Llama API

Eval

eval

Methods

Evaluate Rows ->
EvaluateResponse
post/alpha/eval/evaluate-rows
Run Eval ->
Job
post/alpha/eval/run-eval

Domain types

EvaluateResponse = { generations, scores }
Job = { job_id }
Eval

Jobs

eval.jobs

Methods

Cancel ->
post/alpha/eval/job/cancel
Retrieve ->
EvaluateResponse
get/alpha/eval/job/result
Status -> "completed" | "in_progress" | "failed" | 1 more...
get/alpha/eval/job/status