Patch lora forward in modulated model before applying lora to layers in QA inference

This commit is contained in:
Sarvesh 2026-06-09 17:03:33 +05:30
parent 51ef4f3d8d
commit 09b6068929

View file

@ -613,6 +613,9 @@
" ).strip()\n",
"\n",
" # === Part B: Video2LoRA Inference (zero visual tokens) ===\n",
" # Re-patch the base model layers with the LoRA forward function before binding weights\n",
" model.patch_lora_forward()\n",
"\n",
" apply_lora_to_layers(\n",
" model.base_model,\n",
" model.hypernet.layer_indices,\n",