text-generation-inference
TensorRT-LLM backend bump to latest version + misc fixes
#2791
Merged
Overview
Commits
30
Changes
View On GitHub
Commits
  • misc(cmake) update dependencies
    mfuntowicz committed 1 year ago
  • feat(hardware) enable new hardware.hpp and unittests
    mfuntowicz committed 1 year ago
  • test(ctest) enable address sanitizer
    mfuntowicz committed 1 year ago
  • feat(backend): initial rewrite of the backend for simplicity
    mfuntowicz committed 1 year ago
  • feat(backend): remove all the logs from hardware.hpp
    mfuntowicz committed 1 year ago
  • feat(backend): added some logging
    mfuntowicz committed 1 year ago
  • feat(backend): enable compiler warning if support for RVO not applying
    mfuntowicz committed 1 year ago
  • feat(backend): missing return statement
    mfuntowicz committed 1 year ago
  • feat(backend): introduce backend_workspace_t to store precomputed information from the engine folder
    mfuntowicz committed 1 year ago
  • feat(backend): delete previous backend impl
    mfuntowicz committed 1 year ago
  • feat(backend): more impl
    mfuntowicz committed 1 year ago
  • feat(backend): use latest trtllm main version to have g++ >= 13 compatibility
    mfuntowicz committed 1 year ago
  • feat(backend): allow overriding which Python to use
    mfuntowicz committed 1 year ago
  • feat(backend): fix backend_exception_t -> backend_error_t naming
    mfuntowicz committed 1 year ago
  • feat(backend): impl missing generation_step_t as return value of pull_tokens
    mfuntowicz committed 1 year ago
  • feat(backend): make backend_workspace_t::engines_folder constexpr
    mfuntowicz committed 1 year ago
  • feat(backend): fix main.rs retrieving the tokenizer
    mfuntowicz committed 1 year ago
  • feat(backend): add guard to multiple header definitions
    mfuntowicz committed 1 year ago
  • test(backend): add more unittest
    mfuntowicz committed 1 year ago
  • feat(backend): remove constexpr from par
    mfuntowicz committed 1 year ago
  • feat(backend): remove constexpig
    mfuntowicz committed 1 year ago
  • test(backend): more test coverage
    mfuntowicz committed 1 year ago
  • chore(trtllm): update dependency towards 0.15.0
    mfuntowicz committed 1 year ago
  • effectively cancel the request on the executor
    mfuntowicz committed 1 year ago
  • feat(backend) fix moving backend when pulling
    mfuntowicz committed 1 year ago
  • feat(backend): make sure we can easily cancel request on the executor
    mfuntowicz committed 1 year ago
  • feat(backend): fix missing "0" field access
    mfuntowicz committed 1 year ago
  • misc(backend): fix reborrowing Pin<&mut T> as described in the doc https://doc.rust-lang.org/stable/std/pin/struct.Pin.html#method.as_mut
    mfuntowicz committed 1 year ago
  • chore: Add doc and CI for TRTLLM (#2799)
    Hugoch committed 1 year ago
  • misc(backend): indent
    mfuntowicz committed 1 year ago
FAQ Terms Privacy Refunds Impressum
Loading