ehartford commited on
Commit
7b59dcd
1 Parent(s): a11534e

Upload folder using huggingface_hub

Browse files
config.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/workspace/samantha-30b",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "bos_token_id": 1,
7
+ "eos_token_id": 2,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 6656,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 17920,
12
+ "max_position_embeddings": 2048,
13
+ "max_sequence_length": 2048,
14
+ "model_type": "llama",
15
+ "num_attention_heads": 52,
16
+ "num_hidden_layers": 60,
17
+ "pad_token_id": 0,
18
+ "rms_norm_eps": 1e-06,
19
+ "tie_word_embeddings": false,
20
+ "torch_dtype": "float32",
21
+ "transformers_version": "4.28.1",
22
+ "use_cache": true,
23
+ "vocab_size": 32000
24
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.28.1"
7
+ }
latest ADDED
@@ -0,0 +1 @@
 
 
1
+ global_step135
pytorch_model-00001-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a9858c080dd098169fb146daaf7710a525f429537df18f9e68722a90c0be897
3
+ size 9944399515
pytorch_model-00002-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6376759861419461cf4c401d409e397e0fcf665c66ac1bf39da7878f323d35a
3
+ size 9692217267
pytorch_model-00003-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1013247cbf19a9ae2930107d7d27fbe8701ece4a8aae2624a35e5209fe97c7d0
3
+ size 9746797807
pytorch_model-00004-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b56eb6db9231c8d66e145ff6708fda3e40c8614561236d48901473175443853b
3
+ size 9992163381
pytorch_model-00005-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:70cf39d14d8394af59ea63af3ae43fd7d9b7fdbb48d8f737f33a78141b48dca6
3
+ size 9746743895
pytorch_model-00006-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c4564411acb0e803b6f54c92933b514c19f0fa673bb4ef04009a4c43ffbd5b2
3
+ size 9869480291
pytorch_model-00007-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fcc00921f8800f633fb8f50d0a5f39691a94969e0875443033d3e1a034fc81a0
3
+ size 9869426913
pytorch_model-00008-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:446f50d95e2f5f5f13a243552c4eff54f9579061a6df69170f7e1d1f170d3010
3
+ size 9746797807
pytorch_model-00009-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:418eb626ded7b245765980c67473bdefa452e9bffe4288041118599355007302
3
+ size 9992163381
pytorch_model-00010-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a166ce764af5c34b1b3698675d2bed75135980f2c21b45b79e57cf758b4eb78
3
+ size 9746743895
pytorch_model-00011-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6126e9f4dcc0c73a84ce3d038c06aee8eff3f3ee175175888e3b2ed207da3d7
3
+ size 9869480291
pytorch_model-00012-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:163829b04fc2bd4135ede595a3ce07f7aabd5663f55a3fb1b9d30fc09c5d47c2
3
+ size 9869426913
pytorch_model-00013-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9774ab12ae6469a4a602c4e13203afc401c752551e5e356a78771749ece7c562
3
+ size 9746797807
pytorch_model-00014-of-00014.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b5d4f313b4944271d8135b0de5c781bd23d006d36c6fb2de4596d8f9359c763
3
+ size 2283356880
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,610 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 130115789824
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "pytorch_model-00014-of-00014.bin",
7
+ "model.embed_tokens.weight": "pytorch_model-00001-of-00014.bin",
8
+ "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00014.bin",
9
+ "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00014.bin",
10
+ "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00014.bin",
11
+ "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00014.bin",
12
+ "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00014.bin",
13
+ "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00014.bin",
14
+ "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00014.bin",
15
+ "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00014.bin",
16
+ "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00014.bin",
17
+ "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00014.bin",
18
+ "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00014.bin",
19
+ "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00014.bin",
20
+ "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00014.bin",
21
+ "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00014.bin",
22
+ "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00014.bin",
23
+ "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00014.bin",
24
+ "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00014.bin",
25
+ "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00014.bin",
26
+ "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00014.bin",
27
+ "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00014.bin",
28
+ "model.layers.10.input_layernorm.weight": "pytorch_model-00003-of-00014.bin",
29
+ "model.layers.10.mlp.down_proj.weight": "pytorch_model-00003-of-00014.bin",
30
+ "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00003-of-00014.bin",
31
+ "model.layers.10.mlp.up_proj.weight": "pytorch_model-00003-of-00014.bin",
32
+ "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00003-of-00014.bin",
33
+ "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00003-of-00014.bin",
34
+ "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00003-of-00014.bin",
35
+ "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00003-of-00014.bin",
36
+ "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00014.bin",
37
+ "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00003-of-00014.bin",
38
+ "model.layers.11.input_layernorm.weight": "pytorch_model-00003-of-00014.bin",
39
+ "model.layers.11.mlp.down_proj.weight": "pytorch_model-00003-of-00014.bin",
40
+ "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00003-of-00014.bin",
41
+ "model.layers.11.mlp.up_proj.weight": "pytorch_model-00003-of-00014.bin",
42
+ "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00003-of-00014.bin",
43
+ "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00003-of-00014.bin",
44
+ "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00003-of-00014.bin",
45
+ "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00003-of-00014.bin",
46
+ "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00014.bin",
