Upload folder using huggingface_hub

config.json

@@ -0,0 +1,33 @@
+{
+  "_name_or_path": "openbuddy-falcon-40b-v16.1-4k",
+  "alibi": false,
+  "apply_residual_connection_post_layernorm": false,
+  "architectures": [
+    "FalconForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bias": false,
+  "bos_token_id": 11,
+  "eos_token_id": 11,
+  "hidden_dropout": 0.0,
+  "hidden_size": 8192,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "max_position_embeddings": 2048,
+  "model_type": "falcon",
+  "multi_query": true,
+  "new_decoder_architecture": true,
+  "num_attention_heads": 128,
+  "num_hidden_layers": 60,
+  "num_kv_heads": 8,
+  "parallel_attn": true,
+  "rope_scaling": {
+    "factor": 2.0,
+    "type": "linear"
+  },
+  "rope_theta": 10000.0,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.37.0.dev0",
+  "use_cache": true,
+  "vocab_size": 70656
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 11,
+  "eos_token_id": 11,
+  "transformers_version": "4.37.0.dev0"
+}

model-00001-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21f326f4031b6a17af4a8dd3528918c2245583277489404c632e397becd70bea
+size 4697753840

model-00002-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d20cbbacdb820a0e6a3de41e6725a8b280668cdf346db2ac05393d93c85d7937
+size 4899212696

model-00003-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac88f0d00ef8d05423a52ae40b86d86e2a73946f8d9dd34e45a6089f0412052a
+size 4613933832

model-00004-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:585f82c9c37050dd6be8c68573a60a370b8a86707baf11b37a51cef37b070a20
+size 4899212728

model-00005-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20eb944b7942590cad579cea75ea36bdddc5941c9a74b941a1c48623f84d2c8b
+size 4613933856

model-00006-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0717886afef02f865e6cd820b3a81426af28a553188cca8999f0ea7e47a23dc
+size 4899212728

model-00007-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d2fcd20f706eb01e92cfa45c4f44d37e2f919fd53dc330bf0c58fe90545a4df
+size 4613933856

model-00008-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fece013334f8e9c1395934710a328fc7bf5700428c00ebe2275c9226eda12e3a
+size 4899212728

model-00009-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da5c1b1605e31f0548f5bf687d45c6dd67d1c097330e6761eae220024beaad8d
+size 4613933856

model-00010-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15798442ce0129125d4240e9881e4c03797d3253e8026faa5215539a549f862b
+size 4899212728

model-00011-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d32dd32efbcac544c130ee9989aa63be980a6269dde27b62936a072df9032d2
+size 4613933856

model-00012-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64831a715698222013c6c4b200eae5ac339e3e2bb59fcec70b359a9cf73afb39
+size 4899212728

model-00013-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a7c64b80264d68a9c9e84edb7cd919bf21597e391c4c295c3362cf8c128b0af
+size 4613933856

model-00014-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46743317a5fdc3e882d73ee92e08bdff373d0fa1433df2212651e0457a74e2f4
+size 4899212728

model-00015-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49792c27f6acabe3eb828dfcbea3beda291ec436f7dc8978e158424fbd77bac9
+size 4613933856

model-00016-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c0fb8429efeeabedd78aa61e299e91b7cd571fbf35d762886b7441370f5e652e
+size 4899212728

model-00017-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae0b15ccb54efd6398fcd5cb8ce27f6ebc0af5e02c910e388ea6151d037eda41
+size 4613933856

model-00018-of-00018.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbca326ba673682270c3d7669211c825040c673f711e4d9b4064532ef81b7de5
+size 1895990872

