tiny per-element Laguna test fixture

config.json

@@ -0,0 +1,62 @@
+{
+  "architectures": [
+    "LagunaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": null,
+  "dtype": "float32",
+  "eos_token_id": null,
+  "gating": "per-element",
+  "head_dim": 16,
+  "hidden_act": "silu",
+  "hidden_size": 64,
+  "initializer_range": 0.02,
+  "intermediate_size": 128,
+  "layer_types": [
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 128,
+  "mlp_layer_types": [
+    "dense",
+    "sparse"
+  ],
+  "model_type": "laguna",
+  "moe_apply_router_weight_on_input": false,
+  "moe_intermediate_size": 32,
+  "moe_routed_scaling_factor": 1.0,
+  "moe_router_logit_softcapping": 0.0,
+  "num_attention_heads": 8,
+  "num_attention_heads_per_layer": [
+    8,
+    8
+  ],
+  "num_experts": 8,
+  "num_experts_per_tok": 2,
+  "num_hidden_layers": 2,
+  "num_key_value_heads": 2,
+  "output_router_logits": false,
+  "pad_token_id": null,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "full_attention": {
+      "partial_rotary_factor": 0.5,
+      "rope_theta": 500000.0,
+      "rope_type": "default"
+    },
+    "rope_type": "default",
+    "sliding_attention": {
+      "partial_rotary_factor": 1.0,
+      "rope_theta": 10000.0,
+      "rope_type": "default"
+    }
+  },
+  "router_aux_loss_coef": 0.001,
+  "shared_expert_intermediate_size": 32,
+  "sliding_window": 8,
+  "tie_word_embeddings": false,
+  "transformers_version": "5.13.0.dev0",
+  "use_cache": true,
+  "vocab_size": 256
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "output_attentions": false,
+  "output_hidden_states": false,
+  "transformers_version": "5.13.0.dev0",
+  "use_cache": true
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1f03bda83dbc48f7fe1e3cce9d43344c6f1056e5d1e050dabd279eb03861c44
+size 689208