47
+ "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00003-of-00014.bin",
48
+ "model.layers.12.input_layernorm.weight": "pytorch_model-00003-of-00014.bin",
49
+ "model.layers.12.mlp.down_proj.weight": "pytorch_model-00003-of-00014.bin",
50
+ "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00003-of-00014.bin",
51
+ "model.layers.12.mlp.up_proj.weight": "pytorch_model-00003-of-00014.bin",
52
+ "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00003-of-00014.bin",
53
+ "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00003-of-00014.bin",
54
+ "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00003-of-00014.bin",
55
+ "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00003-of-00014.bin",
56
+ "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00014.bin",
57
+ "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00003-of-00014.bin",
58
+ "model.layers.13.input_layernorm.weight": "pytorch_model-00004-of-00014.bin",
59
+ "model.layers.13.mlp.down_proj.weight": "pytorch_model-00004-of-00014.bin",
60
+ "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00004-of-00014.bin",
61
+ "model.layers.13.mlp.up_proj.weight": "pytorch_model-00004-of-00014.bin",
62
+ "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00004-of-00014.bin",
63
+ "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00003-of-00014.bin",
64
+ "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00003-of-00014.bin",
65
+ "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00003-of-00014.bin",
66
+ "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00014.bin",
67
+ "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00003-of-00014.bin",
68
+ "model.layers.14.input_layernorm.weight": "pytorch_model-00004-of-00014.bin",
69
+ "model.layers.14.mlp.down_proj.weight": "pytorch_model-00004-of-00014.bin",
70
+ "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00004-of-00014.bin",
71
+ "model.layers.14.mlp.up_proj.weight": "pytorch_model-00004-of-00014.bin",
72
+ "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00004-of-00014.bin",
73
+ "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00004-of-00014.bin",
74
+ "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00004-of-00014.bin",
75
+ "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00004-of-00014.bin",
76
+ "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00014.bin",
77
+ "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00004-of-00014.bin",
78
+ "model.layers.15.input_layernorm.weight": "pytorch_model-00004-of-00014.bin",
79
+ "model.layers.15.mlp.down_proj.weight": "pytorch_model-00004-of-00014.bin",
80
+ "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00004-of-00014.bin",
81
+ "model.layers.15.mlp.up_proj.weight": "pytorch_model-00004-of-00014.bin",
82
+ "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00004-of-00014.bin",
83
+ "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00004-of-00014.bin",
84
+ "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00004-of-00014.bin",
85
+ "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00004-of-00014.bin",
86
+ "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00014.bin",
87
+ "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00004-of-00014.bin",
88
+ "model.layers.16.input_layernorm.weight": "pytorch_model-00004-of-00014.bin",
89
+ "model.layers.16.mlp.down_proj.weight": "pytorch_model-00004-of-00014.bin",
90
+ "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00004-of-00014.bin",
91
+ "model.layers.16.mlp.up_proj.weight": "pytorch_model-00004-of-00014.bin",
92
+ "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00004-of-00014.bin",
93
+ "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00004-of-00014.bin",
94
+ "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00004-of-00014.bin",
95
+ "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00004-of-00014.bin",
96
+ "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00014.bin",
97
+ "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00004-of-00014.bin",
98
+ "model.layers.17.input_layernorm.weight": "pytorch_model-00004-of-00014.bin",
99
+ "model.layers.17.mlp.down_proj.weight": "pytorch_model-00004-of-00014.bin",
100
+ "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00004-of-00014.bin",
101
+ "model.layers.17.mlp.up_proj.weight": "pytorch_model-00004-of-00014.bin",
102
+ "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00004-of-00014.bin",
103
+ "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00004-of-00014.bin",
104
+ "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00004-of-00014.bin",
105
+ "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00004-of-00014.bin",
106
+ "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00014.bin",
107
+ "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00004-of-00014.bin",
108
+ "model.layers.18.input_layernorm.weight": "pytorch_model-00005-of-00014.bin",
109
+ "model.layers.18.mlp.down_proj.weight": "pytorch_model-00005-of-00014.bin",
110
+ "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00005-of-00014.bin",
111
+ "model.layers.18.mlp.up_proj.weight": "pytorch_model-00005-of-00014.bin",
112
+ "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00005-of-00014.bin",
113
+ "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00005-of-00014.bin",
114
+ "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00005-of-00014.bin",
115
+ "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00005-of-00014.bin",
116
+ "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00014.bin",
117
+ "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00005-of-00014.bin",
118
+ "model.layers.19.input_layernorm.weight": "pytorch_model-00005-of-00014.bin",
119
+ "model.layers.19.mlp.down_proj.weight": "pytorch_model-00005-of-00014.bin",
120
+ "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00005-of-00014.bin",
121
+ "model.layers.19.mlp.up_proj.weight": "pytorch_model-00005-of-00014.bin",
122
+ "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00005-of-00014.bin",
123
+ "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00005-of-00014.bin",
124
+ "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00005-of-00014.bin",
125
+ "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00005-of-00014.bin",
