Update config.json (#3)

- Update config.json (dcb406270317fd493324903e3a5b303f135b64ac)
- Update config.json (9b0ea887479d408f0eec6ef2ae1ff179afa6ff58)
- Update config.json (9d135aa7ce1c475399351bad15ca9eb8ec99ff82)


Co-authored-by: Younes Belkada <ybelkada@users.noreply.huggingface.co>

config.json

@@ -28,23 +28,27 @@
   "num_hidden_layers": 2,
   "pad_token_id": 0,
   "qk_layer_norms": false,
-  "qk_layer_norms_perceiver": false,
-  "resampler_depth": 2,
-  "resampler_head_dim": 8,
-  "resampler_n_heads": 2,
-  "resampler_n_latents": 16,
   "rms_norm_eps": 1e-06,
   "tie_word_embeddings": false,
   "torch_dtype": "float16",
   "transformers_version": "4.27.0.dev0",
   "use_cache": true,
   "use_resampler": true,
-  "vision_embed_dim": 32,
-  "vision_image_size": 30,
-  "vision_intermediate_size": 37,
-  "vision_patch_size": 2,
-  "vision_num_attention_heads": 4,
-  "vision_num_hidden_layers": 5,
   "vocab_size": 32000,
-  "word_embed_proj_dim": 16
+  "word_embed_proj_dim": 16,
+  "vision_config": {
+    "embed_dim": 32,
+    "image_size": 30,
+    "intermediate_size": 37,
+    "patch_size": 2,
+    "num_attention_heads": 4,
+    "num_hidden_layers": 5
+  },
+  "perceiver_config": {
+    "qk_layer_norms_perceiver": false,
+    "resampler_depth": 2,
+    "resampler_head_dim": 8,
+    "resampler_n_heads": 2,
+    "resampler_n_latents": 16
+  }
 }