model.safetensors.index.json

@@ -0,0 +1,490 @@
+{
+  "metadata": {
+    "total_size": 82698862592
+  },
+  "weight_map": {
+    "transformer.h.0.ln_attn.bias": "model-00001-of-00018.safetensors",
+    "transformer.h.0.ln_attn.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.0.ln_mlp.bias": "model-00001-of-00018.safetensors",
+    "transformer.h.0.ln_mlp.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.0.mlp.dense_4h_to_h.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.0.mlp.dense_h_to_4h.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.0.self_attention.dense.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.0.self_attention.query_key_value.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.1.ln_attn.bias": "model-00001-of-00018.safetensors",
+    "transformer.h.1.ln_attn.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.1.ln_mlp.bias": "model-00001-of-00018.safetensors",
+    "transformer.h.1.ln_mlp.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.1.mlp.dense_4h_to_h.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.1.mlp.dense_h_to_4h.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.1.self_attention.dense.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.1.self_attention.query_key_value.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.10.ln_attn.bias": "model-00004-of-00018.safetensors",
+    "transformer.h.10.ln_attn.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.10.ln_mlp.bias": "model-00004-of-00018.safetensors",
+    "transformer.h.10.ln_mlp.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.10.mlp.dense_4h_to_h.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.10.mlp.dense_h_to_4h.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.10.self_attention.dense.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.10.self_attention.query_key_value.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.11.ln_attn.bias": "model-00004-of-00018.safetensors",
+    "transformer.h.11.ln_attn.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.11.ln_mlp.bias": "model-00004-of-00018.safetensors",
+    "transformer.h.11.ln_mlp.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.11.mlp.dense_4h_to_h.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.11.mlp.dense_h_to_4h.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.11.self_attention.dense.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.11.self_attention.query_key_value.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.12.ln_attn.bias": "model-00004-of-00018.safetensors",
+    "transformer.h.12.ln_attn.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.12.ln_mlp.bias": "model-00004-of-00018.safetensors",
+    "transformer.h.12.ln_mlp.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.12.mlp.dense_4h_to_h.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.12.mlp.dense_h_to_4h.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.12.self_attention.dense.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.12.self_attention.query_key_value.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.13.ln_attn.bias": "model-00005-of-00018.safetensors",
+    "transformer.h.13.ln_attn.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.13.ln_mlp.bias": "model-00005-of-00018.safetensors",
+    "transformer.h.13.ln_mlp.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.13.mlp.dense_4h_to_h.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.13.mlp.dense_h_to_4h.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.13.self_attention.dense.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.13.self_attention.query_key_value.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.14.ln_attn.bias": "model-00005-of-00018.safetensors",
+    "transformer.h.14.ln_attn.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.14.ln_mlp.bias": "model-00005-of-00018.safetensors",
+    "transformer.h.14.ln_mlp.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.14.mlp.dense_4h_to_h.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.14.mlp.dense_h_to_4h.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.14.self_attention.dense.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.14.self_attention.query_key_value.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.15.ln_attn.bias": "model-00005-of-00018.safetensors",
+    "transformer.h.15.ln_attn.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.15.ln_mlp.bias": "model-00005-of-00018.safetensors",
+    "transformer.h.15.ln_mlp.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.15.mlp.dense_4h_to_h.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.15.mlp.dense_h_to_4h.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.15.self_attention.dense.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.15.self_attention.query_key_value.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.16.ln_attn.bias": "model-00006-of-00018.safetensors",
+    "transformer.h.16.ln_attn.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.16.ln_mlp.bias": "model-00006-of-00018.safetensors",
+    "transformer.h.16.ln_mlp.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.16.mlp.dense_4h_to_h.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.16.mlp.dense_h_to_4h.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.16.self_attention.dense.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.16.self_attention.query_key_value.weight": "model-00005-of-00018.safetensors",
+    "transformer.h.17.ln_attn.bias": "model-00006-of-00018.safetensors",
+    "transformer.h.17.ln_attn.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.17.ln_mlp.bias": "model-00006-of-00018.safetensors",
+    "transformer.h.17.ln_mlp.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.17.mlp.dense_4h_to_h.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.17.mlp.dense_h_to_4h.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.17.self_attention.dense.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.17.self_attention.query_key_value.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.18.ln_attn.bias": "model-00006-of-00018.safetensors",
