trinity.algorithm.policy_loss_fn.gspo_policy_loss module

GSPO-token policy loss function.

Implemented from https://arxiv.org/pdf/2507.18071

class trinity.algorithm.policy_loss_fn.gspo_policy_loss.GSPOLossFn(backend: str = 'verl', clip_range: float | None = None, clip_range_low: float | None = None, clip_range_high: float | None = None, loss_agg_mode: str | None = 'seq-mean-token-mean')

基类：PolicyLossFn

__init__(backend: str = 'verl', clip_range: float | None = None, clip_range_low: float | None = None, clip_range_high: float | None = None, loss_agg_mode: str | None = 'seq-mean-token-mean') → None

Initialize the policy loss function.

参数:

backend -- The training framework/backend to use (e.g., "verl")

classmethod default_args() → Dict

Get default initialization arguments for this loss function.

返回:

The default init arguments for the policy loss function.

返回类型:

Dict

property select_keys

Returns parameter keys mapped to the specific training framework's naming convention.