Text translation method, device and equipment based on large language model

The invention provides a text translation method, device and equipment based on a large language model, and can be applied to the technical field of artificial intelligence and big data. The method comprises the steps that based on a to-be-translated text, a target sample is determined from a sample set containing multiple samples, the multiple samples are determined in advance based on multiple historical translation samples, and the historical translation samples comprise historical source translation texts and historical target translation texts; the historical target translation texts are obtained after the historical source translation texts are translated, and the historical source translation texts and the historical target translation texts are in one-to-one correspondence; based on the target sample, generating a prompt text for translation by using a preset template; and inputting the to-be-translated text and the prompt text into the large language model, and outputting a target translation result.