+    "transformer.h.18.ln_attn.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.18.ln_mlp.bias": "model-00006-of-00018.safetensors",
+    "transformer.h.18.ln_mlp.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.18.mlp.dense_4h_to_h.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.18.mlp.dense_h_to_4h.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.18.self_attention.dense.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.18.self_attention.query_key_value.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.19.ln_attn.bias": "model-00006-of-00018.safetensors",
+    "transformer.h.19.ln_attn.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.19.ln_mlp.bias": "model-00006-of-00018.safetensors",
+    "transformer.h.19.ln_mlp.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.19.mlp.dense_4h_to_h.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.19.mlp.dense_h_to_4h.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.19.self_attention.dense.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.19.self_attention.query_key_value.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.2.ln_attn.bias": "model-00002-of-00018.safetensors",
+    "transformer.h.2.ln_attn.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.2.ln_mlp.bias": "model-00002-of-00018.safetensors",
+    "transformer.h.2.ln_mlp.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.2.mlp.dense_4h_to_h.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.2.mlp.dense_h_to_4h.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.2.self_attention.dense.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.2.self_attention.query_key_value.weight": "model-00001-of-00018.safetensors",
+    "transformer.h.20.ln_attn.bias": "model-00007-of-00018.safetensors",
+    "transformer.h.20.ln_attn.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.20.ln_mlp.bias": "model-00007-of-00018.safetensors",
+    "transformer.h.20.ln_mlp.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.20.mlp.dense_4h_to_h.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.20.mlp.dense_h_to_4h.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.20.self_attention.dense.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.20.self_attention.query_key_value.weight": "model-00006-of-00018.safetensors",
+    "transformer.h.21.ln_attn.bias": "model-00007-of-00018.safetensors",
+    "transformer.h.21.ln_attn.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.21.ln_mlp.bias": "model-00007-of-00018.safetensors",
+    "transformer.h.21.ln_mlp.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.21.mlp.dense_4h_to_h.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.21.mlp.dense_h_to_4h.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.21.self_attention.dense.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.21.self_attention.query_key_value.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.22.ln_attn.bias": "model-00007-of-00018.safetensors",
+    "transformer.h.22.ln_attn.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.22.ln_mlp.bias": "model-00007-of-00018.safetensors",
+    "transformer.h.22.ln_mlp.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.22.mlp.dense_4h_to_h.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.22.mlp.dense_h_to_4h.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.22.self_attention.dense.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.22.self_attention.query_key_value.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.23.ln_attn.bias": "model-00008-of-00018.safetensors",
+    "transformer.h.23.ln_attn.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.23.ln_mlp.bias": "model-00008-of-00018.safetensors",
+    "transformer.h.23.ln_mlp.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.23.mlp.dense_4h_to_h.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.23.mlp.dense_h_to_4h.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.23.self_attention.dense.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.23.self_attention.query_key_value.weight": "model-00007-of-00018.safetensors",
+    "transformer.h.24.ln_attn.bias": "model-00008-of-00018.safetensors",
+    "transformer.h.24.ln_attn.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.24.ln_mlp.bias": "model-00008-of-00018.safetensors",
+    "transformer.h.24.ln_mlp.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.24.mlp.dense_4h_to_h.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.24.mlp.dense_h_to_4h.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.24.self_attention.dense.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.24.self_attention.query_key_value.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.25.ln_attn.bias": "model-00008-of-00018.safetensors",
+    "transformer.h.25.ln_attn.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.25.ln_mlp.bias": "model-00008-of-00018.safetensors",
+    "transformer.h.25.ln_mlp.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.25.mlp.dense_4h_to_h.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.25.mlp.dense_h_to_4h.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.25.self_attention.dense.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.25.self_attention.query_key_value.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.26.ln_attn.bias": "model-00008-of-00018.safetensors",
+    "transformer.h.26.ln_attn.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.26.ln_mlp.bias": "model-00008-of-00018.safetensors",
+    "transformer.h.26.ln_mlp.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.26.mlp.dense_4h_to_h.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.26.mlp.dense_h_to_4h.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.26.self_attention.dense.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.26.self_attention.query_key_value.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.27.ln_attn.bias": "model-00009-of-00018.safetensors",
