DeepFellow DOCS

CompletionLegacyRequest

PropertyTypeRequiredDescription
modelstringYesID of the model to use.
promptstring | string[] | integer[] | integer[][]YesThe prompt(s) to generate completions for - can be string, array of strings, tokens, or token arrays.
best_ofinteger | nullNoGenerate best_of completions and return the best one. Must be greater than n if both are set.
echoboolean | nullNoEcho back the prompt in addition to the completion. Default: false.
frequency_penaltynumber | nullNoPenalize new tokens based on their existing frequency (-2.0 to 2.0).
logit_biasobject | nullNoModify likelihood of specified tokens appearing (token ID -> bias from -100 to 100).
logprobsinteger | nullNoInclude log probabilities on the most likely tokens (max 5).
max_tokensinteger | nullNoMaximum number of tokens to generate in the completion.
ninteger | nullNoHow many completions to generate for each prompt.
presence_penaltynumber | nullNoPenalize new tokens based on whether they appear in the text so far (-2.0 to 2.0).
seedinteger | nullNoSeed for deterministic sampling (best effort, not guaranteed).
stopstring | StopSequenceList | nullNoUp to 4 sequences where the API will stop generating.
streamboolean | nullNoStream back partial progress as server-sent events. Default: false.
stream_optionsserver__models__completions__StreamOptions | nullNoOptions for streaming response (only used when stream=True).
suffixstring | nullNoSuffix after completion of inserted text.
temperaturenumber | nullNoSampling temperature (0-2). Higher = more random, lower = more deterministic.
top_pnumber | nullNoNucleus sampling - consider tokens with top_p probability mass (0-1).
userstring | nullNoUnique identifier for end-user to help monitor and detect abuse.

We use cookies on our website. We use them to ensure proper functioning of the site and, if you agree, for purposes such as analytics, marketing, and targeting ads.