126
+ "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00014.bin",
127
+ "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00005-of-00014.bin",
128
+ "model.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00014.bin",
129
+ "model.layers.2.mlp.down_proj.weight": "pytorch_model-00001-of-00014.bin",
130
+ "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00014.bin",
131
+ "model.layers.2.mlp.up_proj.weight": "pytorch_model-00001-of-00014.bin",
132
+ "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00014.bin",
133
+ "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00014.bin",
134
+ "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00014.bin",
135
+ "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00014.bin",
136
+ "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00014.bin",
137
+ "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00014.bin",
138
+ "model.layers.20.input_layernorm.weight": "pytorch_model-00005-of-00014.bin",
139
+ "model.layers.20.mlp.down_proj.weight": "pytorch_model-00005-of-00014.bin",
140
+ "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00005-of-00014.bin",
141
+ "model.layers.20.mlp.up_proj.weight": "pytorch_model-00005-of-00014.bin",
142
+ "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00005-of-00014.bin",
143
+ "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00005-of-00014.bin",
144
+ "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00005-of-00014.bin",
145
+ "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00005-of-00014.bin",
146
+ "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00014.bin",
147
+ "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00005-of-00014.bin",
148
+ "model.layers.21.input_layernorm.weight": "pytorch_model-00005-of-00014.bin",
149
+ "model.layers.21.mlp.down_proj.weight": "pytorch_model-00005-of-00014.bin",
150
+ "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00005-of-00014.bin",
151
+ "model.layers.21.mlp.up_proj.weight": "pytorch_model-00005-of-00014.bin",
152
+ "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00005-of-00014.bin",
153
+ "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00005-of-00014.bin",
154
+ "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00005-of-00014.bin",
155
+ "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00005-of-00014.bin",
156
+ "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00014.bin",
157
+ "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00005-of-00014.bin",
158
+ "model.layers.22.input_layernorm.weight": "pytorch_model-00006-of-00014.bin",
159
+ "model.layers.22.mlp.down_proj.weight": "pytorch_model-00006-of-00014.bin",
160
+ "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00005-of-00014.bin",
161
+ "model.layers.22.mlp.up_proj.weight": "pytorch_model-00006-of-00014.bin",
162
+ "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00006-of-00014.bin",
163
+ "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00005-of-00014.bin",
164
+ "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00005-of-00014.bin",
165
+ "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00005-of-00014.bin",
166
+ "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00014.bin",
167
+ "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00005-of-00014.bin",
168
+ "model.layers.23.input_layernorm.weight": "pytorch_model-00006-of-00014.bin",
169
+ "model.layers.23.mlp.down_proj.weight": "pytorch_model-00006-of-00014.bin",
170
+ "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00006-of-00014.bin",
171
+ "model.layers.23.mlp.up_proj.weight": "pytorch_model-00006-of-00014.bin",
172
+ "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00006-of-00014.bin",
173
+ "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00006-of-00014.bin",
174
+ "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00006-of-00014.bin",
175
+ "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00006-of-00014.bin",
176
+ "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00014.bin",
177
+ "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00006-of-00014.bin",
178
+ "model.layers.24.input_layernorm.weight": "pytorch_model-00006-of-00014.bin",
179
+ "model.layers.24.mlp.down_proj.weight": "pytorch_model-00006-of-00014.bin",
180
+ "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00006-of-00014.bin",
181
+ "model.layers.24.mlp.up_proj.weight": "pytorch_model-00006-of-00014.bin",
182
+ "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00006-of-00014.bin",
183
+ "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00006-of-00014.bin",
184
+ "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00006-of-00014.bin",
185
+ "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00006-of-00014.bin",
186
+ "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00014.bin",
187
+ "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00006-of-00014.bin",
188
+ "model.layers.25.input_layernorm.weight": "pytorch_model-00006-of-00014.bin",
189
+ "model.layers.25.mlp.down_proj.weight": "pytorch_model-00006-of-00014.bin",
190
+ "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00006-of-00014.bin",
191
+ "model.layers.25.mlp.up_proj.weight": "pytorch_model-00006-of-00014.bin",
192
+ "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00006-of-00014.bin",
193
+ "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00006-of-00014.bin",
194
+ "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00006-of-00014.bin",
195
+ "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00006-of-00014.bin",
196
+ "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00014.bin",
197
+ "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00006-of-00014.bin",
198
+ "model.layers.26.input_layernorm.weight": "pytorch_model-00006-of-00014.bin",
199
+ "model.layers.26.mlp.down_proj.weight": "pytorch_model-00006-of-00014.bin",
200
+ "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00006-of-00014.bin",
201
+ "model.layers.26.mlp.up_proj.weight": "pytorch_model-00006-of-00014.bin",
202
+ "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00006-of-00014.bin",
203
+ "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00006-of-00014.bin",
204
+ "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00006-of-00014.bin",
