Add vision_config, cleanup

config.json

@@ -14,17 +14,8 @@
     "mlp_ratio":4,
     "num_attention_heads":12,
     "num_hidden_layers":6,
-    "resolution_before":224,
     "stop_gradient":false,
     "tie_word_embeddings":false,
-    "vit_embed_dim":512,
-    "vit_num_hidden_layers":12,
-    "vit_layernorm_init_from_vit":false,
-    "vit_layernorm_shared":true,
-    "vit_patch_size":16,
-    "vit_remove_last":false,
-    "vit_intermediate_size":512,
-    "vit_hidden_size":768,
     "vocab_size":50265,
     "text_config_dict": null,
     "text_config":{
@@ -47,5 +38,19 @@
       "eos_token_id": 2,
       "position_embedding_type": "absolute",
       "use_cache": true
+    },
+    "vision_config_dict": null,
+    "vision_config":{
+      "architectures": ["BridgeTowerVisionModel"],
+      "embed_dim": 512,
+      "input_resolution": 224,
+      "width": 768,
+      "layers": 12,
+      "patch_size": 16,
+      "transformer_width": 512,
+      "output_resolution": 288,
+      "stop_gradient": false,
+      "vit_layernorm_shared": true,
+      "vit_remove_last": false,
     }
 }