pytorch/ignite

Tracking the metrics with performance gaps on XLA

Open

Aperta il 14 lug 2021

Vedi su GitHub
 (2 commenti) (0 reazioni) (0 assegnatari)Python (4313 star) (602 fork)batch import
HacktoberfestPyDataGlobalenhancementhelp wantedmodule: metrics

Descrizione

🚀 Feature

Since this commit, the metrics have been disabled with xla device due to a performance regression.

https://github.com/pytorch/ignite/blob/7bdf92322a15ba0bea71ad76f3f785b7ea5907c3/ignite/metrics/metric.py#L224-L225

To solve this, it would be interesting to track which metrics work fine and which have lack of performance. Thus, we could think about more specific enable / disable process.

Guida contributor

Tech stack
python
Dominio
performancemachine learning
Tipo issue
feature
DifficoltàQuanto dovrebbe essere impegnativa per un nuovo contributor.
4
Tempo stimatoTempo stimato per completare e verificare uno scope piccolo.
1-2 days
Stato attivitàQuanto è probabile che la issue sia ancora attiva e reviewable.
stale
ChiarezzaQuanto chiaramente la issue descrive problema, scope e risultato atteso.
needs investigation
Prerequisiti
Nessuno
Adatta ai principiantiQuanto la issue è adatta a contributor alla prima esperienza.
25
Direzione di ricerca
Investigate the referenced commit 002b595daa8a8345286c5e096c33e278948686a7 to understand why metrics were disabled on XLA. Examine the code in ignite/metrics/metric.py lines 224-225 to identify the performance regression. Then, measure which metrics (e.g., Accuracy, Loss) perform well on XLA and which cause slowdowns. The goal is to propose a more granular enabling/disabling mechanism for metrics on XLA devices.