maxent_grpo.training.runtime.deps

Dependency loading utilities used by the training runtime.

Functions

_allow_stubbed_deps()	Return True when lightweight dependency stubs are allowed.
_install_accelerate_stub(_hint)	Install a lightweight accelerate stub for tests.
_install_torch_stub(hint)	Install a lightweight torch stub so imports can succeed in tests.
_maybe_create_deepspeed_plugin()	Construct a DeepSpeedPlugin from Accelerate env/config when available.
get_trl_prepare_deepspeed()	Return TRL's prepare_deepspeed helper when available.
maybe_create_deepspeed_plugin()	Public wrapper for DeepSpeed plugin creation.
require_accelerator(context)	Return accelerate.Accelerator or raise a helpful RuntimeError.
require_dataloader(context)	Return torch.utils.data.DataLoader with a descriptive error on failure.
require_deepspeed(context[, module])	Return a DeepSpeed module import or raise a contextual RuntimeError.
require_torch(context)	Return the torch module or raise a helpful RuntimeError.
require_transformer_base_classes(context)	Return (PreTrainedModel, PreTrainedTokenizer) with clear failure messages.

maxent_grpo.training.runtime.deps.maybe_create_deepspeed_plugin()

Public wrapper for DeepSpeed plugin creation.

Return type:

Any | None

maxent_grpo.training.runtime.deps.get_trl_prepare_deepspeed()

Return TRL’s prepare_deepspeed helper when available.

Return type:

Any | None

maxent_grpo.training.runtime.deps.require_accelerator(context)

Return accelerate.Accelerator or raise a helpful RuntimeError.

Parameters:

context (str)

Return type:

Any

maxent_grpo.training.runtime.deps.require_dataloader(context)

Return torch.utils.data.DataLoader with a descriptive error on failure.

Parameters:

context (str)

Return type:

Any

maxent_grpo.training.runtime.deps.require_deepspeed(context, module='deepspeed')

Return a DeepSpeed module import or raise a contextual RuntimeError.

Parameters:

• context (str)

• module (str)

Return type:

ModuleType

maxent_grpo.training.runtime.deps.require_torch(context)

Return the torch module or raise a helpful RuntimeError.

Parameters:

context (str)

Return type:

ModuleType

maxent_grpo.training.runtime.deps.require_transformer_base_classes(context)

Return (PreTrainedModel, PreTrainedTokenizer) with clear failure messages.

Parameters:

context (str)

Return type:

Tuple[Any, Any]