205
+ "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00006-of-00014.bin",
206
+ "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00014.bin",
207
+ "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00006-of-00014.bin",
208
+ "model.layers.27.input_layernorm.weight": "pytorch_model-00007-of-00014.bin",
209
+ "model.layers.27.mlp.down_proj.weight": "pytorch_model-00007-of-00014.bin",
210
+ "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00007-of-00014.bin",
211
+ "model.layers.27.mlp.up_proj.weight": "pytorch_model-00007-of-00014.bin",
212
+ "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00007-of-00014.bin",
213
+ "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00006-of-00014.bin",
214
+ "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00007-of-00014.bin",
215
+ "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00006-of-00014.bin",
216
+ "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00014.bin",
217
+ "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00007-of-00014.bin",
218
+ "model.layers.28.input_layernorm.weight": "pytorch_model-00007-of-00014.bin",
219
+ "model.layers.28.mlp.down_proj.weight": "pytorch_model-00007-of-00014.bin",
220
+ "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00007-of-00014.bin",
221
+ "model.layers.28.mlp.up_proj.weight": "pytorch_model-00007-of-00014.bin",
222
+ "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00007-of-00014.bin",
223
+ "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00007-of-00014.bin",
224
+ "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00007-of-00014.bin",
225
+ "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00007-of-00014.bin",
226
+ "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00014.bin",
227
+ "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00007-of-00014.bin",
228
+ "model.layers.29.input_layernorm.weight": "pytorch_model-00007-of-00014.bin",
229
+ "model.layers.29.mlp.down_proj.weight": "pytorch_model-00007-of-00014.bin",
230
+ "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00007-of-00014.bin",
231
+ "model.layers.29.mlp.up_proj.weight": "pytorch_model-00007-of-00014.bin",
232
+ "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00007-of-00014.bin",
233
+ "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00007-of-00014.bin",
234
+ "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00007-of-00014.bin",
235
+ "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00007-of-00014.bin",
236
+ "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00014.bin",
237
+ "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00007-of-00014.bin",
238
+ "model.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00014.bin",
239
+ "model.layers.3.mlp.down_proj.weight": "pytorch_model-00001-of-00014.bin",
240
+ "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00001-of-00014.bin",
241
+ "model.layers.3.mlp.up_proj.weight": "pytorch_model-00001-of-00014.bin",
242
+ "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00014.bin",
243
+ "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00001-of-00014.bin",
244
+ "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00001-of-00014.bin",
245
+ "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00001-of-00014.bin",
246
+ "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00014.bin",
247
+ "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00001-of-00014.bin",
248
+ "model.layers.30.input_layernorm.weight": "pytorch_model-00007-of-00014.bin",
249
+ "model.layers.30.mlp.down_proj.weight": "pytorch_model-00007-of-00014.bin",
250
+ "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00007-of-00014.bin",
251
+ "model.layers.30.mlp.up_proj.weight": "pytorch_model-00007-of-00014.bin",
252
+ "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00007-of-00014.bin",
253
+ "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00007-of-00014.bin",
254
+ "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00007-of-00014.bin",
255
+ "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00007-of-00014.bin",
256
+ "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00014.bin",
257
+ "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00007-of-00014.bin",
258
+ "model.layers.31.input_layernorm.weight": "pytorch_model-00008-of-00014.bin",
259
+ "model.layers.31.mlp.down_proj.weight": "pytorch_model-00007-of-00014.bin",
260
+ "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00007-of-00014.bin",
261
+ "model.layers.31.mlp.up_proj.weight": "pytorch_model-00008-of-00014.bin",
262
+ "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00008-of-00014.bin",
263
+ "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00007-of-00014.bin",
264
+ "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00007-of-00014.bin",
265
+ "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00007-of-00014.bin",
266
+ "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00014.bin",
267
+ "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00007-of-00014.bin",
268
+ "model.layers.32.input_layernorm.weight": "pytorch_model-00008-of-00014.bin",
269
+ "model.layers.32.mlp.down_proj.weight": "pytorch_model-00008-of-00014.bin",
270
+ "model.layers.32.mlp.gate_proj.weight": "pytorch_model-00008-of-00014.bin",
271
+ "model.layers.32.mlp.up_proj.weight": "pytorch_model-00008-of-00014.bin",
272
+ "model.layers.32.post_attention_layernorm.weight": "pytorch_model-00008-of-00014.bin",
273
+ "model.layers.32.self_attn.k_proj.weight": "pytorch_model-00008-of-00014.bin",
274
+ "model.layers.32.self_attn.o_proj.weight": "pytorch_model-00008-of-00014.bin",
275
+ "model.layers.32.self_attn.q_proj.weight": "pytorch_model-00008-of-00014.bin",
276
+ "model.layers.32.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00014.bin",
277
+ "model.layers.32.self_attn.v_proj.weight": "pytorch_model-00008-of-00014.bin",
278
+ "model.layers.33.input_layernorm.weight": "pytorch_model-00008-of-00014.bin",
279
+ "model.layers.33.mlp.down_proj.weight": "pytorch_model-00008-of-00014.bin",
280
+ "model.layers.33.mlp.gate_proj.weight": "pytorch_model-00008-of-00014.bin",
281
+ "model.layers.33.mlp.up_proj.weight": "pytorch_model-00008-of-00014.bin",
282
+ "model.layers.33.post_attention_layernorm.weight": "pytorch_model-00008-of-00014.bin",
