diff --git a/models--LLM360--K2/blobs/0e5448b8c9a7a705501c5103af02f076036414b4cf54083c9ad17048d97221b5 b/models--LLM360--K2/blobs/0e5448b8c9a7a705501c5103af02f076036414b4cf54083c9ad17048d97221b5
deleted file mode 100644
index 4334414923310eb5f2c2a36e9aded6c0a746f067..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/0e5448b8c9a7a705501c5103af02f076036414b4cf54083c9ad17048d97221b5
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0e5448b8c9a7a705501c5103af02f076036414b4cf54083c9ad17048d97221b5
-size 4857111932
diff --git a/models--LLM360--K2/blobs/0ef2f96b2f9b4943f45dc1ba39f0594eb14cf128c5ff5f41db34ad07cb556e7f b/models--LLM360--K2/blobs/0ef2f96b2f9b4943f45dc1ba39f0594eb14cf128c5ff5f41db34ad07cb556e7f
deleted file mode 100644
index 88d73ac21e498fb2dd3444269a572e0b06803f71..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/0ef2f96b2f9b4943f45dc1ba39f0594eb14cf128c5ff5f41db34ad07cb556e7f
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0ef2f96b2f9b4943f45dc1ba39f0594eb14cf128c5ff5f41db34ad07cb556e7f
-size 4857111932
diff --git a/models--LLM360--K2/blobs/197922ce7f7807ba59bfdd42fbbe9c67e2ff8370ffda566614cdf74463363f2b b/models--LLM360--K2/blobs/197922ce7f7807ba59bfdd42fbbe9c67e2ff8370ffda566614cdf74463363f2b
deleted file mode 100644
index 6c8c03014934a9b3c5489dcc704790d9c8fc1dff..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/197922ce7f7807ba59bfdd42fbbe9c67e2ff8370ffda566614cdf74463363f2b
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:197922ce7f7807ba59bfdd42fbbe9c67e2ff8370ffda566614cdf74463363f2b
-size 4857111932
diff --git a/models--LLM360--K2/blobs/1991b47bdfe3cb7e05d4a6f12d1bfb38a22fedf2da68d22269e675978940b57f b/models--LLM360--K2/blobs/1991b47bdfe3cb7e05d4a6f12d1bfb38a22fedf2da68d22269e675978940b57f
deleted file mode 100644
index b6d4f7cb89e6d3607d16078e5054a828e471a95a..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/1991b47bdfe3cb7e05d4a6f12d1bfb38a22fedf2da68d22269e675978940b57f
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1991b47bdfe3cb7e05d4a6f12d1bfb38a22fedf2da68d22269e675978940b57f
-size 4857111932
diff --git a/models--LLM360--K2/blobs/2c9ca528e203129d5e7e95b0f0a915ceb8e0faea b/models--LLM360--K2/blobs/2c9ca528e203129d5e7e95b0f0a915ceb8e0faea
deleted file mode 100644
index 2c9ca528e203129d5e7e95b0f0a915ceb8e0faea..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/2c9ca528e203129d5e7e95b0f0a915ceb8e0faea
+++ /dev/null
@@ -1,730 +0,0 @@
-{
-  "metadata": {
-    "total_size": 130572369920
-  },
-  "weight_map": {
-    "lm_head.weight": "pytorch_model-00027-of-00027.bin",
-    "model.embed_tokens.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.10.input_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.mlp.down_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.mlp.up_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.input_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.11.mlp.down_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.mlp.up_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.12.input_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.mlp.down_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.mlp.up_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.input_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.mlp.down_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.mlp.up_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.input_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.14.mlp.down_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.mlp.up_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.15.input_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.mlp.down_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.mlp.up_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.input_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.mlp.down_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.mlp.up_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.input_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.17.mlp.down_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.mlp.up_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.18.input_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.mlp.down_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.mlp.up_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.input_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.mlp.down_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.mlp.up_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.2.input_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.2.mlp.down_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.2.mlp.up_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.20.input_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.20.mlp.down_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.20.mlp.up_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.21.input_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.mlp.down_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.mlp.up_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.input_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.mlp.down_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.mlp.up_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.input_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.23.mlp.down_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.mlp.up_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.24.input_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.mlp.down_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.mlp.up_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.input_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.mlp.down_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.mlp.up_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.input_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.26.mlp.down_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.mlp.up_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.27.input_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.mlp.down_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.mlp.up_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.input_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.mlp.down_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.mlp.up_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.input_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.29.mlp.down_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.mlp.up_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.3.input_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.mlp.down_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.mlp.up_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.30.input_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.mlp.down_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.mlp.up_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.input_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.mlp.down_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.mlp.up_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.input_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.32.mlp.down_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.32.mlp.gate_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.mlp.up_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.32.post_attention_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.32.self_attn.k_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.self_attn.o_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.self_attn.q_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.self_attn.v_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.33.input_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.mlp.down_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.mlp.gate_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.mlp.up_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.post_attention_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.self_attn.k_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.self_attn.o_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.self_attn.q_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.self_attn.v_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.input_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.mlp.down_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.mlp.gate_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.mlp.up_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.post_attention_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.self_attn.k_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.self_attn.o_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.self_attn.q_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.self_attn.v_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.input_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.35.mlp.down_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.35.mlp.gate_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.mlp.up_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.35.post_attention_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.35.self_attn.k_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.self_attn.o_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.self_attn.q_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.self_attn.v_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.36.input_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.mlp.down_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.mlp.gate_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.mlp.up_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.post_attention_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.self_attn.k_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.self_attn.o_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.self_attn.q_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.self_attn.v_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.input_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.mlp.down_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.mlp.gate_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.mlp.up_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.post_attention_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.self_attn.k_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.self_attn.o_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.self_attn.q_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.self_attn.v_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.input_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.38.mlp.down_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.38.mlp.gate_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.mlp.up_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.38.post_attention_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.38.self_attn.k_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.self_attn.o_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.self_attn.q_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.self_attn.v_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.39.input_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.mlp.down_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.mlp.gate_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.mlp.up_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.post_attention_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.self_attn.k_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.self_attn.o_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.self_attn.q_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.self_attn.v_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.4.input_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.mlp.down_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.mlp.up_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.40.input_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.mlp.down_