Add Leopard-LLaVA

config.json

@@ -0,0 +1,46 @@
+{
+    "architectures": [
+        "LlavaForConditionalGeneration"
+    ],
+    "ignore_index": -100,
+    "image_token_index": 128200,
+    "model_type": "llava",
+    "projector_hidden_act": "gelu",
+    "text_config": {
+        "architectures": [
+            "LlavaLlamaForCausalLM"
+        ],
+        "bos_token_id": 128000,
+        "eos_token_id": 128001,
+        "intermediate_size": 14336,
+        "max_position_embeddings": 16384,
+        "max_sequence_length": 16384,
+        "model_type": "llama",
+        "num_key_value_heads": 8,
+        "rms_norm_eps": 1e-05,
+        "rope_scaling": {
+            "factor": 8.0,
+            "high_freq_factor": 4.0,
+            "low_freq_factor": 1.0,
+            "original_max_position_embeddings": 8192,
+            "rope_type": "llama3"
+        },
+        "rope_theta": 500000,
+        "torch_dtype": "float16",
+        "vocab_size": 128256
+    },
+    "transformers_version": "4.44.2",
+    "vision_config": {
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1152,
+        "image_size": 364,
+        "intermediate_size": 4304,
+        "layer_norm_eps": 1e-06,
+        "model_type": "siglip_vision_model",
+        "num_attention_heads": 16,
+        "num_hidden_layers": 27,
+        "patch_size": 14
+    },
+    "vision_feature_layer": -2,
+    "vision_feature_select_strategy": "default"
+}