Training in progress, step 1

config.json

@@ -0,0 +1,31 @@
+{
+  "act_format": "linear",
+  "act_on": false,
+  "act_type": "associative",
+  "architectures": [
+    "InnerLoopARMTForCausalLM"
+  ],
+  "attend_to_previous_input": false,
+  "base_model_config": null,
+  "base_model_name": "meta-llama/Llama-3.2-1B",
+  "constant_depth": false,
+  "correction": true,
+  "d_mem": 64,
+  "dtype": "bfloat16",
+  "freeze_mem": false,
+  "gating": false,
+  "layers_attr": "model.layers",
+  "max_hop": 4,
+  "model_type": "armt",
+  "n_heads": 1,
+  "noisy_halting": false,
+  "num_mem_tokens": 32,
+  "segment_alignment": "left",
+  "segment_size": 1024,
+  "sliding_window": true,
+  "time_penalty": 0.0,
+  "transformers_version": "4.57.1",
+  "use_denom": true,
+  "use_sink": true,
+  "wrap_pos": false
+}

model.safetensors.index.json

@@ -0,0 +1,236 @@
+{
+  "metadata": {
+    "total_size": 5228871744
+  },
+  "weight_map": {
+    "memory": "pytorch_model-00001-of-00002.safetensors",
+    "model.lm_head.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.embed_tokens.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.0.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.0.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.0.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.0.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.1.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.1.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.1.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.1.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.10.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.10.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.10.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.10.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.11.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.11.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.11.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.11.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.12.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.12.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.12.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.12.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.13.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.13.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.13.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.13.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.14.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.14.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.14.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.14.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.15.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.W_mb.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.W_mk.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.W_mq.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.W_mv.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.layer.mlp.down_proj.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.layer.mlp.gate_proj.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.layer.mlp.up_proj.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.15.layer.self_attn.o_proj.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.15.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.15.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.2.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.2.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.2.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.2.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.3.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.3.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.3.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.3.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.4.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.4.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.4.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.4.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.5.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.5.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.5.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.5.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.6.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.6.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.6.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.6.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.7.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.7.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.7.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.7.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.8.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.8.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.8.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.8.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.W_mb.bias": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.9.W_mb.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.W_mk.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.W_mq.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.W_mv.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.9.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "model.model.layers.9.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.layers.9.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.safetensors",
+    "model.model.norm.weight": "pytorch_model-00002-of-00002.safetensors",
+    "sink": "pytorch_model-00001-of-00002.safetensors"
+  }
+}

pytorch_model.bin.index.json

@@ -0,0 +1,236 @@
+{
+  "metadata": {
+    "total_size": 5228871744
+  },
+  "weight_map": {
+    "memory": "pytorch_model-00001-of-00002.bin",
+    "model.lm_head.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.embed_tokens.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.0.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.0.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.0.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.0.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.1.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.1.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.1.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.1.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.10.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.10.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.10.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.10.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.11.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.11.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.11.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.11.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.12.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.12.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.12.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.12.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.13.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.13.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.13.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.13.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.14.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.14.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.14.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.14.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.15.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.W_mb.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.W_mk.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.W_mq.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.W_mv.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.layer.mlp.down_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.layer.mlp.gate_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.layer.mlp.up_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.15.layer.self_attn.o_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.15.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.15.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.2.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.2.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.2.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.2.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.3.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.3.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.3.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.3.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.4.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.4.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.4.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.4.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.5.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.5.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.5.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.5.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.6.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.6.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.6.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.6.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.7.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.7.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.7.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.7.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.8.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.8.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.8.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.8.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.W_mb.bias": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.9.W_mb.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.W_mk.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.W_mq.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.W_mv.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.layer.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.9.layer.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.layer.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.layer.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.layer.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
+    "model.model.layers.9.layer.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.layer.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.layer.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.layers.9.layer.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "model.model.norm.weight": "pytorch_model-00002-of-00002.bin",
+    "sink": "pytorch_model-00001-of-00002.bin"
+  }
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d7c1cf719c3cbfcaf2869369bc5a368ff2ac939ba88a474e66e4b78a0bd9c08
+size 6904