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.mlp.gate_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.mlp.up_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.post_attention_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.self_attn.k_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.self_attn.o_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.self_attn.q_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.self_attn.v_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.input_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.41.mlp.down_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.41.mlp.gate_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.mlp.up_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.41.post_attention_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.41.self_attn.k_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.self_attn.o_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.self_attn.q_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.self_attn.v_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.42.input_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.mlp.down_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.mlp.gate_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.mlp.up_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.post_attention_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.self_attn.k_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.self_attn.o_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.self_attn.q_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.self_attn.v_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.input_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.mlp.down_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.mlp.gate_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.mlp.up_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.post_attention_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.self_attn.k_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.self_attn.o_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.self_attn.q_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.self_attn.v_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.input_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.44.mlp.down_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.44.mlp.gate_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.mlp.up_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.44.post_attention_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.44.self_attn.k_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.self_attn.o_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.self_attn.q_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.self_attn.v_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.45.input_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.mlp.down_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.mlp.gate_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.mlp.up_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.post_attention_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.self_attn.k_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.self_attn.o_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.self_attn.q_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.self_attn.v_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.input_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.mlp.down_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.mlp.gate_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.mlp.up_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.post_attention_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.self_attn.k_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.self_attn.o_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.self_attn.q_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.self_attn.v_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.input_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.47.mlp.down_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.47.mlp.gate_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.mlp.up_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.47.post_attention_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.47.self_attn.k_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.self_attn.o_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.self_attn.q_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.self_attn.v_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.48.input_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.mlp.down_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.mlp.gate_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.mlp.up_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.post_attention_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.self_attn.k_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.self_attn.o_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.self_attn.q_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.self_attn.v_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.input_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.mlp.down_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.mlp.gate_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.mlp.up_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.post_attention_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.self_attn.k_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.self_attn.o_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.self_attn.q_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.self_attn.v_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.5.input_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.5.mlp.down_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.5.mlp.up_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.50.input_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.50.mlp.down_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.50.mlp.gate_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.50.mlp.up_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.50.post_attention_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.50.self_attn.k_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.50.self_attn.o_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.50.self_attn.q_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.50.self_attn.v_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.51.input_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.mlp.down_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.mlp.gate_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.mlp.up_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.post_attention_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.self_attn.k_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.self_attn.o_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.self_attn.q_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.self_attn.v_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.input_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.mlp.down_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.mlp.gate_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.mlp.up_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.post_attention_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.self_attn.k_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.self_attn.o_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.self_attn.q_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.self_attn.v_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.input_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.53.mlp.down_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.53.mlp.gate_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.mlp.up_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.53.post_attention_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.53.self_attn.k_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.self_attn.o_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.self_attn.q_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.self_attn.v_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.54.input_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.mlp.down_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.mlp.gate_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.mlp.up_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.post_attention_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.self_attn.k_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.self_attn.o_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.self_attn.q_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.self_attn.v_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.input_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.mlp.down_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.mlp.gate_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.mlp.up_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.post_attention_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.self_attn.k_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.self_attn.o_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.self_attn.q_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.self_attn.v_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.input_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.56.mlp.down_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.56.mlp.gate_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.mlp.up_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.56.post_attention_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.56.self_attn.k_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.self_attn.o_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.self_attn.q_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.self_attn.v_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.57.input_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.mlp.down_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.mlp.gate_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.mlp.up_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.post_attention_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.self_attn.k_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.self_attn.o_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.self_attn.q_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.self_attn.v_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.input_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.mlp.down_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.mlp.gate_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.mlp.up_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.post_attention_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.self_attn.k_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.self_attn.o_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.self_attn.q_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.self_attn.v_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.input_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.59.mlp.down_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.59.mlp.gate_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.mlp.up_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.59.post_attention_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.59.self_attn.k_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.self_attn.o_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.self_attn.q_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.self_attn.v_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.6.input_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.mlp.down_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.mlp.up_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.60.input_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