lighteval
Homogeneize generation params
#428
Merged

Commits
  • adding input generation config
    clefourrier committed 1 year ago
  • added tgi model
    clefourrier committed 1 year ago
  • grammar is task dependant, removed from the cofnig
    clefourrier committed 1 year ago
  • added openai config + moved everything to dict
    clefourrier committed 1 year ago
  • added generation configs to models
    clefourrier committed 1 year ago
  • added generation configs to models
    clefourrier committed 1 year ago
  • fix
    clefourrier committed 1 year ago
  • fix
    clefourrier committed 1 year ago
  • added doc
    clefourrier committed 1 year ago
  • Saved GenerationParameter class in model config classes, then saved in the models to use other attributes later
    clefourrier committed 1 year ago
  • changed model args
    clefourrier committed 1 year ago
  • test
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
  • updated launchers
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
  • rename base_model to transformers_model
    clefourrier committed 1 year ago
  • removed the use of a GenerationConfig object, as it's got lots of params set by default which slow down generations
    clefourrier committed 1 year ago
  • revert
    clefourrier committed 1 year ago
  • fix docs
    clefourrier committed 1 year ago
  • fix #16 by also allowing a generationconfig object to be passed programmatically if needed
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
  • Apply suggestions from code review
    clefourrier committed 1 year ago
  • Update src/lighteval/models/transformers/transformers_model.py
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
  • removed temperature from default vllm params as it should be passed via the generationparams instead
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
  • Update src/lighteval/models/transformers/transformers_model.py
    clefourrier committed 1 year ago
  • logging fix
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
  • added default gen params
    clefourrier committed 1 year ago
  • Apply suggestions from code review
    clefourrier committed 1 year ago
  • rename file
    clefourrier committed 1 year ago
  • added from path to openai model
    clefourrier committed 1 year ago
  • style
    clefourrier committed 1 year ago
  • Update src/lighteval/models/transformers/transformers_model.py
    clefourrier committed 1 year ago
  • inferenceendpoint renamed to ie
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
  • style 2
    clefourrier committed 1 year ago
  • fix vllm
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    NathanHB committed 1 year ago
  • restore line
    clefourrier committed 1 year ago
  • Merge branch 'main' into clem_homogeneize_generation_params
    clefourrier committed 1 year ago
Loading