monkeypatch.llama_expand_mask
-monkeypatch.llama_expand_mask
expands the binary attention mask per 3.2.2 of https://arxiv.org/pdf/2107.02027.pdf
- - -diff --git a/.nojekyll b/.nojekyll index 009d60b87..feab296d3 100644 --- a/.nojekyll +++ b/.nojekyll @@ -1 +1 @@ -f9f423f8 \ No newline at end of file +5526dc52 \ No newline at end of file diff --git a/FAQS.html b/FAQS.html index 9f44d69c3..56f951e85 100644 --- a/FAQS.html +++ b/FAQS.html @@ -2,13 +2,16 @@
- +monkeypatch.llama_expand_mask
expands the binary attention mask per 3.2.2 of https://arxiv.org/pdf/2107.02027.pdf
- - -monkeypatch.llama_patch_multipack
Patched LlamaAttention to use torch.nn.functional.scaled_dot_product_attention
- - -