.mlp.down_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.mlp.gate_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.mlp.up_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.post_attention_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.self_attn.k_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.self_attn.o_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.self_attn.q_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.self_attn.v_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.input_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.mlp.down_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.mlp.gate_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.mlp.up_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.post_attention_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.self_attn.k_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.self_attn.o_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.self_attn.q_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.self_attn.v_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.input_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.62.mlp.down_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.62.mlp.gate_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.mlp.up_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.62.post_attention_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.62.self_attn.k_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.self_attn.o_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.self_attn.q_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.self_attn.v_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.63.input_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.mlp.down_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.mlp.gate_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.mlp.up_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.post_attention_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.self_attn.k_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.self_attn.o_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.self_attn.q_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.self_attn.v_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.input_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.mlp.down_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.mlp.gate_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.mlp.up_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.post_attention_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.self_attn.k_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.self_attn.o_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.self_attn.q_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.self_attn.v_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.input_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.65.mlp.down_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.65.mlp.gate_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.mlp.up_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.65.post_attention_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.65.self_attn.k_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.self_attn.o_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.self_attn.q_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.self_attn.v_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.66.input_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.mlp.down_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.mlp.gate_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.mlp.up_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.post_attention_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.self_attn.k_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.self_attn.o_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.self_attn.q_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.self_attn.v_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.input_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.mlp.down_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.mlp.gate_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.mlp.up_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.post_attention_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.self_attn.k_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.self_attn.o_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.self_attn.q_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.self_attn.v_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.input_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.68.mlp.down_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.68.mlp.gate_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.mlp.up_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.68.post_attention_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.68.self_attn.k_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.self_attn.o_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.self_attn.q_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.self_attn.v_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.69.input_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.mlp.down_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.mlp.gate_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.mlp.up_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.post_attention_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.self_attn.k_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.self_attn.o_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.self_attn.q_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.self_attn.v_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.7.input_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.mlp.down_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.mlp.up_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.70.input_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.mlp.down_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.mlp.gate_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.mlp.up_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.post_attention_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.self_attn.k_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.self_attn.o_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.self_attn.q_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.self_attn.v_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.input_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.71.mlp.down_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.71.mlp.gate_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.mlp.up_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.71.post_attention_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.71.self_attn.k_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.self_attn.o_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.self_attn.q_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.self_attn.v_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.72.input_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.mlp.down_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.mlp.gate_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.mlp.up_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.post_attention_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.self_attn.k_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.self_attn.o_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.self_attn.q_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.self_attn.v_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.input_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.mlp.down_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.mlp.gate_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.mlp.up_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.post_attention_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.self_attn.k_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.self_attn.o_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.self_attn.q_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.self_attn.v_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.input_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.74.mlp.down_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.74.mlp.gate_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.mlp.up_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.74.post_attention_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.74.self_attn.k_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.self_attn.o_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.self_attn.q_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.self_attn.v_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.75.input_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.mlp.down_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.mlp.gate_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.mlp.up_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.post_attention_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.self_attn.k_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.self_attn.o_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.self_attn.q_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.self_attn.v_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.input_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.mlp.down_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.mlp.gate_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.mlp.up_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.post_attention_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.self_attn.k_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.self_attn.o_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.self_attn.q_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.self_attn.v_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.input_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.77.mlp.down_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.77.mlp.gate_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.mlp.up_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.77.post_attention_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.77.self_attn.k_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.self_attn.o_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.self_attn.q_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.self_attn.v_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.78.input_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.mlp.down_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.mlp.gate_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.mlp.up_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.post_attention_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.self_attn.k_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.self_attn.o_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.self_attn.q_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.self_attn.v_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.input_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.mlp.down_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.mlp.gate_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.mlp.up_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.post_attention_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.self_attn.k_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.self_attn.o_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.self_attn.q_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.self_attn.v_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.8.input_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.8.mlp.down_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.8.mlp.up_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.9.input_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.mlp.down_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.mlp.up_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.norm.weight": "pytorch_model-00027-of-00027.bin"
