fit_length#
- hybrid_aif.train.fit_length(ep, seq_len, pad=None)[source]#
One episode at exactly
seq_lensteps: truncated, or padded by freezing.Padding repeats the final observation and
contwith zero action and zero reward;is_lastmarks the final stored step. Task columns pad bypad[name]("hold"or"zero"; zeros when not given).