“Google” adds a new feature to “Jimni” to reduce the cost of using artificial intelligence models

Google launches an advantage in applying artificial intelligence programming, to reduce the cost of external developers use of the latest artificial intelligence models offered by the American company.

Google calls the new feature “Extensive Timi Storage”, and indicates that it provides 75% of the cost of “repeated context” that is passed to the artificial intelligence models via the Jimni Applications interface. This feature supports the version of the artificial intelligence model Jimni 2.5 Pro and Jimni 2.5 Flash from Google.

Tech Crash, which specializes in technology topics, stated that this is likely that this is a good news for developers, with the continued high cost of using artificial intelligence models.

It is noteworthy that the temporary storage is a wide range of practices in the field of artificial intelligence, as it re -using the data used in abundance or pre -calculated from models to reduce computing and cost requirements when performing each task.

For example, the cache can keep the answers of the questions that users usually ask on the form, which enriches the form for re -searching for answers every time the request is raised itself.

Google had previously made a temporary storage of requests submitted to artificial intelligence models, but it was an explicit temporary storage, which means that developers should determine their most frequent claims. While it was supposed to ensure costs, the explicit storage of demands usually requires a great manual effort.

Some developers were not satisfied with the performance of the Google app for explicit interim storage in the Jimlin 2.5 Pro version, which they said could cause the cost of use unexpectedly. The complaints reached their climax last week, prompting the Jimni team to apologize and pledge changes.

Unlike explicit temporary storage, implicit temporary storage is automatic.

It is activated by default in Jimlini 2.5 models and transmits cost savings if a repeated request in the Jimni applications interface arrives to an artificial intelligence model in a temporary storage memory.