transformers
🚨 Exporters: decompose generative models into one decode graph
#50252
Open

🚨 Exporters: decompose generative models into one decode graph #50252

IlyasMoutawwakil
IlyasMoutawwakil Cache: a cross-attention layer type, and shaped empty tensors on lazy…
858e915b
IlyasMoutawwakil Generation and masking: export-safe causal skip and padding, a non-sl…
53207635
IlyasMoutawwakil Keep VibeVoice's own static-cache override; no attribute-name paramet…
2f4ff6e1
IlyasMoutawwakil T5Gemma2: drop its cross-attention cache override for the base one
661f914b
IlyasMoutawwakil Merge remote-tracking branch 'origin/cache-cross-attention-layer' int…
d8b5652a
IlyasMoutawwakil Merge remote-tracking branch 'origin/generation-masking-export' into …
d1dbc5d0
IlyasMoutawwakil Exporters: take the generation decomposition's cache, masking and vis…
3c7b654d
IlyasMoutawwakil Exporters: decompose generative models into one decode graph
733203f0
github-actions
IlyasMoutawwakil Exporters: format the ExecuTorch runner
7cfeaefb
HuggingFaceDocBuilderDev
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
39cf5d5b
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
ac8749f9
IlyasMoutawwakil Exporters: keep a greedy generation config saveable after completing …
dc359c68
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
4b8a9387
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
c387d668
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
427f7408
IlyasMoutawwakil Exporters: let the generation decode graph take deepstack features
289d20a5
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
d1332663
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
3d9cbc60
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
831e5c03
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
810d75d3
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
b62e4908
IlyasMoutawwakil Exporters: decode a token at a time for linear attention that switche…
78d52c58
IlyasMoutawwakil Exporters: size a captured static cache for the generation its config…
26d9a4cf
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
93c1f7f9
IlyasMoutawwakil Merge branch 'exporters-multimodal-runtime' into exporters-generation…
2b8d4295

Login to write a write a comment.

Login via GitHub

Reviewers
No reviews
Assignees
No one assigned
Labels
Milestone