Back to SimpleMetrics page.
Calculate the pseudo-perplexity from a PerResidueProbabilitiesMetric, which is a value defining the likelihood of a sequence given the prediction of a model (smaller is better). It is defined as the exponentiation of the average negative logarithm of the predicted probabilities.
Autogenerated Tag Syntax Documentation:
A metric for estimating the likeliness of a sequence given some predicted probabilities.
References and author information for the PseudoPerplexityMetric simple metric:
PseudoPerplexityMetric SimpleMetric's author(s): Moritz Ertelt, University of Leipzig [moritz.ertelt@gmail.com]
<PseudoPerplexityMetric name="(&string;)" custom_type="(&string;)"
metric="(&string;)" use_cached_data="(false &bool;)"
cache_prefix="(&string;)" cache_suffix="(&string;)"
fail_on_missing_cache="(true &bool;)" />
<ROSETTASCRIPTS>
<RESIDUE_SELECTORS>
<Chain name="res" chains="A" />
<Index name="mask" resnums="25"/>
</RESIDUE_SELECTORS>
<SIMPLE_METRICS>
<PerResidueEsmProbabilitiesMetric name="prediction" residue_selector="res" attention_mask_selection="mask" write_pssm="test.pssm" model="esm2_t6_8M_UR50D" multirun="true"/>
<PseudoPerplexityMetric name="perplex" metric="prediction"/>
</SIMPLE_METRICS>
<FILTERS>
</FILTERS>
<MOVERS>
<RunSimpleMetrics name="run" metrics="perplex"/>
</MOVERS>
<PROTOCOLS>
<Add mover_name="run"/>
</PROTOCOLS>
</ROSETTASCRIPTS>
This is currently unpublished.