283
+ "model.layers.33.self_attn.k_proj.weight": "pytorch_model-00008-of-00014.bin",
284
+ "model.layers.33.self_attn.o_proj.weight": "pytorch_model-00008-of-00014.bin",
285
+ "model.layers.33.self_attn.q_proj.weight": "pytorch_model-00008-of-00014.bin",
286
+ "model.layers.33.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00014.bin",
287
+ "model.layers.33.self_attn.v_proj.weight": "pytorch_model-00008-of-00014.bin",
288
+ "model.layers.34.input_layernorm.weight": "pytorch_model-00008-of-00014.bin",
289
+ "model.layers.34.mlp.down_proj.weight": "pytorch_model-00008-of-00014.bin",
290
+ "model.layers.34.mlp.gate_proj.weight": "pytorch_model-00008-of-00014.bin",
291
+ "model.layers.34.mlp.up_proj.weight": "pytorch_model-00008-of-00014.bin",
292
+ "model.layers.34.post_attention_layernorm.weight": "pytorch_model-00008-of-00014.bin",
293
+ "model.layers.34.self_attn.k_proj.weight": "pytorch_model-00008-of-00014.bin",
294
+ "model.layers.34.self_attn.o_proj.weight": "pytorch_model-00008-of-00014.bin",
295
+ "model.layers.34.self_attn.q_proj.weight": "pytorch_model-00008-of-00014.bin",
296
+ "model.layers.34.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00014.bin",
297
+ "model.layers.34.self_attn.v_proj.weight": "pytorch_model-00008-of-00014.bin",
298
+ "model.layers.35.input_layernorm.weight": "pytorch_model-00008-of-00014.bin",
299
+ "model.layers.35.mlp.down_proj.weight": "pytorch_model-00008-of-00014.bin",
300
+ "model.layers.35.mlp.gate_proj.weight": "pytorch_model-00008-of-00014.bin",
301
+ "model.layers.35.mlp.up_proj.weight": "pytorch_model-00008-of-00014.bin",
302
+ "model.layers.35.post_attention_layernorm.weight": "pytorch_model-00008-of-00014.bin",
303
+ "model.layers.35.self_attn.k_proj.weight": "pytorch_model-00008-of-00014.bin",
304
+ "model.layers.35.self_attn.o_proj.weight": "pytorch_model-00008-of-00014.bin",
305
+ "model.layers.35.self_attn.q_proj.weight": "pytorch_model-00008-of-00014.bin",
306
+ "model.layers.35.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00014.bin",
307
+ "model.layers.35.self_attn.v_proj.weight": "pytorch_model-00008-of-00014.bin",
308
+ "model.layers.36.input_layernorm.weight": "pytorch_model-00009-of-00014.bin",
309
+ "model.layers.36.mlp.down_proj.weight": "pytorch_model-00009-of-00014.bin",
310
+ "model.layers.36.mlp.gate_proj.weight": "pytorch_model-00009-of-00014.bin",
311
+ "model.layers.36.mlp.up_proj.weight": "pytorch_model-00009-of-00014.bin",
312
+ "model.layers.36.post_attention_layernorm.weight": "pytorch_model-00009-of-00014.bin",
313
+ "model.layers.36.self_attn.k_proj.weight": "pytorch_model-00008-of-00014.bin",
314
+ "model.layers.36.self_attn.o_proj.weight": "pytorch_model-00008-of-00014.bin",
315
+ "model.layers.36.self_attn.q_proj.weight": "pytorch_model-00008-of-00014.bin",
316
+ "model.layers.36.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00014.bin",
317
+ "model.layers.36.self_attn.v_proj.weight": "pytorch_model-00008-of-00014.bin",
318
+ "model.layers.37.input_layernorm.weight": "pytorch_model-00009-of-00014.bin",
319
+ "model.layers.37.mlp.down_proj.weight": "pytorch_model-00009-of-00014.bin",
320
+ "model.layers.37.mlp.gate_proj.weight": "pytorch_model-00009-of-00014.bin",
321
+ "model.layers.37.mlp.up_proj.weight": "pytorch_model-00009-of-00014.bin",
322
+ "model.layers.37.post_attention_layernorm.weight": "pytorch_model-00009-of-00014.bin",
323
+ "model.layers.37.self_attn.k_proj.weight": "pytorch_model-00009-of-00014.bin",
324
+ "model.layers.37.self_attn.o_proj.weight": "pytorch_model-00009-of-00014.bin",
325
+ "model.layers.37.self_attn.q_proj.weight": "pytorch_model-00009-of-00014.bin",
326
+ "model.layers.37.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00014.bin",
327
+ "model.layers.37.self_attn.v_proj.weight": "pytorch_model-00009-of-00014.bin",
328
+ "model.layers.38.input_layernorm.weight": "pytorch_model-00009-of-00014.bin",
329
+ "model.layers.38.mlp.down_proj.weight": "pytorch_model-00009-of-00014.bin",
330
+ "model.layers.38.mlp.gate_proj.weight": "pytorch_model-00009-of-00014.bin",
331
+ "model.layers.38.mlp.up_proj.weight": "pytorch_model-00009-of-00014.bin",
332
+ "model.layers.38.post_attention_layernorm.weight": "pytorch_model-00009-of-00014.bin",
333
+ "model.layers.38.self_attn.k_proj.weight": "pytorch_model-00009-of-00014.bin",
334
+ "model.layers.38.self_attn.o_proj.weight": "pytorch_model-00009-of-00014.bin",
335
+ "model.layers.38.self_attn.q_proj.weight": "pytorch_model-00009-of-00014.bin",
336
+ "model.layers.38.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00014.bin",
337
+ "model.layers.38.self_attn.v_proj.weight": "pytorch_model-00009-of-00014.bin",
338
+ "model.layers.39.input_layernorm.weight": "pytorch_model-00009-of-00014.bin",
339
+ "model.layers.39.mlp.down_proj.weight": "pytorch_model-00009-of-00014.bin",
340
+ "model.layers.39.mlp.gate_proj.weight": "pytorch_model-00009-of-00014.bin",
341
+ "model.layers.39.mlp.up_proj.weight": "pytorch_model-00009-of-00014.bin",
342
+ "model.layers.39.post_attention_layernorm.weight": "pytorch_model-00009-of-00014.bin",
343
+ "model.layers.39.self_attn.k_proj.weight": "pytorch_model-00009-of-00014.bin",
344
+ "model.layers.39.self_attn.o_proj.weight": "pytorch_model-00009-of-00014.bin",
345
+ "model.layers.39.self_attn.q_proj.weight": "pytorch_model-00009-of-00014.bin",
346
+ "model.layers.39.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00014.bin",
347
+ "model.layers.39.self_attn.v_proj.weight": "pytorch_model-00009-of-00014.bin",
348
+ "model.layers.4.input_layernorm.weight": "pytorch_model-00002-of-00014.bin",
349
+ "model.layers.4.mlp.down_proj.weight": "pytorch_model-00002-of-00014.bin",
350
+ "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00002-of-00014.bin",
351
+ "model.layers.4.mlp.up_proj.weight": "pytorch_model-00002-of-00014.bin",
352
+ "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00002-of-00014.bin",
353
+ "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00001-of-00014.bin",
354
+ "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00002-of-00014.bin",
355
+ "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00001-of-00014.bin",
356
+ "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00014.bin",
357