+    "transformer.h.27.ln_attn.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.27.ln_mlp.bias": "model-00009-of-00018.safetensors",
+    "transformer.h.27.ln_mlp.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.27.mlp.dense_4h_to_h.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.27.mlp.dense_h_to_4h.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.27.self_attention.dense.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.27.self_attention.query_key_value.weight": "model-00008-of-00018.safetensors",
+    "transformer.h.28.ln_attn.bias": "model-00009-of-00018.safetensors",
+    "transformer.h.28.ln_attn.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.28.ln_mlp.bias": "model-00009-of-00018.safetensors",
+    "transformer.h.28.ln_mlp.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.28.mlp.dense_4h_to_h.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.28.mlp.dense_h_to_4h.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.28.self_attention.dense.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.28.self_attention.query_key_value.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.29.ln_attn.bias": "model-00009-of-00018.safetensors",
+    "transformer.h.29.ln_attn.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.29.ln_mlp.bias": "model-00009-of-00018.safetensors",
+    "transformer.h.29.ln_mlp.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.29.mlp.dense_4h_to_h.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.29.mlp.dense_h_to_4h.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.29.self_attention.dense.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.29.self_attention.query_key_value.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.3.ln_attn.bias": "model-00002-of-00018.safetensors",
+    "transformer.h.3.ln_attn.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.3.ln_mlp.bias": "model-00002-of-00018.safetensors",
+    "transformer.h.3.ln_mlp.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.3.mlp.dense_4h_to_h.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.3.mlp.dense_h_to_4h.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.3.self_attention.dense.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.3.self_attention.query_key_value.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.30.ln_attn.bias": "model-00010-of-00018.safetensors",
+    "transformer.h.30.ln_attn.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.30.ln_mlp.bias": "model-00010-of-00018.safetensors",
+    "transformer.h.30.ln_mlp.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.30.mlp.dense_4h_to_h.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.30.mlp.dense_h_to_4h.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.30.self_attention.dense.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.30.self_attention.query_key_value.weight": "model-00009-of-00018.safetensors",
+    "transformer.h.31.ln_attn.bias": "model-00010-of-00018.safetensors",
+    "transformer.h.31.ln_attn.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.31.ln_mlp.bias": "model-00010-of-00018.safetensors",
+    "transformer.h.31.ln_mlp.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.31.mlp.dense_4h_to_h.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.31.mlp.dense_h_to_4h.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.31.self_attention.dense.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.31.self_attention.query_key_value.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.32.ln_attn.bias": "model-00010-of-00018.safetensors",
+    "transformer.h.32.ln_attn.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.32.ln_mlp.bias": "model-00010-of-00018.safetensors",
+    "transformer.h.32.ln_mlp.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.32.mlp.dense_4h_to_h.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.32.mlp.dense_h_to_4h.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.32.self_attention.dense.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.32.self_attention.query_key_value.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.33.ln_attn.bias": "model-00010-of-00018.safetensors",
+    "transformer.h.33.ln_attn.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.33.ln_mlp.bias": "model-00010-of-00018.safetensors",
+    "transformer.h.33.ln_mlp.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.33.mlp.dense_4h_to_h.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.33.mlp.dense_h_to_4h.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.33.self_attention.dense.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.33.self_attention.query_key_value.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.34.ln_attn.bias": "model-00011-of-00018.safetensors",
+    "transformer.h.34.ln_attn.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.34.ln_mlp.bias": "model-00011-of-00018.safetensors",
+    "transformer.h.34.ln_mlp.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.34.mlp.dense_4h_to_h.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.34.mlp.dense_h_to_4h.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.34.self_attention.dense.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.34.self_attention.query_key_value.weight": "model-00010-of-00018.safetensors",
+    "transformer.h.35.ln_attn.bias": "model-00011-of-00018.safetensors",
+    "transformer.h.35.ln_attn.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.35.ln_mlp.bias": "model-00011-of-00018.safetensors",
+    "transformer.h.35.ln_mlp.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.35.mlp.dense_4h_to_h.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.35.mlp.dense_h_to_4h.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.35.self_attention.dense.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.35.self_attention.query_key_value.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.36.ln_attn.bias": "model-00011-of-00018.safetensors",
