26 Révisions (4e0d85205a62fc52d44a7fd065f422f1329a6561)

Auteur SHA1 Message Date
  orangels 70ced01112 feat: record response debug cases in reports il y a 1 mois
  orangels a15a62264a fix: set bench max tokens default to 64k il y a 1 mois
  orangels 739524628c feat: report effective benchmark request params il y a 1 mois
  orangels 986dc4e088 fix: restore bench max tokens default il y a 1 mois
  orangels 1a98836b94 fix: set bench max tokens default to 200k il y a 1 mois
  orangels 948e32d64a feat: log request errors for debug il y a 1 mois
  orangels e0aca08f36 feat: align gpqa scoring with simple evals il y a 1 mois
  orangels 898d02f7ef feat: record upstream error debug payloads il y a 1 mois
  orangels 18cc7ff716 feat: add provider thinking parameters il y a 1 mois
  orangels ee13739c5e feat: add raw response debug mode il y a 1 mois
  orangels 60cc009c83 fix: support bare stream flag and cover SSE parsing il y a 1 mois
  orangels f21ab8bd4c feat: add --max-tokens CLI param and correct_samples to bench reports il y a 1 mois
  orangels 4988a86654 feat: add progress bars to benchmark and RPM execution il y a 1 mois
  orangels 527b1e8fc7 feat: add params field to RPM and benchmark reports il y a 1 mois
  orangels fdb437a732 fix: correct streaming bugs and RPM diagnose gaps il y a 1 mois
  orangels 20217b2250 feat: add streaming (SSE) support with TTFT metrics il y a 1 mois
  orangels 2b9fae7ebd feat: report rpm limiter mode details il y a 1 mois
  orangels 0a96e11e98 feat: run rpm limiter modes il y a 1 mois
  orangels 65ce102004 fix: resolve only selected provider env refs il y a 1 mois
  orangels 31c41cb8ac fix: satisfy final clippy checks il y a 1 mois
  orangels 83edc789df fix: make benchmark and rpm execution reliable il y a 1 mois
  orangels 5d1163d635 feat: run benchmark and rpm tests il y a 1 mois
  orangels 1e0926cfba feat: add openai and anthropic request adapters il y a 1 mois
  orangels 2064c69924 fix: tighten benchmark dataset helpers il y a 1 mois
  orangels 67a4ea8af4 feat: add benchmark dataset fetch and prompts il y a 1 mois
  orangels c0ebdea077 chore: initialize cli project skeleton il y a 1 mois