+ "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00001-of-00014.bin",
358
+ "model.layers.40.input_layernorm.weight": "pytorch_model-00009-of-00014.bin",
359
+ "model.layers.40.mlp.down_proj.weight": "pytorch_model-00009-of-00014.bin",
360
+ "model.layers.40.mlp.gate_proj.weight": "pytorch_model-00009-of-00014.bin",
361
+ "model.layers.40.mlp.up_proj.weight": "pytorch_model-00009-of-00014.bin",
362
+ "model.layers.40.post_attention_layernorm.weight": "pytorch_model-00009-of-00014.bin",
363
+ "model.layers.40.self_attn.k_proj.weight": "pytorch_model-00009-of-00014.bin",
364
+ "model.layers.40.self_attn.o_proj.weight": "pytorch_model-00009-of-00014.bin",
365
+ "model.layers.40.self_attn.q_proj.weight": "pytorch_model-00009-of-00014.bin",
366
+ "model.layers.40.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00014.bin",
367
+ "model.layers.40.self_attn.v_proj.weight": "pytorch_model-00009-of-00014.bin",
368
+ "model.layers.41.input_layernorm.weight": "pytorch_model-00010-of-00014.bin",
369
+ "model.layers.41.mlp.down_proj.weight": "pytorch_model-00010-of-00014.bin",
370
+ "model.layers.41.mlp.gate_proj.weight": "pytorch_model-00010-of-00014.bin",
371
+ "model.layers.41.mlp.up_proj.weight": "pytorch_model-00010-of-00014.bin",
372
+ "model.layers.41.post_attention_layernorm.weight": "pytorch_model-00010-of-00014.bin",
373
+ "model.layers.41.self_attn.k_proj.weight": "pytorch_model-00010-of-00014.bin",
374
+ "model.layers.41.self_attn.o_proj.weight": "pytorch_model-00010-of-00014.bin",
375
+ "model.layers.41.self_attn.q_proj.weight": "pytorch_model-00010-of-00014.bin",
376
+ "model.layers.41.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00014.bin",
377
+ "model.layers.41.self_attn.v_proj.weight": "pytorch_model-00010-of-00014.bin",
378
+ "model.layers.42.input_layernorm.weight": "pytorch_model-00010-of-00014.bin",
379
+ "model.layers.42.mlp.down_proj.weight": "pytorch_model-00010-of-00014.bin",
380
+ "model.layers.42.mlp.gate_proj.weight": "pytorch_model-00010-of-00014.bin",
381
+ "model.layers.42.mlp.up_proj.weight": "pytorch_model-00010-of-00014.bin",
382
+ "model.layers.42.post_attention_layernorm.weight": "pytorch_model-00010-of-00014.bin",
383
+ "model.layers.42.self_attn.k_proj.weight": "pytorch_model-00010-of-00014.bin",
384
+ "model.layers.42.self_attn.o_proj.weight": "pytorch_model-00010-of-00014.bin",
385
+ "model.layers.42.self_attn.q_proj.weight": "pytorch_model-00010-of-00014.bin",
386
+ "model.layers.42.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00014.bin",
387
+ "model.layers.42.self_attn.v_proj.weight": "pytorch_model-00010-of-00014.bin",
388
+ "model.layers.43.input_layernorm.weight": "pytorch_model-00010-of-00014.bin",
389
+ "model.layers.43.mlp.down_proj.weight": "pytorch_model-00010-of-00014.bin",
390
+ "model.layers.43.mlp.gate_proj.weight": "pytorch_model-00010-of-00014.bin",
391
+ "model.layers.43.mlp.up_proj.weight": "pytorch_model-00010-of-00014.bin",
392
+ "model.layers.43.post_attention_layernorm.weight": "pytorch_model-00010-of-00014.bin",
393
+ "model.layers.43.self_attn.k_proj.weight": "pytorch_model-00010-of-00014.bin",
394
+ "model.layers.43.self_attn.o_proj.weight": "pytorch_model-00010-of-00014.bin",
395
+ "model.layers.43.self_attn.q_proj.weight": "pytorch_model-00010-of-00014.bin",
396
+ "model.layers.43.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00014.bin",
397
+ "model.layers.43.self_attn.v_proj.weight": "pytorch_model-00010-of-00014.bin",
398
+ "model.layers.44.input_layernorm.weight": "pytorch_model-00010-of-00014.bin",
399
+ "model.layers.44.mlp.down_proj.weight": "pytorch_model-00010-of-00014.bin",
400
+ "model.layers.44.mlp.gate_proj.weight": "pytorch_model-00010-of-00014.bin",
401
+ "model.layers.44.mlp.up_proj.weight": "pytorch_model-00010-of-00014.bin",
402
+ "model.layers.44.post_attention_layernorm.weight": "pytorch_model-00010-of-00014.bin",
403
+ "model.layers.44.self_attn.k_proj.weight": "pytorch_model-00010-of-00014.bin",
404
+ "model.layers.44.self_attn.o_proj.weight": "pytorch_model-00010-of-00014.bin",
405
+ "model.layers.44.self_attn.q_proj.weight": "pytorch_model-00010-of-00014.bin",
406
+ "model.layers.44.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00014.bin",
407
+ "model.layers.44.self_attn.v_proj.weight": "pytorch_model-00010-of-00014.bin",
408
+ "model.layers.45.input_layernorm.weight": "pytorch_model-00011-of-00014.bin",
409
+ "model.layers.45.mlp.down_proj.weight": "pytorch_model-00011-of-00014.bin",
410
+ "model.layers.45.mlp.gate_proj.weight": "pytorch_model-00010-of-00014.bin",
411
+ "model.layers.45.mlp.up_proj.weight": "pytorch_model-00011-of-00014.bin",
412
+ "model.layers.45.post_attention_layernorm.weight": "pytorch_model-00011-of-00014.bin",
413
+ "model.layers.45.self_attn.k_proj.weight": "pytorch_model-00010-of-00014.bin",
414
+ "model.layers.45.self_attn.o_proj.weight": "pytorch_model-00010-of-00014.bin",
415
+ "model.layers.45.self_attn.q_proj.weight": "pytorch_model-00010-of-00014.bin",
416
+ "model.layers.45.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00014.bin",
417
+ "model.layers.45.self_attn.v_proj.weight": "pytorch_model-00010-of-00014.bin",
418
+ "model.layers.46.input_layernorm.weight": "pytorch_model-00011-of-00014.bin",
419
+ "model.layers.46.mlp.down_proj.weight": "pytorch_model-00011-of-00014.bin",
420
+ "model.layers.46.mlp.gate_proj.weight": "pytorch_model-00011-of-00014.bin",
421
+ "model.layers.46.mlp.up_proj.weight": "pytorch_model-00011-of-00014.bin",
422
+ "model.layers.46.post_attention_layernorm.weight": "pytorch_model-00011-of-00014.bin",
423
+ "model.layers.46.self_attn.k_proj.weight": "pytorch_model-00011-of-00014.bin",
424
+ "model.layers.46.self_attn.o_proj.weight": "pytorch_model-00011-of-00014.bin",
425
+ "model.layers.46.self_attn.q_proj.weight": "pytorch_model-00011-of-00014.bin",
426
+ "model.layers.46.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00014.bin",
427
+ "model.layers.46.self_attn.v_proj.weight": "pytorch_model-00011-of-00014.bin",
428
+ "model.layers.47.input_layernorm.weight": "pytorch_model-00011-of-00014.bin",
429
+ "model.layers.47.mlp.down_proj.weight": "pytorch_model-00011-of-00014.bin",
430
+ "model.layers.47.mlp.gate_proj.weight": "pytorch_model-00011-of-00014.bin",
431
+ "model.layers.47.mlp.up_proj.weight": "pytorch_model-00011-of-00014.bin",