-  }
-}
diff --git a/models--LLM360--K2/blobs/3bfc25078191c7795def871046f6cb059474fcef43440f4f6b9db0fd757a7f64 b/models--LLM360--K2/blobs/3bfc25078191c7795def871046f6cb059474fcef43440f4f6b9db0fd757a7f64
deleted file mode 100644
index de4ee1d5deabc19cac865bbc30e629233af680c7..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/3bfc25078191c7795def871046f6cb059474fcef43440f4f6b9db0fd757a7f64
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3bfc25078191c7795def871046f6cb059474fcef43440f4f6b9db0fd757a7f64
-size 4857111932
diff --git a/models--LLM360--K2/blobs/3df9e12db23090ffcb07219d5c388eabdd3cc780587d368242b398fdf8e90b4d b/models--LLM360--K2/blobs/3df9e12db23090ffcb07219d5c388eabdd3cc780587d368242b398fdf8e90b4d
deleted file mode 100644
index 8bcff7df49feabcf225a10adaf05bbd9994bab79..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/3df9e12db23090ffcb07219d5c388eabdd3cc780587d368242b398fdf8e90b4d
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3df9e12db23090ffcb07219d5c388eabdd3cc780587d368242b398fdf8e90b4d
-size 4660470238
diff --git a/models--LLM360--K2/blobs/434fd7e790c2995a1f8c0ca46f6a3a5dea2c1d7d0366ed1b47e5bdec80314442 b/models--LLM360--K2/blobs/434fd7e790c2995a1f8c0ca46f6a3a5dea2c1d7d0366ed1b47e5bdec80314442
deleted file mode 100644
index 8a71578f194be800af457ba0eb59b0bbc134f0a4..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/434fd7e790c2995a1f8c0ca46f6a3a5dea2c1d7d0366ed1b47e5bdec80314442
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:434fd7e790c2995a1f8c0ca46f6a3a5dea2c1d7d0366ed1b47e5bdec80314442
-size 4857111932
diff --git a/models--LLM360--K2/blobs/4a1ff5c1228e7d11248f2fcfbd70dc59bb26a4ed737becb63e2e130b028e1698 b/models--LLM360--K2/blobs/4a1ff5c1228e7d11248f2fcfbd70dc59bb26a4ed737becb63e2e130b028e1698
deleted file mode 100644
index 8c7e0b459545099a7d1db2de56fe132440a78dee..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/4a1ff5c1228e7d11248f2fcfbd70dc59bb26a4ed737becb63e2e130b028e1698
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4a1ff5c1228e7d11248f2fcfbd70dc59bb26a4ed737becb63e2e130b028e1698
-size 4857111932
diff --git a/models--LLM360--K2/blobs/4f3b40b1a1fef50d05c3f48ce8a22bd543fd3abe376153290bacc3d37134b463 b/models--LLM360--K2/blobs/4f3b40b1a1fef50d05c3f48ce8a22bd543fd3abe376153290bacc3d37134b463
deleted file mode 100644
index d4027dcc4904f4c76fdd6e6f54a6cf14a68a9e2c..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/4f3b40b1a1fef50d05c3f48ce8a22bd543fd3abe376153290bacc3d37134b463
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4f3b40b1a1fef50d05c3f48ce8a22bd543fd3abe376153290bacc3d37134b463
-size 4857111932
diff --git a/models--LLM360--K2/blobs/583722a14a478af056eb7190401b2fa95c90d75ab1de83acb4d5f20941a224d1 b/models--LLM360--K2/blobs/583722a14a478af056eb7190401b2fa95c90d75ab1de83acb4d5f20941a224d1
deleted file mode 100644
index e2479dce1209d428346e80f6471e127a130ae0db..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/583722a14a478af056eb7190401b2fa95c90d75ab1de83acb4d5f20941a224d1
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:583722a14a478af056eb7190401b2fa95c90d75ab1de83acb4d5f20941a224d1
-size 4857111932
diff --git a/models--LLM360--K2/blobs/5afe6368480862ccd7e5fab96e17159a6a56a6ef2f75083ac04b69df977784da b/models--LLM360--K2/blobs/5afe6368480862ccd7e5fab96e17159a6a56a6ef2f75083ac04b69df977784da
deleted file mode 100644
index f8e0ce37cf7d233a4ea4e6f4dd13ab9aa622d4bc..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/5afe6368480862ccd7e5fab96e17159a6a56a6ef2f75083ac04b69df977784da
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5afe6368480862ccd7e5fab96e17159a6a56a6ef2f75083ac04b69df977784da
-size 4857111932
diff --git a/models--LLM360--K2/blobs/5b4779f96a5161dccf4a59cc5479918f931c47342fa44ab01a9272f476963083 b/models--LLM360--K2/blobs/5b4779f96a5161dccf4a59cc5479918f931c47342fa44ab01a9272f476963083
deleted file mode 100644
index 55734c036070462728de909e93be3f7ddda3f368..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/5b4779f96a5161dccf4a59cc5479918f931c47342fa44ab01a9272f476963083
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5b4779f96a5161dccf4a59cc5479918f931c47342fa44ab01a9272f476963083
-size 4857111932
diff --git a/models--LLM360--K2/blobs/617edceaa3b26e387a4d9aec6e8104db2c81ac35e5bf47f9e518af22e956094a b/models--LLM360--K2/blobs/617edceaa3b26e387a4d9aec6e8104db2c81ac35e5bf47f9e518af22e956094a
deleted file mode 100644
index f31a3868f90bd8ea859270c3e40813a39146c316..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/617edceaa3b26e387a4d9aec6e8104db2c81ac35e5bf47f9e518af22e956094a
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:617edceaa3b26e387a4d9aec6e8104db2c81ac35e5bf47f9e518af22e956094a
-size 4857111932
diff --git a/models--LLM360--K2/blobs/66a1a4140058b1e5040be2ebbb351ba3c73051295dbaf353487860631be06bce b/models--LLM360--K2/blobs/66a1a4140058b1e5040be2ebbb351ba3c73051295dbaf353487860631be06bce
deleted file mode 100644
index 4f4f87cdff64a76266b38ee423da29f413503d80..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/66a1a4140058b1e5040be2ebbb351ba3c73051295dbaf353487860631be06bce
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:66a1a4140058b1e5040be2ebbb351ba3c73051295dbaf353487860631be06bce
-size 4857111932
diff --git a/models--LLM360--K2/blobs/72b7ab829379c5ce42fddf9b8ebae0ff50f967d0cd049af34c796be1dbcc3837 b/models--LLM360--K2/blobs/72b7ab829379c5ce42fddf9b8ebae0ff50f967d0cd049af34c796be1dbcc3837
deleted file mode 100644
index 7c342878228fdd8814ef91a4e76152d2f92c62c5..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/72b7ab829379c5ce42fddf9b8ebae0ff50f967d0cd049af34c796be1dbcc3837
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:72b7ab829379c5ce42fddf9b8ebae0ff50f967d0cd049af34c796be1dbcc3837
-size 4857111932
diff --git a/models--LLM360--K2/blobs/7c546f28758f14aea461f4eedba1dc6380ffb525bcb06328c2e7ccf4bc0914e3 b/models--LLM360--K2/blobs/7c546f28758f14aea461f4eedba1dc6380ffb525bcb06328c2e7ccf4bc0914e3
deleted file mode 100644
index adccb90f2cead3388be14f91c681ed1568989e41..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/7c546f28758f14aea461f4eedba1dc6380ffb525bcb06328c2e7ccf4bc0914e3
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7c546f28758f14aea461f4eedba1dc6380ffb525bcb06328c2e7ccf4bc0914e3
-size 4857111932
diff --git a/models--LLM360--K2/blobs/7ef2e098b4cf22b9facdb8f1df3f5ef4d1416f16 b/models--LLM360--K2/blobs/7ef2e098b4cf22b9facdb8f1df3f5ef4d1416f16
deleted file mode 100644
index 7ef2e098b4cf22b9facdb8f1df3f5ef4d1416f16..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/7ef2e098b4cf22b9facdb8f1df3f5ef4d1416f16
+++ /dev/null
@@ -1,30 +0,0 @@
-{
-  "_name_or_path": "huggyllama/llama-65b",
-  "architectures": [
-    "LlamaForCausalLM"
-  ],
-  "attention_bias": false,
-  "attention_dropout": 0.0,
-  "bos_token_id": null,
-  "eos_token_id": 2,
-  "hidden_act": "silu",
-  "hidden_size": 8192,
-  "initializer_range": 0.02,
-  "intermediate_size": 22016,
-  "max_position_embeddings": 2048,
-  "max_sequence_length": 2048,
-  "model_type": "llama",
-  "num_attention_heads": 64,
-  "num_hidden_layers": 80,
-  "num_key_value_heads": 64,
-  "pad_token_id": 0,
-  "pretraining_tp": 1,
-  "rms_norm_eps": 1e-05,
-  "rope_scaling": null,
-  "rope_theta": 10000.0,
-  "tie_word_embeddings": false,
-  "torch_dtype": "float16",
-  "transformers_version": "4.39.3",
-  "use_cache": true,
-  "vocab_size": 32032
-}
diff --git a/models--LLM360--K2/blobs/80860954ce526545b5eaa951ed98a9dc87576e0ce5243b53001e79a90f107ac2 b/models--LLM360--K2/blobs/80860954ce526545b5eaa951ed98a9dc87576e0ce5243b53001e79a90f107ac2
deleted file mode 100644
index 473828bc795a1cfe82dd06384d9e310e54bcb93b..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/80860954ce526545b5eaa951ed98a9dc87576e0ce5243b53001e79a90f107ac2
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:80860954ce526545b5eaa951ed98a9dc87576e0ce5243b53001e79a90f107ac2
-size 4857111932
diff --git a/models--LLM360--K2/blobs/8fecd1dfe9745e2b6546298051b39dd02ace8b88eb0232a9d6a9beccb935f9b6 b/models--LLM360--K2/blobs/8fecd1dfe9745e2b6546298051b39dd02ace8b88eb0232a9d6a9beccb935f9b6
deleted file mode 100644
index 927da3dfc925b6b1e4a2bfa0c2ed5c33b20ee8a8..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/8fecd1dfe9745e2b6546298051b39dd02ace8b88eb0232a9d6a9beccb935f9b6
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:8fecd1dfe9745e2b6546298051b39dd02ace8b88eb0232a9d6a9beccb935f9b6
-size 4857111932
diff --git a/models--LLM360--K2/blobs/9844405d62d9f809278ed1577b4364f68327282f461b4afaf75a9cebf074c325 b/models--LLM360--K2/blobs/9844405d62d9f809278ed1577b4364f68327282f461b4afaf75a9cebf074c325
deleted file mode 100644
index 96e6b2fe882b10c748aff2749319da02d5e10070..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/9844405d62d9f809278ed1577b4364f68327282f461b4afaf75a9cebf074c325
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9844405d62d9f809278ed1577b4364f68327282f461b4afaf75a9cebf074c325
-size 4857111932
diff --git a/models--LLM360--K2/blobs/a258e4bed92e87cfeacce9077d9e2989e45436d6259e6f16f1cace7d7fcdd6d8 b/models--LLM360--K2/blobs/a258e4bed92e87cfeacce9077d9e2989e45436d6259e6f16f1cace7d7fcdd6d8
deleted file mode 100644
index 32a98e1e90f4e5e1e55ccdb2f7cc90af1db05a28..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/a258e4bed92e87cfeacce9077d9e2989e45436d6259e6f16f1cace7d7fcdd6d8
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a258e4bed92e87cfeacce9077d9e2989e45436d6259e6f16f1cace7d7fcdd6d8
-size 4857111932
diff --git a/models--LLM360--K2/blobs/b6c7e1e011596a62423becf7831de6885bd4cb66835c9b12a1890d82341caa5d b/models--LLM360--K2/blobs/b6c7e1e011596a62423becf7831de6885bd4cb66835c9b12a1890d82341caa5d
deleted file mode 100644
index 6d69d5a1e997993d3e8db4ff0b4d8ecfeef3fc79..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/b6c7e1e011596a62423becf7831de6885bd4cb66835c9b12a1890d82341caa5d
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b6c7e1e011596a62423becf7831de6885bd4cb66835c9b12a1890d82341caa5d
-size 4857111932
diff --git a/models--LLM360--K2/blobs/c0c42dfdce41209444c998d0f30f5bdabf5e747f9ba18f5a917a3e72c8a4956b b/models--LLM360--K2/blobs/c0c42dfdce41209444c998d0f30f5bdabf5e747f9ba18f5a917a3e72c8a4956b
deleted file mode 100644
index 7e0c204c4154f8e1a08d3420af0fbde0363bdbee..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/c0c42dfdce41209444c998d0f30f5bdabf5e747f9ba18f5a917a3e72c8a4956b
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c0c42dfdce41209444c998d0f30f5bdabf5e747f9ba18f5a917a3e72c8a4956b
-size 4484358522
diff --git a/models--LLM360--K2/blobs/c1f7a86a6d975dad016a1692a17b3c8116c1ff4ef04362060addfc1b65aa2c08 b/models--LLM360--K2/blobs/c1f7a86a6d975dad016a1692a17b3c8116c1ff4ef04362060addfc1b65aa2c08
deleted file mode 100644
index 4f09592e86545f1e79e6062d42b0c0f1a9de5781..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/c1f7a86a6d975dad016a1692a17b3c8116c1ff4ef04362060addfc1b65aa2c08
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c1f7a86a6d975dad016a1692a17b3c8116c1ff4ef04362060addfc1b65aa2c08
-size 4857111932
diff --git a/models--LLM360--K2/blobs/c55743fc3d25e389dc4e29f7bb2db5e21f033fc8292ffe411eed32fdff0b2498 b/models--LLM360--K2/blobs/c55743fc3d25e389dc4e29f7bb2db5e21f033fc8292ffe411eed32fdff0b2498
deleted file mode 100644
index d9b593b7286dc65ba7c5bd87b9a457be971dda75..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/c55743fc3d25e389dc4e29f7bb2db5e21f033fc8292ffe411eed32fdff0b2498
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c55743fc3d25e389dc4e29f7bb2db5e21f033fc8292ffe411eed32fdff0b2498
-size 4857111932
diff --git a/models--LLM360--K2/blobs/c9a048c6dfcef5561324dbded792521edfb1d9c7a03404189d2a09157c628760 b/models--LLM360--K2/blobs/c9a048c6dfcef5561324dbded792521edfb1d9c7a03404189d2a09157c628760
deleted file mode 100644
index c8ae1d76b8fa5ee28b52bd966aaa954def23d19a..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/c9a048c6dfcef5561324dbded792521edfb1d9c7a03404189d2a09157c628760
