Upload config.json with huggingface_hub

config.json

@@ -1,21 +1,18 @@
 {
   "model_type": "oculus",
   "architectures": ["OculusForConditionalGeneration"],
-  "variant": "Unified",
   "version": "0.1",
 
   "vision_encoders": {
     "dinov3": {
-      "path": "vision_encoders/dinov3-large",
-      "model_id": "facebook/dinov3-vith16plus-pretrain-lvd1689m",
-      "hidden_size": 1280,
-      "num_layers": 32,
+      "model_id": "facebook/dinov3-vitl16-pretrain-lvd1689m",
+      "hidden_size": 1024,
+      "num_layers": 24,
       "num_heads": 16,
       "image_size": 224,
       "patch_size": 16
     },
     "siglip2": {
-      "path": "vision_encoders/siglip2-so400m",
       "model_id": "google/siglip2-so400m-patch16-naflex",
       "hidden_size": 1152,
       "num_layers": 27,
@@ -26,7 +23,6 @@
   },
 
   "language_model": {
-    "path": "language_model/lfm2.5-1.2b",
     "model_id": "LiquidAI/LFM2.5-1.2B-Base",
     "hidden_size": 1536,
     "num_layers": 16,
@@ -41,8 +37,8 @@
   },
 
   "projector_config": {
-    "fused_vision_dim": 2432,
-    "hidden_dim": 4864,
+    "fused_vision_dim": 2176,
+    "hidden_dim": 4352,
     "output_dim": 1536
   },
 
@@ -51,10 +47,8 @@
     "detection_classes": 80
   },
 
-  "instruct_enabled": true,
   "reasoning_enabled": true,
   "thinking_token": "<think>",
   "thinking_end_token": "</think>",
-  "max_thinking_tokens": 256,
-  "standalone": true
+  "max_thinking_tokens": 256
 }