432
+ "model.layers.47.post_attention_layernorm.weight": "pytorch_model-00011-of-00014.bin",
433
+ "model.layers.47.self_attn.k_proj.weight": "pytorch_model-00011-of-00014.bin",
434
+ "model.layers.47.self_attn.o_proj.weight": "pytorch_model-00011-of-00014.bin",
435
+ "model.layers.47.self_attn.q_proj.weight": "pytorch_model-00011-of-00014.bin",
436
+ "model.layers.47.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00014.bin",
437
+ "model.layers.47.self_attn.v_proj.weight": "pytorch_model-00011-of-00014.bin",
438
+ "model.layers.48.input_layernorm.weight": "pytorch_model-00011-of-00014.bin",
439
+ "model.layers.48.mlp.down_proj.weight": "pytorch_model-00011-of-00014.bin",
440
+ "model.layers.48.mlp.gate_proj.weight": "pytorch_model-00011-of-00014.bin",
441
+ "model.layers.48.mlp.up_proj.weight": "pytorch_model-00011-of-00014.bin",
442
+ "model.layers.48.post_attention_layernorm.weight": "pytorch_model-00011-of-00014.bin",
443
+ "model.layers.48.self_attn.k_proj.weight": "pytorch_model-00011-of-00014.bin",
444
+ "model.layers.48.self_attn.o_proj.weight": "pytorch_model-00011-of-00014.bin",
445
+ "model.layers.48.self_attn.q_proj.weight": "pytorch_model-00011-of-00014.bin",
446
+ "model.layers.48.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00014.bin",
447
+ "model.layers.48.self_attn.v_proj.weight": "pytorch_model-00011-of-00014.bin",
448
+ "model.layers.49.input_layernorm.weight": "pytorch_model-00011-of-00014.bin",
449
+ "model.layers.49.mlp.down_proj.weight": "pytorch_model-00011-of-00014.bin",
450
+ "model.layers.49.mlp.gate_proj.weight": "pytorch_model-00011-of-00014.bin",
451
+ "model.layers.49.mlp.up_proj.weight": "pytorch_model-00011-of-00014.bin",
452
+ "model.layers.49.post_attention_layernorm.weight": "pytorch_model-00011-of-00014.bin",
453
+ "model.layers.49.self_attn.k_proj.weight": "pytorch_model-00011-of-00014.bin",
454
+ "model.layers.49.self_attn.o_proj.weight": "pytorch_model-00011-of-00014.bin",
455
+ "model.layers.49.self_attn.q_proj.weight": "pytorch_model-00011-of-00014.bin",
456
+ "model.layers.49.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00014.bin",
457
+ "model.layers.49.self_attn.v_proj.weight": "pytorch_model-00011-of-00014.bin",
458
+ "model.layers.5.input_layernorm.weight": "pytorch_model-00002-of-00014.bin",
459
+ "model.layers.5.mlp.down_proj.weight": "pytorch_model-00002-of-00014.bin",
460
+ "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00002-of-00014.bin",
461
+ "model.layers.5.mlp.up_proj.weight": "pytorch_model-00002-of-00014.bin",
462
+ "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00002-of-00014.bin",
463
+ "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00002-of-00014.bin",
464
+ "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00002-of-00014.bin",
465
+ "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00002-of-00014.bin",
466
+ "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00014.bin",
467
+ "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00002-of-00014.bin",
468
+ "model.layers.50.input_layernorm.weight": "pytorch_model-00012-of-00014.bin",
469
+ "model.layers.50.mlp.down_proj.weight": "pytorch_model-00012-of-00014.bin",
470
+ "model.layers.50.mlp.gate_proj.weight": "pytorch_model-00012-of-00014.bin",
471
+ "model.layers.50.mlp.up_proj.weight": "pytorch_model-00012-of-00014.bin",
472
+ "model.layers.50.post_attention_layernorm.weight": "pytorch_model-00012-of-00014.bin",
473
+ "model.layers.50.self_attn.k_proj.weight": "pytorch_model-00011-of-00014.bin",
474
+ "model.layers.50.self_attn.o_proj.weight": "pytorch_model-00012-of-00014.bin",
475
+ "model.layers.50.self_attn.q_proj.weight": "pytorch_model-00011-of-00014.bin",
476
+ "model.layers.50.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00014.bin",
477
+ "model.layers.50.self_attn.v_proj.weight": "pytorch_model-00012-of-00014.bin",
478
+ "model.layers.51.input_layernorm.weight": "pytorch_model-00012-of-00014.bin",
479
+ "model.layers.51.mlp.down_proj.weight": "pytorch_model-00012-of-00014.bin",
480
+ "model.layers.51.mlp.gate_proj.weight": "pytorch_model-00012-of-00014.bin",
481
+ "model.layers.51.mlp.up_proj.weight": "pytorch_model-00012-of-00014.bin",
482
+ "model.layers.51.post_attention_layernorm.weight": "pytorch_model-00012-of-00014.bin",
483
+ "model.layers.51.self_attn.k_proj.weight": "pytorch_model-00012-of-00014.bin",
484
+ "model.layers.51.self_attn.o_proj.weight": "pytorch_model-00012-of-00014.bin",
485
+ "model.layers.51.self_attn.q_proj.weight": "pytorch_model-00012-of-00014.bin",
486
+ "model.layers.51.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00014.bin",
487
+ "model.layers.51.self_attn.v_proj.weight": "pytorch_model-00012-of-00014.bin",
488
+ "model.layers.52.input_layernorm.weight": "pytorch_model-00012-of-00014.bin",
489
+ "model.layers.52.mlp.down_proj.weight": "pytorch_model-00012-of-00014.bin",
490
+ "model.layers.52.mlp.gate_proj.weight": "pytorch_model-00012-of-00014.bin",
491
+ "model.layers.52.mlp.up_proj.weight": "pytorch_model-00012-of-00014.bin",
492
+ "model.layers.52.post_attention_layernorm.weight": "pytorch_model-00012-of-00014.bin",
493
+ "model.layers.52.self_attn.k_proj.weight": "pytorch_model-00012-of-00014.bin",
494
+ "model.layers.52.self_attn.o_proj.weight": "pytorch_model-00012-of-00014.bin",
495
+ "model.layers.52.self_attn.q_proj.weight": "pytorch_model-00012-of-00014.bin",
496
+ "model.layers.52.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00014.bin",
497
+ "model.layers.52.self_attn.v_proj.weight": "pytorch_model-00012-of-00014.bin",
498
+ "model.layers.53.input_layernorm.weight": "pytorch_model-00012-of-00014.bin",
499
+ "model.layers.53.mlp.down_proj.weight": "pytorch_model-00012-of-00014.bin",
500
+ "model.layers.53.mlp.gate_proj.weight": "pytorch_model-00012-of-00014.bin",
501
+ "model.layers.53.mlp.up_proj.weight": "pytorch_model-00012-of-00014.bin",
502
+ "model.layers.53.post_attention_layernorm.weight": "pytorch_model-00012-of-00014.bin",
503
+ "model.layers.53.self_attn.k_proj.weight": "pytorch_model-00012-of-00014.bin",
504
+ "model.layers.53.self_attn.o_proj.weight": "pytorch_model-00012-of-00014.bin",
505
+ "model.layers.53.self_attn.q_proj.weight": "pytorch_model-00012-of-00014.bin",
506
