Update config.json

config.json

@@ -1,7 +1,7 @@
 {
-  "_name_or_path": "./llava-language-model",
+  "_name_or_path": "qnguyen3/nanoLLaVA-1.5",
   "architectures": [
-    "LlavaQwen2Model"
+    "BunnyQwenForCausalLM"
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 151645,
@@ -12,16 +12,55 @@
   "image_aspect_ratio": "pad",
   "initializer_range": 0.02,
   "intermediate_size": 2816,
-  "language_model": "vilm/Quyen-SE-v0.1",
   "max_position_embeddings": 32768,
   "max_window_layers": 21,
   "mm_hidden_size": 1152,
   "mm_projector_lr": null,
   "mm_projector_type": "mlp2x_gelu",
   "mm_vision_tower": "google/siglip-so400m-patch14-384",
-  "model_type": "llava-qwen2",
+  "language_model": "vilm/Quyen-SE-v0.1",
+  "model_type": "llava",
+  "text_config": {
+    "_name_or_path": "vilm/Quyen-SE-v0.1",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "attention_dropout": 0.0,
+    "eos_token_id": 151645,
+    "hidden_act": "silu",
+    "hidden_size": 1024,
+    "initializer_range": 0.02,
+    "intermediate_size": 2816,
+    "max_position_embeddings": 32768,
+    "max_window_layers": 21,
+    "model_type": "qwen2",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 24,
+	  "num_image_tokens": 729,
+    "num_key_value_heads": 16,
+    "rms_norm_eps": 1e-06,
+    "rope_theta": 1000000.0,
+    "sliding_window": 4096,
+    "tie_word_embeddings": false,
+    "torch_dtype": "bfloat16",
+    "transformers_version": "4.38.1",
+    "use_cache": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "vision_config": {
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+	  "num_image_tokens": 729,
+    "patch_size": 14
+  },
   "num_attention_heads": 16,
   "num_hidden_layers": 24,
+  "num_image_tokens": 729,
   "num_key_value_heads": 16,
   "rms_norm_eps": 1e-06,
   "rope_theta": 1000000.0,
@@ -29,10 +68,12 @@
   "tie_word_embeddings": false,
   "tokenizer_model_max_length": 4096,
   "tokenizer_padding_side": "right",
-  "transformers_version": "4.42.3",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.39.2",
   "tune_mm_mlp_adapter": false,
-  "use_cache": true,
+  "use_cache": false,
   "use_mm_proj": true,
   "use_sliding_window": false,
-  "vocab_size": 151936
-}
+  "vocab_size": 151936,
+  "image_token_index": -200
+}