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c9a048c6dfcef5561324dbded792521edfb1d9c7a03404189d2a09157c628760
-size 4857111932
diff --git a/models--LLM360--K2/blobs/ca618179df960063f7086853a0de59fd2fcfa2b7a220fe089d66743aa6648b8e b/models--LLM360--K2/blobs/ca618179df960063f7086853a0de59fd2fcfa2b7a220fe089d66743aa6648b8e
deleted file mode 100644
index 3c866c22eac79894d0a18bc0a8239fe697810e71..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/ca618179df960063f7086853a0de59fd2fcfa2b7a220fe089d66743aa6648b8e
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ca618179df960063f7086853a0de59fd2fcfa2b7a220fe089d66743aa6648b8e
-size 4857111932
diff --git a/models--LLM360--K2/blobs/d3ebaa9cb6f5e87aa968768383824375290dc33e56a1852e48821fe285fd8448 b/models--LLM360--K2/blobs/d3ebaa9cb6f5e87aa968768383824375290dc33e56a1852e48821fe285fd8448
deleted file mode 100644
index 7f6e74fcd068d36f896bf19154e1cb2fdfa1304a..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/blobs/d3ebaa9cb6f5e87aa968768383824375290dc33e56a1852e48821fe285fd8448
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d3ebaa9cb6f5e87aa968768383824375290dc33e56a1852e48821fe285fd8448
-size 4857111932
diff --git a/models--LLM360--K2/refs/main b/models--LLM360--K2/refs/main
deleted file mode 100644
index 7a0ce76bf946b4efce99069a01002bf06ddd8bff..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/refs/main
+++ /dev/null
@@ -1 +0,0 @@
-ccd3ced964e211a9c31cb346444bce761d26519e
\ No newline at end of file
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/config.json b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/config.json
deleted file mode 100644
index 7ef2e098b4cf22b9facdb8f1df3f5ef4d1416f16..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/config.json
+++ /dev/null
@@ -1,30 +0,0 @@
-{
-  "_name_or_path": "huggyllama/llama-65b",
-  "architectures": [
-    "LlamaForCausalLM"
-  ],
-  "attention_bias": false,
-  "attention_dropout": 0.0,
-  "bos_token_id": null,
-  "eos_token_id": 2,
-  "hidden_act": "silu",
-  "hidden_size": 8192,
-  "initializer_range": 0.02,
-  "intermediate_size": 22016,
-  "max_position_embeddings": 2048,
-  "max_sequence_length": 2048,
-  "model_type": "llama",
-  "num_attention_heads": 64,
-  "num_hidden_layers": 80,
-  "num_key_value_heads": 64,
-  "pad_token_id": 0,
-  "pretraining_tp": 1,
-  "rms_norm_eps": 1e-05,
-  "rope_scaling": null,
-  "rope_theta": 10000.0,
-  "tie_word_embeddings": false,
-  "torch_dtype": "float16",
-  "transformers_version": "4.39.3",
-  "use_cache": true,
-  "vocab_size": 32032
-}
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00001-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00001-of-00027.bin
deleted file mode 100644
index 8bcff7df49feabcf225a10adaf05bbd9994bab79..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00001-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3df9e12db23090ffcb07219d5c388eabdd3cc780587d368242b398fdf8e90b4d
-size 4660470238
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00002-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00002-of-00027.bin
deleted file mode 100644
index b6d4f7cb89e6d3607d16078e5054a828e471a95a..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00002-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1991b47bdfe3cb7e05d4a6f12d1bfb38a22fedf2da68d22269e675978940b57f
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00003-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00003-of-00027.bin
deleted file mode 100644
index 96e6b2fe882b10c748aff2749319da02d5e10070..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00003-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9844405d62d9f809278ed1577b4364f68327282f461b4afaf75a9cebf074c325
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00004-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00004-of-00027.bin
deleted file mode 100644
index c8ae1d76b8fa5ee28b52bd966aaa954def23d19a..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00004-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c9a048c6dfcef5561324dbded792521edfb1d9c7a03404189d2a09157c628760
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00005-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00005-of-00027.bin
deleted file mode 100644
index 4334414923310eb5f2c2a36e9aded6c0a746f067..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00005-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0e5448b8c9a7a705501c5103af02f076036414b4cf54083c9ad17048d97221b5
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00006-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00006-of-00027.bin
deleted file mode 100644
index e2479dce1209d428346e80f6471e127a130ae0db..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00006-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:583722a14a478af056eb7190401b2fa95c90d75ab1de83acb4d5f20941a224d1
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00007-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00007-of-00027.bin
deleted file mode 100644
index 4f09592e86545f1e79e6062d42b0c0f1a9de5781..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00007-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c1f7a86a6d975dad016a1692a17b3c8116c1ff4ef04362060addfc1b65aa2c08
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00008-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00008-of-00027.bin
deleted file mode 100644
index f8e0ce37cf7d233a4ea4e6f4dd13ab9aa622d4bc..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00008-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5afe6368480862ccd7e5fab96e17159a6a56a6ef2f75083ac04b69df977784da
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00009-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00009-of-00027.bin
deleted file mode 100644
index 6d69d5a1e997993d3e8db4ff0b4d8ecfeef3fc79..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00009-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b6c7e1e011596a62423becf7831de6885bd4cb66835c9b12a1890d82341caa5d
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00010-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00010-of-00027.bin
deleted file mode 100644
index 6c8c03014934a9b3c5489dcc704790d9c8fc1dff..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00010-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:197922ce7f7807ba59bfdd42fbbe9c67e2ff8370ffda566614cdf74463363f2b
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00011-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00011-of-00027.bin
deleted file mode 100644
index 7c342878228fdd8814ef91a4e76152d2f92c62c5..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00011-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:72b7ab829379c5ce42fddf9b8ebae0ff50f967d0cd049af34c796be1dbcc3837
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00012-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00012-of-00027.bin
deleted file mode 100644
index 7f6e74fcd068d36f896bf19154e1cb2fdfa1304a..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00012-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d3ebaa9cb6f5e87aa968768383824375290dc33e56a1852e48821fe285fd8448
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00013-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00013-of-00027.bin
deleted file mode 100644
index 3c866c22eac79894d0a18bc0a8239fe697810e71..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00013-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ca618179df960063f7086853a0de59fd2fcfa2b7a220fe089d66743aa6648b8e
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00014-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00014-of-00027.bin
deleted file mode 100644
index 32a98e1e90f4e5e1e55ccdb2f7cc90af1db05a28..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00014-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a258e4bed92e87cfeacce9077d9e2989e45436d6259e6f16f1cace7d7fcdd6d8
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00015-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00015-of-00027.bin
deleted file mode 100644
index de4ee1d5deabc19cac865bbc30e629233af680c7..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00015-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3bfc25078191c7795def871046f6cb059474fcef43440f4f6b9db0fd757a7f64
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00016-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00016-of-00027.bin
deleted file mode 100644
index 927da3dfc925b6b1e4a2bfa0c2ed5c33b20ee8a8..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00016-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:8fecd1dfe9745e2b6546298051b39dd02ace8b88eb0232a9d6a9beccb935f9b6
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00017-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00017-of-00027.bin
deleted file mode 100644
index d9b593b7286dc65ba7c5bd87b9a457be971dda75..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00017-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c55743fc3d25e389dc4e29f7bb2db5e21f033fc8292ffe411eed32fdff0b2498
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00018-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00018-of-00027.bin
deleted file mode 100644
index f31a3868f90bd8ea859270c3e40813a39146c316..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00018-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:617edceaa3b26e387a4d9aec6e8104db2c81ac35e5bf47f9e518af22e956094a
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00019-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00019-of-00027.bin
deleted file mode 100644
index 8a71578f194be800af457ba0eb59b0bbc134f0a4..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00019-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:434fd7e790c2995a1f8c0ca46f6a3a5dea2c1d7d0366ed1b47e5bdec80314442
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00020-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00020-of-00027.bin
deleted file mode 100644
index d4027dcc4904f4c76fdd6e6f54a6cf14a68a9e2c..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00020-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4f3b40b1a1fef50d05c3f48ce8a22bd543fd3abe376153290bacc3d37134b463
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00021-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00021-of-00027.bin
deleted file mode 100644
index 55734c036070462728de909e93be3f7ddda3f368..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00021-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5b4779f96a5161dccf4a59cc5479918f931c47342fa44ab01a9272f476963083
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00022-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00022-of-00027.bin
deleted file mode 100644
index 4f4f87cdff64a76266b38ee423da29f413503d80..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00022-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:66a1a4140058b1e5040be2ebbb351ba3c73051295dbaf353487860631be06bce
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00023-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00023-of-00027.bin
deleted file mode 100644
index 8c7e0b459545099a7d1db2de56fe132440a78dee..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00023-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4a1ff5c1228e7d11248f2fcfbd70dc59bb26a4ed737becb63e2e130b028e1698
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00024-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00024-of-00027.bin
deleted file mode 100644
index adccb90f2cead3388be14f91c681ed1568989e41..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00024-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7c546f28758f14aea461f4eedba1dc6380ffb525bcb06328c2e7ccf4bc0914e3
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00025-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00025-of-00027.bin
deleted file mode 100644
index 473828bc795a1cfe82dd06384d9e310e54bcb93b..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00025-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:80860954ce526545b5eaa951ed98a9dc87576e0ce5243b53001e79a90f107ac2
