Upload 9 files

config.json

@@ -87,7 +87,7 @@
       "sliding_attention",
       "full_attention"
     ],
-    "max_position_embeddings": 256,
+    "max_position_embeddings": 2048,
     "model_type": "gemma3n_text",
     "num_attention_heads": 8,
     "num_hidden_layers": 2,
@@ -102,7 +102,7 @@
     "vocab_size": 262400,
     "vocab_size_per_layer_input": 262144
   },
-  "transformers_version": "4.57.1",
+  "transformers_version": "4.57.6",
   "vision_config": {
     "architecture": "mobilenetv3_small_050",
     "do_pooling": false,

generation_config.json

@@ -6,5 +6,5 @@
     106
   ],
   "pad_token_id": 0,
-  "transformers_version": "4.57.1"
+  "transformers_version": "4.57.6"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3c1b4484f7d3a0217d800bb537d4fac4e27e4eff1ad448e2124a6709c422b3db
+oid sha256:b7d85c328ad44ff4ee16f50de3a67287ce610f5ad9e0423d59f2ce19c53c39ed
 size 73145268

preprocessor_config.json

@@ -22,7 +22,7 @@
     0.5
   ],
   "image_processor_type": "SiglipImageProcessorFast",
-  "image_seq_length": 4,
+  "image_seq_length": 256,
   "image_std": [
     0.5,
     0.5,
@@ -47,7 +47,7 @@
   "return_tensors": null,
   "sampling_rate": 16000,
   "size": {
-    "height": 64,
-    "width": 64
+    "height": 768,
+    "width": 768
   }
 }