests/verl/utils/dataset/test_rl_dataset.py:29
Functiontest_rm_dataset
()
tests/verl/utils/dataset/test_rm_dataset.py:30
Functiontest_rvdz
()
tests/ray/test_rvdz.py:37
Functiontest_seqlen_balancing
()
tests/utility/test_tensor_dict_utilities.py:265
Functiontest_sft_cot_dataset
()
tests/verl/utils/dataset/test_sft_dataset.py:29
Functiontest_sft_dataset
()
tests/verl/utils/dataset/test_sft_dataset.py:46
Functiontest_single_controller_import
()
tests/sanity/test_import.py:21
Functiontest_tensor_dict_constructor
()
tests/utility/test_tensor_dict_utilities.py:52
Functiontest_tensor_dict_make_iterator
()
tests/utility/test_tensor_dict_utilities.py:66
Functiontest_torch_save_data_proto
()
tests/utility/test_tensor_dict_utilities.py:229
Functiontest_union_tensor_dict
()
tests/utility/test_tensor_dict_utilities.py:26
Functiontest_vllm_with_hf
()
tests/rollout/test_vllm_hf_loader.py:72
Methodthis_rank_models
(self)
verl/workers/sharding_manager/megatron_vllm.py:191
Functiontokenize_and_postprocess_data
input_data is the output from tokenizer.
verl/utils/torch_functional.py:225
Methodtp_size
(self)
verl/single_controller/base/megatron/worker_group.py:36
Functiontwo_to_all_dispatch_fn
Assume the input is a list of 2. Duplicate the input interleaved and pass to each worker.
tests/ray/test_worker_group_basics.py:26
Methodupdate
(self, current_kl, n_steps)
verl/trainer/ppo/core_algos.py:39
Methodupdate_actor
(self, data: DataProto)
verl/workers/megatron_workers.py:326
Methodupdate_critic
(self, data: DataProto)
verl/workers/megatron_workers.py:559
Methodupdate_critic
Update the critic
verl/workers/critic/base.py:38
Methodupdate_critic
(self, dataloader: Iterable[DataProto])
verl/workers/critic/megatron_critic.py:204
Methodupdate_critic
(self, data: DataProto)
verl/workers/critic/dp_critic.py:146
Functionupdate_dict_with_config
(dictionary: Dict, config: DictConfig)
verl/utils/config.py:20
Functionupdate_kwargs_with_config
(dictionary: Dict, config: ModelParallelConfig)
verl/utils/megatron/tensor_parallel.py:27
Methodupdate_policy
Update the policy with an iterator of DataProto Args: dataloader (Iterable[DataProto]): an iterator over the DataProto that retur
verl/workers/actor/megatron_actor.py:329
Methodupdate_policy
(self, data: DataProto)
verl/workers/actor/dp_actor.py:203
Methodupdate_this_rank_models
(self, new_models)
verl/workers/sharding_manager/megatron_vllm.py:186
Methodvalidate_input
(keys)
verl/protocol.py:402
Functionvocab_parallel_log_probs_from_logits_response_rmpad
Similar to log_probs_from_logits_response_rmpad, but the logits_rmpad is now spliited across tensor parallel region. This will further reduce the
verl/utils/megatron/tensor_parallel.py:141
Methodvocab_size
(self)
tests/e2e/envs/digit_completion/tokenizer.py:86
Methodvocab_size
`int`: Size of the base vocabulary (without the added tokens).
verl/workers/rollout/tokenizer.py:28
Functionwait
()
verl/utils/ulysses.py:146
Methodwait_for_hdfs
(self)
verl/utils/debug/trajectory_tracker.py:63
Methodweight_buffers
(self)
verl/utils/memory_buffer.py:209
Methodworker_names
(self)
verl/single_controller/ray/base.py:281
Methodworkers
(self)
verl/single_controller/ray/base.py:362
Methodworld_size
(self)
verl/single_controller/ray/base.py:366
Methodworld_size
(self)
verl/single_controller/base/worker_group.py:39
Methodworld_size
(self)
verl/single_controller/base/worker_group.py:130
Methodworld_size
(self)
verl/single_controller/base/worker.py:171
Methodzero
Reset the buffer to zero.
verl/utils/memory_buffer.py:36
Methodzero
Reset the buffer to zero.
verl/utils/megatron/memory.py:29
← previous801–858 of 858, ranked by callers