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00026-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00026-of-00027.bin
deleted file mode 100644
index 88d73ac21e498fb2dd3444269a572e0b06803f71..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00026-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0ef2f96b2f9b4943f45dc1ba39f0594eb14cf128c5ff5f41db34ad07cb556e7f
-size 4857111932
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00027-of-00027.bin b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00027-of-00027.bin
deleted file mode 100644
index 7e0c204c4154f8e1a08d3420af0fbde0363bdbee..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model-00027-of-00027.bin
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c0c42dfdce41209444c998d0f30f5bdabf5e747f9ba18f5a917a3e72c8a4956b
-size 4484358522
diff --git a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model.bin.index.json b/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model.bin.index.json
deleted file mode 100644
index 2c9ca528e203129d5e7e95b0f0a915ceb8e0faea..0000000000000000000000000000000000000000
--- a/models--LLM360--K2/snapshots/ccd3ced964e211a9c31cb346444bce761d26519e/pytorch_model.bin.index.json
+++ /dev/null
@@ -1,730 +0,0 @@
-{
-  "metadata": {
-    "total_size": 130572369920
-  },
-  "weight_map": {
-    "lm_head.weight": "pytorch_model-00027-of-00027.bin",
-    "model.embed_tokens.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.10.input_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.mlp.down_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.mlp.up_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.input_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.11.mlp.down_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.mlp.up_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.12.input_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.mlp.down_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.mlp.up_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.input_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.mlp.down_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.mlp.up_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.input_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.14.mlp.down_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.mlp.up_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00005-of-00027.bin",
-    "model.layers.15.input_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.mlp.down_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.mlp.up_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.input_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.mlp.down_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.mlp.up_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.input_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.17.mlp.down_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.mlp.up_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00006-of-00027.bin",
-    "model.layers.18.input_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.mlp.down_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.mlp.up_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.input_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.mlp.down_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.mlp.up_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.2.input_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.2.mlp.down_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.2.mlp.up_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00027.bin",
-    "model.layers.20.input_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.20.mlp.down_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.20.mlp.up_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00007-of-00027.bin",
-    "model.layers.21.input_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.mlp.down_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.mlp.up_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.input_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.mlp.down_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.mlp.up_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.input_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.23.mlp.down_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.mlp.up_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00008-of-00027.bin",
-    "model.layers.24.input_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.mlp.down_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.mlp.up_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.input_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.mlp.down_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.mlp.up_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.input_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.26.mlp.down_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.mlp.up_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00009-of-00027.bin",
-    "model.layers.27.input_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.mlp.down_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.mlp.up_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.input_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.mlp.down_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.mlp.up_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.input_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.29.mlp.down_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.mlp.up_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00010-of-00027.bin",
-    "model.layers.3.input_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.mlp.down_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.mlp.up_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.30.input_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.mlp.down_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.mlp.up_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.input_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.mlp.down_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.mlp.up_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.input_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.32.mlp.down_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.32.mlp.gate_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.mlp.up_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.32.post_attention_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.32.self_attn.k_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.self_attn.o_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.self_attn.q_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.32.self_attn.v_proj.weight": "pytorch_model-00011-of-00027.bin",
-    "model.layers.33.input_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.mlp.down_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.mlp.gate_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.mlp.up_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.post_attention_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.self_attn.k_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.self_attn.o_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.self_attn.q_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.33.self_attn.v_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.input_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.mlp.down_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.mlp.gate_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.mlp.up_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.post_attention_layernorm.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.self_attn.k_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.self_attn.o_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.self_attn.q_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.34.self_attn.v_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.input_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.35.mlp.down_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.35.mlp.gate_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.mlp.up_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.35.post_attention_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.35.self_attn.k_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.self_attn.o_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.self_attn.q_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.35.self_attn.v_proj.weight": "pytorch_model-00012-of-00027.bin",
-    "model.layers.36.input_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.mlp.down_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.mlp.gate_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.mlp.up_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.post_attention_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.self_attn.k_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.self_attn.o_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.self_attn.q_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.36.self_attn.v_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.input_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.mlp.down_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.mlp.gate_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.mlp.up_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.post_attention_layernorm.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.self_attn.k_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.self_attn.o_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.self_attn.q_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.37.self_attn.v_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.input_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.38.mlp.down_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.38.mlp.gate_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.mlp.up_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.38.post_attention_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.38.self_attn.k_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.self_attn.o_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.self_attn.q_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.38.self_attn.v_proj.weight": "pytorch_model-00013-of-00027.bin",
-    "model.layers.39.input_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.mlp.down_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.mlp.gate_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.mlp.up_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.post_attention_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.self_attn.k_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.self_attn.o_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.self_attn.q_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.39.self_attn.v_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.4.input_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.mlp.down_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.mlp.up_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.40.input_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.mlp.down_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.mlp.gate_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.mlp.up_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.post_attention_layernorm.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.self_attn.k_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.self_attn.o_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.self_attn.q_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.40.self_attn.v_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.input_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.41.mlp.down_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.41.mlp.gate_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.mlp.up_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.41.post_attention_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.41.self_attn.k_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.self_attn.o_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.self_attn.q_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.41.self_attn.v_proj.weight": "pytorch_model-00014-of-00027.bin",