+    "transformer.h.36.ln_attn.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.36.ln_mlp.bias": "model-00011-of-00018.safetensors",
+    "transformer.h.36.ln_mlp.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.36.mlp.dense_4h_to_h.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.36.mlp.dense_h_to_4h.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.36.self_attention.dense.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.36.self_attention.query_key_value.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.37.ln_attn.bias": "model-00012-of-00018.safetensors",
+    "transformer.h.37.ln_attn.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.37.ln_mlp.bias": "model-00012-of-00018.safetensors",
+    "transformer.h.37.ln_mlp.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.37.mlp.dense_4h_to_h.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.37.mlp.dense_h_to_4h.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.37.self_attention.dense.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.37.self_attention.query_key_value.weight": "model-00011-of-00018.safetensors",
+    "transformer.h.38.ln_attn.bias": "model-00012-of-00018.safetensors",
+    "transformer.h.38.ln_attn.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.38.ln_mlp.bias": "model-00012-of-00018.safetensors",
+    "transformer.h.38.ln_mlp.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.38.mlp.dense_4h_to_h.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.38.mlp.dense_h_to_4h.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.38.self_attention.dense.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.38.self_attention.query_key_value.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.39.ln_attn.bias": "model-00012-of-00018.safetensors",
+    "transformer.h.39.ln_attn.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.39.ln_mlp.bias": "model-00012-of-00018.safetensors",
+    "transformer.h.39.ln_mlp.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.39.mlp.dense_4h_to_h.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.39.mlp.dense_h_to_4h.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.39.self_attention.dense.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.39.self_attention.query_key_value.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.4.ln_attn.bias": "model-00002-of-00018.safetensors",
+    "transformer.h.4.ln_attn.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.4.ln_mlp.bias": "model-00002-of-00018.safetensors",
+    "transformer.h.4.ln_mlp.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.4.mlp.dense_4h_to_h.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.4.mlp.dense_h_to_4h.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.4.self_attention.dense.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.4.self_attention.query_key_value.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.40.ln_attn.bias": "model-00012-of-00018.safetensors",
+    "transformer.h.40.ln_attn.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.40.ln_mlp.bias": "model-00012-of-00018.safetensors",
+    "transformer.h.40.ln_mlp.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.40.mlp.dense_4h_to_h.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.40.mlp.dense_h_to_4h.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.40.self_attention.dense.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.40.self_attention.query_key_value.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.41.ln_attn.bias": "model-00013-of-00018.safetensors",
+    "transformer.h.41.ln_attn.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.41.ln_mlp.bias": "model-00013-of-00018.safetensors",
+    "transformer.h.41.ln_mlp.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.41.mlp.dense_4h_to_h.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.41.mlp.dense_h_to_4h.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.41.self_attention.dense.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.41.self_attention.query_key_value.weight": "model-00012-of-00018.safetensors",
+    "transformer.h.42.ln_attn.bias": "model-00013-of-00018.safetensors",
+    "transformer.h.42.ln_attn.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.42.ln_mlp.bias": "model-00013-of-00018.safetensors",
+    "transformer.h.42.ln_mlp.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.42.mlp.dense_4h_to_h.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.42.mlp.dense_h_to_4h.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.42.self_attention.dense.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.42.self_attention.query_key_value.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.43.ln_attn.bias": "model-00013-of-00018.safetensors",
+    "transformer.h.43.ln_attn.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.43.ln_mlp.bias": "model-00013-of-00018.safetensors",
+    "transformer.h.43.ln_mlp.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.43.mlp.dense_4h_to_h.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.43.mlp.dense_h_to_4h.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.43.self_attention.dense.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.43.self_attention.query_key_value.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.44.ln_attn.bias": "model-00014-of-00018.safetensors",
+    "transformer.h.44.ln_attn.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.44.ln_mlp.bias": "model-00014-of-00018.safetensors",
+    "transformer.h.44.ln_mlp.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.44.mlp.dense_4h_to_h.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.44.mlp.dense_h_to_4h.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.44.self_attention.dense.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.44.self_attention.query_key_value.weight": "model-00013-of-00018.safetensors",
+    "transformer.h.45.ln_attn.bias": "model-00014-of-00018.safetensors",
+    "transformer.h.45.ln_attn.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.45.ln_mlp.bias": "model-00014-of-00018.safetensors",
