Support the siglip 2 naflex model as a clip vision model. (#11831)

Not useful yet.
2026-01-12 14:05:54 -08:00
parent a3b5d4996a
commit c881a1d689
3 changed files with 91 additions and 10 deletions
--- a/comfy/clip_vision_siglip2_base_naflex.json
+++ b/comfy/clip_vision_siglip2_base_naflex.json
@@ -0,0 +1,14 @@
+{
+  "num_channels": 3,
+  "hidden_act": "gelu_pytorch_tanh",
+  "hidden_size": 1152,
+  "image_size": -1,
+  "intermediate_size": 4304,
+  "model_type": "siglip2_vision_model",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 27,
+  "patch_size": 16,
+  "num_patches": 256,
+  "image_mean": [0.5, 0.5, 0.5],
+  "image_std": [0.5, 0.5, 0.5]
+}