-    "model.layers.42.input_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.mlp.down_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.mlp.gate_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.mlp.up_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.post_attention_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.self_attn.k_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.self_attn.o_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.self_attn.q_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.42.self_attn.v_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.input_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.mlp.down_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.mlp.gate_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.mlp.up_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.post_attention_layernorm.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.self_attn.k_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.self_attn.o_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.self_attn.q_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.43.self_attn.v_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.input_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.44.mlp.down_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.44.mlp.gate_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.mlp.up_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.44.post_attention_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.44.self_attn.k_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.self_attn.o_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.self_attn.q_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.44.self_attn.v_proj.weight": "pytorch_model-00015-of-00027.bin",
-    "model.layers.45.input_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.mlp.down_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.mlp.gate_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.mlp.up_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.post_attention_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.self_attn.k_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.self_attn.o_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.self_attn.q_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.45.self_attn.v_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.input_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.mlp.down_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.mlp.gate_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.mlp.up_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.post_attention_layernorm.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.self_attn.k_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.self_attn.o_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.self_attn.q_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.46.self_attn.v_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.input_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.47.mlp.down_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.47.mlp.gate_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.mlp.up_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.47.post_attention_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.47.self_attn.k_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.self_attn.o_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.self_attn.q_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.47.self_attn.v_proj.weight": "pytorch_model-00016-of-00027.bin",
-    "model.layers.48.input_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.mlp.down_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.mlp.gate_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.mlp.up_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.post_attention_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.self_attn.k_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.self_attn.o_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.self_attn.q_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.48.self_attn.v_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.input_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.mlp.down_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.mlp.gate_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.mlp.up_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.post_attention_layernorm.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.self_attn.k_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.self_attn.o_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.self_attn.q_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.49.self_attn.v_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.5.input_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.5.mlp.down_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.5.mlp.up_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00002-of-00027.bin",
-    "model.layers.50.input_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.50.mlp.down_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.50.mlp.gate_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.50.mlp.up_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.50.post_attention_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.50.self_attn.k_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.50.self_attn.o_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.50.self_attn.q_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.50.self_attn.v_proj.weight": "pytorch_model-00017-of-00027.bin",
-    "model.layers.51.input_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.mlp.down_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.mlp.gate_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.mlp.up_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.post_attention_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.self_attn.k_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.self_attn.o_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.self_attn.q_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.51.self_attn.v_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.input_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.mlp.down_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.mlp.gate_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.mlp.up_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.post_attention_layernorm.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.self_attn.k_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.self_attn.o_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.self_attn.q_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.52.self_attn.v_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.input_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.53.mlp.down_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.53.mlp.gate_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.mlp.up_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.53.post_attention_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.53.self_attn.k_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.self_attn.o_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.self_attn.q_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.53.self_attn.v_proj.weight": "pytorch_model-00018-of-00027.bin",
-    "model.layers.54.input_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.mlp.down_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.mlp.gate_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.mlp.up_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.post_attention_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.self_attn.k_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.self_attn.o_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.self_attn.q_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.54.self_attn.v_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.input_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.mlp.down_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.mlp.gate_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.mlp.up_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.post_attention_layernorm.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.self_attn.k_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.self_attn.o_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.self_attn.q_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.55.self_attn.v_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.input_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.56.mlp.down_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.56.mlp.gate_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.mlp.up_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.56.post_attention_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.56.self_attn.k_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.self_attn.o_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.self_attn.q_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.56.self_attn.v_proj.weight": "pytorch_model-00019-of-00027.bin",
-    "model.layers.57.input_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.mlp.down_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.mlp.gate_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.mlp.up_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.post_attention_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.self_attn.k_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.self_attn.o_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.self_attn.q_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.57.self_attn.v_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.input_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.mlp.down_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.mlp.gate_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.mlp.up_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.post_attention_layernorm.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.self_attn.k_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.self_attn.o_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.self_attn.q_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.58.self_attn.v_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.input_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.59.mlp.down_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.59.mlp.gate_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.mlp.up_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.59.post_attention_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.59.self_attn.k_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.self_attn.o_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.self_attn.q_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.59.self_attn.v_proj.weight": "pytorch_model-00020-of-00027.bin",
