Truncation of output?
#26
by FlareRebellion - opened
Using sglang-omni or vllm, the model seems to always cut of generation prematurely (around the 40 minute mark maybe)? Am I dong something wrong or is this a known limitation? The description claims up to 90 minutes.
FlareRebellion changed discussion title from Truncation of output to Truncation of output?
we will fix it in next version