Commit History

Autor SHA1 Mensaxe Data
  AlpinDale 709628a74d fix hai 5 meses
  AlpinDale b0e420e711 some debug statements hai 5 meses
  AlpinDale 31b8636e89 don't do regular top-k/top-p sampling if kernels are enabled hai 5 meses
  AlpinDale 815736fc54 feat: add cuda kernels for sampling hai 5 meses
  AlpinDale dd18c5042c move prepare_inputs to the GPU (#596) hai 5 meses
  AlpinDale be8154a8a0 feat: proper embeddings API with e5-mistral-7b support hai 6 meses
  AlpinDale d287afd917 optimize get_logprobs hai 6 meses
  AlpinDale 79901b76de logprobs for target model (spec decoding) hai 6 meses
  AlpinDale 35ae01d7ba refactor: attention metadata term hai 6 meses
  AlpinDale ccf4d5cab6 disable banned tokens hai 6 meses
  AlpinDale 9ce319b03c fix: sampler indexing issues in distributed environments (#546) hai 6 meses
  AlpinDale 772b4a4504 temporarily disable dynatemp hai 6 meses
  AlpinDale b178bc12b3 fix min_tokens when eos_token_id is None hai 6 meses
  AlpinDale 7d3194e7f4 revert #244 hai 6 meses
  AlpinDale aed64884c6 feat: prompt logprobs with chunked prefill (#539) hai 6 meses
  AlpinDale fca911ee0a vLLM Upstream Sync (#526) hai 7 meses
  AlpinDale 9d81716bfd [v0.5.3] Release Candidate (#388) hai 9 meses
  50h100a f663d3fccc Merge pull request #397 from 50h100a/pr_samplerasserts hai 10 meses
  50h100a 85ae23ac3c Missed .items() and assert hai 10 meses
  50h100a 43c9858854 Merge pull request #244 from PygmalionAI/faster_topk hai 10 meses
  50h100a bd564148e2 Merge branch 'main' of https://github.com/PygmalionAI/aphrodite-engine into ffs hai 10 meses
  50h100a d3dd170a7d merge main hai 10 meses
  AlpinDale 78d66f16d1 Chunked Prefill Part 1 (#384) hai 10 meses
  AlpinDale 9181fa0396 feat: Triton kernels for sampling (#383) hai 10 meses
  50h100a dc09dc2b4d Merge branch 'main' into pr_samplers hai 10 meses
  AlpinDale f8dfac6372 chore: attention refactor and upstream sync apr01 (#365) hai 10 meses
  50h100a 7ed57e318d Overhauled SamplingTensors construction. hai 10 meses
  50h100a d5dbd29db4 hoist sampler internals into a single function. hai 10 meses
  AlpinDale da223153c6 feat&fix: cohere support and missing GPU blocks (#333) hai 11 meses
  AlpinDale e42a78381a feat: switch from pylint to ruff (#322) hai 11 meses