-    "model.layers.6.input_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.mlp.down_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.mlp.up_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.60.input_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.mlp.down_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.mlp.gate_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.mlp.up_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.post_attention_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.self_attn.k_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.self_attn.o_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.self_attn.q_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.60.self_attn.v_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.input_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.mlp.down_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.mlp.gate_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.mlp.up_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.post_attention_layernorm.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.self_attn.k_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.self_attn.o_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.self_attn.q_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.61.self_attn.v_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.input_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.62.mlp.down_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.62.mlp.gate_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.mlp.up_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.62.post_attention_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.62.self_attn.k_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.self_attn.o_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.self_attn.q_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.62.self_attn.v_proj.weight": "pytorch_model-00021-of-00027.bin",
-    "model.layers.63.input_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.mlp.down_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.mlp.gate_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.mlp.up_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.post_attention_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.self_attn.k_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.self_attn.o_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.self_attn.q_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.63.self_attn.v_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.input_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.mlp.down_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.mlp.gate_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.mlp.up_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.post_attention_layernorm.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.self_attn.k_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.self_attn.o_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.self_attn.q_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.64.self_attn.v_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.input_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.65.mlp.down_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.65.mlp.gate_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.mlp.up_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.65.post_attention_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.65.self_attn.k_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.self_attn.o_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.self_attn.q_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.65.self_attn.v_proj.weight": "pytorch_model-00022-of-00027.bin",
-    "model.layers.66.input_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.mlp.down_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.mlp.gate_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.mlp.up_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.post_attention_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.self_attn.k_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.self_attn.o_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.self_attn.q_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.66.self_attn.v_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.input_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.mlp.down_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.mlp.gate_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.mlp.up_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.post_attention_layernorm.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.self_attn.k_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.self_attn.o_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.self_attn.q_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.67.self_attn.v_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.input_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.68.mlp.down_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.68.mlp.gate_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.mlp.up_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.68.post_attention_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.68.self_attn.k_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.self_attn.o_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.self_attn.q_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.68.self_attn.v_proj.weight": "pytorch_model-00023-of-00027.bin",
-    "model.layers.69.input_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.mlp.down_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.mlp.gate_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.mlp.up_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.post_attention_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.self_attn.k_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.self_attn.o_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.self_attn.q_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.69.self_attn.v_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.7.input_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.mlp.down_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.mlp.up_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.70.input_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.mlp.down_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.mlp.gate_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.mlp.up_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.post_attention_layernorm.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.self_attn.k_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.self_attn.o_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.self_attn.q_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.70.self_attn.v_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.input_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.71.mlp.down_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.71.mlp.gate_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.mlp.up_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.71.post_attention_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.71.self_attn.k_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.self_attn.o_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.self_attn.q_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.71.self_attn.v_proj.weight": "pytorch_model-00024-of-00027.bin",
-    "model.layers.72.input_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.mlp.down_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.mlp.gate_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.mlp.up_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.post_attention_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.self_attn.k_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.self_attn.o_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.self_attn.q_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.72.self_attn.v_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.input_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.mlp.down_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.mlp.gate_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.mlp.up_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.post_attention_layernorm.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.self_attn.k_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.self_attn.o_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.self_attn.q_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.73.self_attn.v_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.input_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.74.mlp.down_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.74.mlp.gate_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.mlp.up_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.74.post_attention_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.74.self_attn.k_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.self_attn.o_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.self_attn.q_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.74.self_attn.v_proj.weight": "pytorch_model-00025-of-00027.bin",
-    "model.layers.75.input_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.mlp.down_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.mlp.gate_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.mlp.up_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.post_attention_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.self_attn.k_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.self_attn.o_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.self_attn.q_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.75.self_attn.v_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.input_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.mlp.down_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.mlp.gate_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.mlp.up_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.post_attention_layernorm.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.self_attn.k_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.self_attn.o_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.self_attn.q_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.76.self_attn.v_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.input_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.77.mlp.down_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.77.mlp.gate_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.mlp.up_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.77.post_attention_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.77.self_attn.k_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.self_attn.o_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.self_attn.q_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.77.self_attn.v_proj.weight": "pytorch_model-00026-of-00027.bin",
-    "model.layers.78.input_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.mlp.down_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.mlp.gate_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.mlp.up_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.post_attention_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.self_attn.k_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.self_attn.o_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.self_attn.q_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.78.self_attn.v_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.input_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.mlp.down_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.mlp.gate_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.mlp.up_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.post_attention_layernorm.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.self_attn.k_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.self_attn.o_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.self_attn.q_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.79.self_attn.v_proj.weight": "pytorch_model-00027-of-00027.bin",
-    "model.layers.8.input_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.8.mlp.down_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.8.mlp.up_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00003-of-00027.bin",
-    "model.layers.9.input_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.mlp.down_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.mlp.up_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00004-of-00027.bin",
-    "model.norm.weight": "pytorch_model-00027-of-00027.bin"
-  }
-}