transformers
Add Zamba2
#34517
Merged
Overview
Commits
90
Changes
View On GitHub
Commits
  • First commit
    pglorio committed 1 year ago
  • Finish model implementation
    pglorio committed 1 year ago
  • First commit
    pglorio committed 1 year ago
  • Finish model implementation
    pglorio committed 1 year ago
  • Merge branch 'zamba2' of https://github.com/Zyphra/transformers_zamba into zamba2
    pglorio committed 1 year ago
  • Register zamba2
    pglorio committed 1 year ago
  • generated modeling and configuration
    pglorio committed 1 year ago
  • Merge pull request #2 from Zyphra/main
    pglorio committed 1 year ago
  • generated modeling and configuration
    pglorio committed 1 year ago
  • added hybrid cache
    pglorio committed 1 year ago
  • fix attention_mask in mamba
    pglorio committed 1 year ago
  • dropped unused loras
    pglorio committed 1 year ago
  • fix flash2
    pglorio committed 1 year ago
  • Merge pull request #3 from Zyphra/main
    pglorio committed 1 year ago
  • config docstrings
    root committed 1 year ago
  • fix config and fwd pass
    pglorio committed 1 year ago
  • make fixup fixes
    pglorio committed 1 year ago
  • text_modeling_zamba2
    pglorio committed 1 year ago
  • Merge pull request #4 from Zyphra/main
    pglorio committed 1 year ago
  • small fixes
    pglorio committed 1 year ago
  • make fixup fixes
    pglorio committed 1 year ago
  • Merge pull request #5 from Zyphra/main
    pglorio committed 1 year ago
  • Fix modular model converter
    pglorio committed 1 year ago
  • added inheritances in modular, renamed zamba cache
    pglorio committed 1 year ago
  • Merge pull request #6 from Zyphra/main
    pglorio committed 1 year ago
  • modular rebase
    pglorio committed 1 year ago
  • Rebase
    pglorio committed 1 year ago
  • new modular conversion
    pglorio committed 1 year ago
  • fix generated modeling file
    pglorio committed 1 year ago
  • fixed import for Zamba2RMSNormGated
    pglorio committed 1 year ago
  • modular file cleanup
    pglorio committed 1 year ago
  • rebase
    pglorio committed 1 year ago
  • make fixup and model tests
    pglorio committed 1 year ago
  • dropped inheritance for Zamba2PreTrainedModel
    pglorio committed 1 year ago
  • make fixup and unit tests
    pglorio committed 1 year ago
  • Add inheritance of rope from GemmaRotaryEmbedding
    pglorio committed 1 year ago
  • moved rope to model init
    pglorio committed 1 year ago
  • drop del self.self_attn and del self.feed_forward
    pglorio committed 1 year ago
  • Rebase onto upstream
    pglorio committed 1 year ago
  • fix tests
    pglorio committed 1 year ago
  • renamed lora -> adapter
    pglorio committed 1 year ago
  • rewrote adapter implementation
    pglorio committed 1 year ago
  • rebase
    pglorio committed 1 year ago
  • fixed tests
    pglorio committed 1 year ago
  • Merge branch 'main' into zamba2
    pglorio committed 1 year ago
  • Fix torch_forward in mamba2 layer
    pglorio committed 1 year ago
  • Fix torch_forward in mamba2 layer
    pglorio committed 1 year ago
  • Fix torch_forward in mamba2 layer
    pglorio committed 1 year ago
  • Dropped adapter in-place sum
    pglorio committed 1 year ago
  • removed rope from attention init
    pglorio committed 1 year ago
  • updated rope
    pglorio committed 1 year ago
  • created get_layers method
    pglorio committed 1 year ago
  • rebase
    pglorio committed 1 year ago
  • make fixup fix
    pglorio committed 1 year ago
  • make fixup fixes
    pglorio committed 1 year ago
  • make fixup fixes
    pglorio committed 1 year ago
  • fix merge conflicts
    pglorio committed 1 year ago
  • update to new attention standard
    pglorio committed 1 year ago
  • fixes for merge
    pglorio committed 1 year ago
  • update to new attention standard
    pglorio committed 1 year ago
  • make fixup fixes
    pglorio committed 1 year ago
  • rebase
    pglorio committed 1 year ago
  • minor fixes
    pglorio committed 1 year ago
  • cache_position
    pglorio committed 1 year ago
  • removed cache_position postion_ids use_cache
    pglorio committed 1 year ago
  • remove config from modular
    pglorio committed 1 year ago
  • removed config from modular (2)
    pglorio committed 1 year ago
  • rebase
    pglorio committed 1 year ago
  • import apply_rotary_pos_emb from llama
    pglorio committed 1 year ago
  • fixed rope_kwargs
    pglorio committed 1 year ago
  • Instantiate cache in Zamba2Model
    pglorio committed 1 year ago
  • fix cache
    pglorio committed 1 year ago
  • fix @slow decorator
    pglorio committed 1 year ago
  • rebase
    pglorio committed 1 year ago
  • rebase
    pglorio committed 1 year ago
  • small fix in modular file
    pglorio committed 1 year ago
  • Update docs/source/en/model_doc/zamba2.md
    pglorio committed 1 year ago
  • several minor fixes
    pglorio committed 1 year ago
  • inherit mamba2decoder fwd and drop position_ids in mamba
    pglorio committed 1 year ago
  • removed docstrings from modular
    pglorio committed 1 year ago
  • rebase
    pglorio committed 1 year ago
  • reinstate zamba2 attention decoder fwd
    pglorio committed 1 year ago
  • use regex for tied keys
    pglorio committed 1 year ago
  • Revert "use regex for tied keys"
    pglorio committed 1 year ago
  • use regex for tied keys
    pglorio committed 1 year ago
  • add cpu to slow forward tests
    pglorio committed 1 year ago
  • dropped config.use_shared_mlp_adapter
    pglorio committed 1 year ago
  • Update docs/source/en/model_doc/zamba2.md
    pglorio committed 1 year ago
  • rebase
    pglorio committed 1 year ago
  • re-convert from modular
    pglorio committed 1 year ago
FAQ Terms Privacy Refunds Impressum
Loading