Remove torch.compile from model loaders entirely

Empty backend string still triggered inductor. Removing compile_args
connection from WanVideoModelLoader nodes (131, 132) ensures no
torch.compile is applied to the transformer blocks.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

workflows/Wan22-I2V-Remix-API.json

@@ -85,10 +85,6 @@
       "load_device": "offload_device",
       "attention_mode": "sageattn",
       "rms_norm_function": "default",
-      "compile_args": [
-        "127",
-        0
-      ],
       "block_swap_args": [
         "128",
         0
@@ -107,10 +103,6 @@
       "load_device": "offload_device",
       "attention_mode": "sageattn",
       "rms_norm_function": "default",
-      "compile_args": [
-        "127",
-        0
-      ],
       "block_swap_args": [
         "128",
         0