(Trained with Unsloth)

config.json

@@ -22,9 +22,9 @@
         "final_logit_softcapping": null,
         "head_dim": 256,
         "hidden_activation": "gelu_pytorch_tanh",
-        "hidden_size": 2560,
+        "hidden_size": 3840,
         "initializer_range": 0.02,
-        "intermediate_size": 10240,
+        "intermediate_size": 15360,
         "layer_types": [
             "sliding_attention",
             "sliding_attention",
@@ -59,13 +59,27 @@
             "sliding_attention",
             "sliding_attention",
             "sliding_attention",
-            "sliding_attention"
+            "sliding_attention",
+            "sliding_attention",
+            "full_attention",
+            "sliding_attention",
+            "sliding_attention",
+            "sliding_attention",
+            "sliding_attention",
+            "sliding_attention",
+            "full_attention",
+            "sliding_attention",
+            "sliding_attention",
+            "sliding_attention",
+            "sliding_attention",
+            "sliding_attention",
+            "full_attention"
         ],
         "max_position_embeddings": 131072,
         "model_type": "gemma3_text",
-        "num_attention_heads": 8,
-        "num_hidden_layers": 34,
-        "num_key_value_heads": 4,
+        "num_attention_heads": 16,
+        "num_hidden_layers": 48,
+        "num_key_value_heads": 8,
         "query_pre_attn_scalar": 256,
         "rms_norm_eps": 1e-06,
         "rope_local_base_freq": 10000.0,