transformers
Faster generation using AWQ + Fused modules
#27411
Merged

Commits
  • v1 fusing modules
    younesbelkada committed 2 years ago
  • add fused mlp support
    younesbelkada committed 2 years ago
  • Merge remote-tracking branch 'upstream/main' into awq-fused-modules
    younesbelkada committed 2 years ago
  • up
    younesbelkada committed 2 years ago
  • fix CI
    younesbelkada committed 2 years ago
  • block save_pretrained
    younesbelkada committed 2 years ago
  • fixup
    younesbelkada committed 2 years ago
  • Merge remote-tracking branch 'upstream/main' into awq-fused-modules
    younesbelkada committed 2 years ago
  • small fix
    younesbelkada committed 2 years ago
  • add new condition
    younesbelkada committed 2 years ago
  • Merge remote-tracking branch 'upstream/main' into awq-fused-modules
    younesbelkada committed 2 years ago
  • Merge branch 'awq-fused-modules' of https://github.com/younesbelkada/transformers into awq-fused-modules
    younesbelkada committed 2 years ago
  • add v1 docs
    younesbelkada committed 2 years ago
  • add some comments
    younesbelkada committed 2 years ago
  • Merge branch 'main' into awq-fused-modules
    younesbelkada committed 2 years ago
  • Merge remote-tracking branch 'upstream/main' into awq-fused-modules
    younesbelkada committed 2 years ago
  • style
    younesbelkada committed 2 years ago
  • fix nit
    younesbelkada committed 2 years ago
  • adapt from suggestion
    younesbelkada committed 2 years ago
  • add check
    younesbelkada committed 2 years ago
  • change arg names
    younesbelkada committed 2 years ago
  • change variables name
    younesbelkada committed 2 years ago
  • Update src/transformers/integrations/awq.py
    younesbelkada committed 2 years ago
  • style
    younesbelkada committed 2 years ago
  • split up into 3 different private methods
    younesbelkada committed 2 years ago
  • more conditions
    younesbelkada committed 2 years ago
  • more checks
    younesbelkada committed 2 years ago
  • add fused tests for custom models
    younesbelkada committed 2 years ago
  • fix
    younesbelkada committed 2 years ago
  • fix tests
    younesbelkada committed 2 years ago
  • final update docs
    younesbelkada committed 2 years ago
  • final fixes
    younesbelkada committed 2 years ago
  • fix importlib metadata
    younesbelkada committed 2 years ago
  • Merge remote-tracking branch 'upstream/main' into awq-fused-modules
    younesbelkada committed 2 years ago
  • Merge branch 'awq-fused-modules' of https://github.com/younesbelkada/transformers into awq-fused-modules
    younesbelkada committed 2 years ago
  • Update src/transformers/utils/quantization_config.py
    younesbelkada committed 2 years ago
  • change it to `do_fuse`
    younesbelkada committed 2 years ago
  • nit
    younesbelkada committed 2 years ago
  • Update src/transformers/utils/quantization_config.py
    younesbelkada committed 2 years ago
  • Update src/transformers/utils/quantization_config.py
    younesbelkada committed 2 years ago
  • Update src/transformers/utils/quantization_config.py
    younesbelkada committed 2 years ago
  • Merge branch 'awq-fused-modules' of https://github.com/younesbelkada/transformers into awq-fused-modules
    younesbelkada committed 2 years ago
  • few fixes
    younesbelkada committed 2 years ago
  • revert
    younesbelkada committed 2 years ago
  • fix test
    younesbelkada committed 2 years ago
  • fix copies
    younesbelkada committed 2 years ago
  • Merge remote-tracking branch 'upstream/main' into awq-fused-modules
    younesbelkada committed 2 years ago
  • raise error if model is not quantized
    younesbelkada committed 2 years ago
  • add test
    younesbelkada committed 2 years ago
  • use quantization_config.config when fusing
    younesbelkada committed 2 years ago
  • Update src/transformers/modeling_utils.py
    younesbelkada committed 2 years ago
Loading