+    "transformer.h.45.ln_mlp.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.45.mlp.dense_4h_to_h.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.45.mlp.dense_h_to_4h.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.45.self_attention.dense.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.45.self_attention.query_key_value.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.46.ln_attn.bias": "model-00014-of-00018.safetensors",
+    "transformer.h.46.ln_attn.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.46.ln_mlp.bias": "model-00014-of-00018.safetensors",
+    "transformer.h.46.ln_mlp.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.46.mlp.dense_4h_to_h.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.46.mlp.dense_h_to_4h.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.46.self_attention.dense.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.46.self_attention.query_key_value.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.47.ln_attn.bias": "model-00014-of-00018.safetensors",
+    "transformer.h.47.ln_attn.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.47.ln_mlp.bias": "model-00014-of-00018.safetensors",
+    "transformer.h.47.ln_mlp.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.47.mlp.dense_4h_to_h.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.47.mlp.dense_h_to_4h.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.47.self_attention.dense.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.47.self_attention.query_key_value.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.48.ln_attn.bias": "model-00015-of-00018.safetensors",
+    "transformer.h.48.ln_attn.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.48.ln_mlp.bias": "model-00015-of-00018.safetensors",
+    "transformer.h.48.ln_mlp.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.48.mlp.dense_4h_to_h.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.48.mlp.dense_h_to_4h.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.48.self_attention.dense.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.48.self_attention.query_key_value.weight": "model-00014-of-00018.safetensors",
+    "transformer.h.49.ln_attn.bias": "model-00015-of-00018.safetensors",
+    "transformer.h.49.ln_attn.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.49.ln_mlp.bias": "model-00015-of-00018.safetensors",
+    "transformer.h.49.ln_mlp.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.49.mlp.dense_4h_to_h.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.49.mlp.dense_h_to_4h.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.49.self_attention.dense.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.49.self_attention.query_key_value.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.5.ln_attn.bias": "model-00002-of-00018.safetensors",
+    "transformer.h.5.ln_attn.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.5.ln_mlp.bias": "model-00002-of-00018.safetensors",
+    "transformer.h.5.ln_mlp.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.5.mlp.dense_4h_to_h.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.5.mlp.dense_h_to_4h.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.5.self_attention.dense.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.5.self_attention.query_key_value.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.50.ln_attn.bias": "model-00015-of-00018.safetensors",
+    "transformer.h.50.ln_attn.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.50.ln_mlp.bias": "model-00015-of-00018.safetensors",
+    "transformer.h.50.ln_mlp.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.50.mlp.dense_4h_to_h.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.50.mlp.dense_h_to_4h.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.50.self_attention.dense.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.50.self_attention.query_key_value.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.51.ln_attn.bias": "model-00016-of-00018.safetensors",
+    "transformer.h.51.ln_attn.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.51.ln_mlp.bias": "model-00016-of-00018.safetensors",
+    "transformer.h.51.ln_mlp.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.51.mlp.dense_4h_to_h.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.51.mlp.dense_h_to_4h.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.51.self_attention.dense.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.51.self_attention.query_key_value.weight": "model-00015-of-00018.safetensors",
+    "transformer.h.52.ln_attn.bias": "model-00016-of-00018.safetensors",
+    "transformer.h.52.ln_attn.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.52.ln_mlp.bias": "model-00016-of-00018.safetensors",
+    "transformer.h.52.ln_mlp.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.52.mlp.dense_4h_to_h.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.52.mlp.dense_h_to_4h.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.52.self_attention.dense.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.52.self_attention.query_key_value.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.53.ln_attn.bias": "model-00016-of-00018.safetensors",
+    "transformer.h.53.ln_attn.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.53.ln_mlp.bias": "model-00016-of-00018.safetensors",
+    "transformer.h.53.ln_mlp.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.53.mlp.dense_4h_to_h.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.53.mlp.dense_h_to_4h.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.53.self_attention.dense.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.53.self_attention.query_key_value.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.54.ln_attn.bias": "model-00016-of-00018.safetensors",
+    "transformer.h.54.ln_attn.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.54.ln_mlp.bias": "model-00016-of-00018.safetensors",
