Upload config.json with huggingface_hub

config.json

@@ -6,31 +6,33 @@
 
   "vision_encoders": {
     "dinov3": {
-      "path": "vision_encoders/dinov2-large",
-      "model_id": "facebook/dinov3-large",
-      "hidden_size": 1024,
-      "num_layers": 24,
-      "num_heads": 16
+      "path": "vision_encoders/dinov3-large",
+      "model_id": "facebook/dinov3-vith16plus-pretrain-lvd1689m",
+      "hidden_size": 1280,
+      "num_layers": 32,
+      "num_heads": 16,
+      "image_size": 224,
+      "patch_size": 16
     },
-    "siglip": {
-      "path": "vision_encoders/siglip-base",
-      "model_id": "google/siglip-base-patch16-224",
-      "hidden_size": 768,
-      "num_layers": 12,
-      "num_heads": 12
+    "siglip2": {
+      "path": "vision_encoders/siglip2-so400m",
+      "model_id": "google/siglip2-so400m-patch16-naflex",
+      "hidden_size": 1152,
+      "num_layers": 27,
+      "num_heads": 16,
+      "image_size": 384,
+      "patch_size": 16
     }
   },
 
   "language_model": {
-    "captioning": {
-      "path": "language_model/blip-captioning",
-      "model_id": "Salesforce/blip-image-captioning-base"
-    },
-    "vqa": {
-      "path": "language_model/blip-vqa-finetuned",
-      "base_model_id": "Salesforce/blip-vqa-base",
-      "finetuned": true
-    }
+    "path": "language_model/lfm2.5-1.2b",
+    "model_id": "LiquidAI/LFM2.5-1.2B-Base",
+    "hidden_size": 1536,
+    "num_layers": 16,
+    "num_heads": 24,
+    "vocab_size": 131072,
+    "max_position_embeddings": 32768
   },
 
   "trained_components": {
@@ -39,15 +41,14 @@
   },
 
   "projector_config": {
-    "fused_vision_dim": 1792,
-    "hidden_dim": 2048,
-    "num_tokens": 64,
-    "output_dim": 768
+    "fused_vision_dim": 2432,
+    "hidden_dim": 4864,
+    "output_dim": 1536
   },
 
   "task_heads": {
-    "detection_classes": 80,
-    "segmentation_classes": 150
+    "segmentation_classes": 150,
+    "detection_classes": 80
   },
 
   "instruct_enabled": true,