+    "transformer.h.54.ln_mlp.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.54.mlp.dense_4h_to_h.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.54.mlp.dense_h_to_4h.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.54.self_attention.dense.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.54.self_attention.query_key_value.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.55.ln_attn.bias": "model-00017-of-00018.safetensors",
+    "transformer.h.55.ln_attn.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.55.ln_mlp.bias": "model-00017-of-00018.safetensors",
+    "transformer.h.55.ln_mlp.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.55.mlp.dense_4h_to_h.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.55.mlp.dense_h_to_4h.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.55.self_attention.dense.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.55.self_attention.query_key_value.weight": "model-00016-of-00018.safetensors",
+    "transformer.h.56.ln_attn.bias": "model-00017-of-00018.safetensors",
+    "transformer.h.56.ln_attn.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.56.ln_mlp.bias": "model-00017-of-00018.safetensors",
+    "transformer.h.56.ln_mlp.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.56.mlp.dense_4h_to_h.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.56.mlp.dense_h_to_4h.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.56.self_attention.dense.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.56.self_attention.query_key_value.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.57.ln_attn.bias": "model-00017-of-00018.safetensors",
+    "transformer.h.57.ln_attn.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.57.ln_mlp.bias": "model-00017-of-00018.safetensors",
+    "transformer.h.57.ln_mlp.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.57.mlp.dense_4h_to_h.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.57.mlp.dense_h_to_4h.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.57.self_attention.dense.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.57.self_attention.query_key_value.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.58.ln_attn.bias": "model-00018-of-00018.safetensors",
+    "transformer.h.58.ln_attn.weight": "model-00018-of-00018.safetensors",
+    "transformer.h.58.ln_mlp.bias": "model-00018-of-00018.safetensors",
+    "transformer.h.58.ln_mlp.weight": "model-00018-of-00018.safetensors",
+    "transformer.h.58.mlp.dense_4h_to_h.weight": "model-00018-of-00018.safetensors",
+    "transformer.h.58.mlp.dense_h_to_4h.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.58.self_attention.dense.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.58.self_attention.query_key_value.weight": "model-00017-of-00018.safetensors",
+    "transformer.h.59.ln_attn.bias": "model-00018-of-00018.safetensors",
+    "transformer.h.59.ln_attn.weight": "model-00018-of-00018.safetensors",
+    "transformer.h.59.ln_mlp.bias": "model-00018-of-00018.safetensors",
+    "transformer.h.59.ln_mlp.weight": "model-00018-of-00018.safetensors",
+    "transformer.h.59.mlp.dense_4h_to_h.weight": "model-00018-of-00018.safetensors",
+    "transformer.h.59.mlp.dense_h_to_4h.weight": "model-00018-of-00018.safetensors",
+    "transformer.h.59.self_attention.dense.weight": "model-00018-of-00018.safetensors",
+    "transformer.h.59.self_attention.query_key_value.weight": "model-00018-of-00018.safetensors",
+    "transformer.h.6.ln_attn.bias": "model-00003-of-00018.safetensors",
+    "transformer.h.6.ln_attn.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.6.ln_mlp.bias": "model-00003-of-00018.safetensors",
+    "transformer.h.6.ln_mlp.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.6.mlp.dense_4h_to_h.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.6.mlp.dense_h_to_4h.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.6.self_attention.dense.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.6.self_attention.query_key_value.weight": "model-00002-of-00018.safetensors",
+    "transformer.h.7.ln_attn.bias": "model-00003-of-00018.safetensors",
+    "transformer.h.7.ln_attn.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.7.ln_mlp.bias": "model-00003-of-00018.safetensors",
+    "transformer.h.7.ln_mlp.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.7.mlp.dense_4h_to_h.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.7.mlp.dense_h_to_4h.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.7.self_attention.dense.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.7.self_attention.query_key_value.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.8.ln_attn.bias": "model-00003-of-00018.safetensors",
+    "transformer.h.8.ln_attn.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.8.ln_mlp.bias": "model-00003-of-00018.safetensors",
+    "transformer.h.8.ln_mlp.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.8.mlp.dense_4h_to_h.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.8.mlp.dense_h_to_4h.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.8.self_attention.dense.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.8.self_attention.query_key_value.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.9.ln_attn.bias": "model-00004-of-00018.safetensors",
+    "transformer.h.9.ln_attn.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.9.ln_mlp.bias": "model-00004-of-00018.safetensors",
+    "transformer.h.9.ln_mlp.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.9.mlp.dense_4h_to_h.weight": "model-00004-of-00018.safetensors",
+    "transformer.h.9.mlp.dense_h_to_4h.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.9.self_attention.dense.weight": "model-00003-of-00018.safetensors",
+    "transformer.h.9.self_attention.query_key_value.weight": "model-00003-of-00018.safetensors",
+    "transformer.ln_f.bias": "model-00018-of-00018.safetensors",
+    "transformer.ln_f.weight": "model-00018-of-00018.safetensors",
+    "transformer.word_embeddings.weight": "model-00001-of-00018.safetensors"
+  }
+}