+ "model.layers.53.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00014.bin",
507
+ "model.layers.53.self_attn.v_proj.weight": "pytorch_model-00012-of-00014.bin",
508
+ "model.layers.54.input_layernorm.weight": "pytorch_model-00013-of-00014.bin",
509
+ "model.layers.54.mlp.down_proj.weight": "pytorch_model-00012-of-00014.bin",
510
+ "model.layers.54.mlp.gate_proj.weight": "pytorch_model-00012-of-00014.bin",
511
+ "model.layers.54.mlp.up_proj.weight": "pytorch_model-00013-of-00014.bin",
512
+ "model.layers.54.post_attention_layernorm.weight": "pytorch_model-00013-of-00014.bin",
513
+ "model.layers.54.self_attn.k_proj.weight": "pytorch_model-00012-of-00014.bin",
514
+ "model.layers.54.self_attn.o_proj.weight": "pytorch_model-00012-of-00014.bin",
515
+ "model.layers.54.self_attn.q_proj.weight": "pytorch_model-00012-of-00014.bin",
516
+ "model.layers.54.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00014.bin",
517
+ "model.layers.54.self_attn.v_proj.weight": "pytorch_model-00012-of-00014.bin",
518
+ "model.layers.55.input_layernorm.weight": "pytorch_model-00013-of-00014.bin",
519
+ "model.layers.55.mlp.down_proj.weight": "pytorch_model-00013-of-00014.bin",
520
+ "model.layers.55.mlp.gate_proj.weight": "pytorch_model-00013-of-00014.bin",
521
+ "model.layers.55.mlp.up_proj.weight": "pytorch_model-00013-of-00014.bin",
522
+ "model.layers.55.post_attention_layernorm.weight": "pytorch_model-00013-of-00014.bin",
523
+ "model.layers.55.self_attn.k_proj.weight": "pytorch_model-00013-of-00014.bin",
524
+ "model.layers.55.self_attn.o_proj.weight": "pytorch_model-00013-of-00014.bin",
525
+ "model.layers.55.self_attn.q_proj.weight": "pytorch_model-00013-of-00014.bin",
526
+ "model.layers.55.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00014.bin",
527
+ "model.layers.55.self_attn.v_proj.weight": "pytorch_model-00013-of-00014.bin",
528
+ "model.layers.56.input_layernorm.weight": "pytorch_model-00013-of-00014.bin",
529
+ "model.layers.56.mlp.down_proj.weight": "pytorch_model-00013-of-00014.bin",
530
+ "model.layers.56.mlp.gate_proj.weight": "pytorch_model-00013-of-00014.bin",
531
+ "model.layers.56.mlp.up_proj.weight": "pytorch_model-00013-of-00014.bin",
532
+ "model.layers.56.post_attention_layernorm.weight": "pytorch_model-00013-of-00014.bin",
533
+ "model.layers.56.self_attn.k_proj.weight": "pytorch_model-00013-of-00014.bin",
534
+ "model.layers.56.self_attn.o_proj.weight": "pytorch_model-00013-of-00014.bin",
535
+ "model.layers.56.self_attn.q_proj.weight": "pytorch_model-00013-of-00014.bin",
536
+ "model.layers.56.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00014.bin",
537
+ "model.layers.56.self_attn.v_proj.weight": "pytorch_model-00013-of-00014.bin",
538
+ "model.layers.57.input_layernorm.weight": "pytorch_model-00013-of-00014.bin",
539
+ "model.layers.57.mlp.down_proj.weight": "pytorch_model-00013-of-00014.bin",
540
+ "model.layers.57.mlp.gate_proj.weight": "pytorch_model-00013-of-00014.bin",
541
+ "model.layers.57.mlp.up_proj.weight": "pytorch_model-00013-of-00014.bin",
542
+ "model.layers.57.post_attention_layernorm.weight": "pytorch_model-00013-of-00014.bin",
543
+ "model.layers.57.self_attn.k_proj.weight": "pytorch_model-00013-of-00014.bin",
544
+ "model.layers.57.self_attn.o_proj.weight": "pytorch_model-00013-of-00014.bin",
545
+ "model.layers.57.self_attn.q_proj.weight": "pytorch_model-00013-of-00014.bin",
546
+ "model.layers.57.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00014.bin",
547
+ "model.layers.57.self_attn.v_proj.weight": "pytorch_model-00013-of-00014.bin",
548
+ "model.layers.58.input_layernorm.weight": "pytorch_model-00013-of-00014.bin",
549
+ "model.layers.58.mlp.down_proj.weight": "pytorch_model-00013-of-00014.bin",
550
+ "model.layers.58.mlp.gate_proj.weight": "pytorch_model-00013-of-00014.bin",
551
+ "model.layers.58.mlp.up_proj.weight": "pytorch_model-00013-of-00014.bin",
552
+ "model.layers.58.post_attention_layernorm.weight": "pytorch_model-00013-of-00014.bin",
553
+ "model.layers.58.self_attn.k_proj.weight": "pytorch_model-00013-of-00014.bin",
554
+ "model.layers.58.self_attn.o_proj.weight": "pytorch_model-00013-of-00014.bin",
555
+ "model.layers.58.self_attn.q_proj.weight": "pytorch_model-00013-of-00014.bin",
556
+ "model.layers.58.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00014.bin",
557
+ "model.layers.58.self_attn.v_proj.weight": "pytorch_model-00013-of-00014.bin",
558
+ "model.layers.59.input_layernorm.weight": "pytorch_model-00014-of-00014.bin",
559
+ "model.layers.59.mlp.down_proj.weight": "pytorch_model-00014-of-00014.bin",
560
+ "model.layers.59.mlp.gate_proj.weight": "pytorch_model-00014-of-00014.bin",
561
+ "model.layers.59.mlp.up_proj.weight": "pytorch_model-00014-of-00014.bin",
562
+ "model.layers.59.post_attention_layernorm.weight": "pytorch_model-00014-of-00014.bin",
563
+ "model.layers.59.self_attn.k_proj.weight": "pytorch_model-00013-of-00014.bin",
564
+ "model.layers.59.self_attn.o_proj.weight": "pytorch_model-00013-of-00014.bin",
565
+ "model.layers.59.self_attn.q_proj.weight": "pytorch_model-00013-of-00014.bin",
566
+ "model.layers.59.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00014.bin",
567
+ "model.layers.59.self_attn.v_proj.weight": "pytorch_model-00013-of-00014.bin",
568
+ "model.layers.6.input_layernorm.weight": "pytorch_model-00002-of-00014.bin",
569
+ "model.layers.6.mlp.down_proj.weight": "pytorch_model-00002-of-00014.bin",
570
+ "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00002-of-00014.bin",
571
+ "model.layers.6.mlp.up_proj.weight": "pytorch_model-00002-of-00014.bin",
572
+ "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00002-of-00014.bin",
573
+ "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00002-of-00014.bin",
574
+ "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00002-of-00014.bin",
575
+ "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00002-of-00014.bin",
576
+ "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00014.bin",
577
+ "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00002-of-00014.bin",
578
+ "model.layers.7.input_layernorm.weight": "pytorch_model-00002-of-00014.bin",
579
+ "model.layers.7.mlp.down_proj.weight": "pytorch_model-00002-of-00014.bin",
580
+ "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00002-of-00014.bin",
581
+ "model.layers.7.mlp.up_proj.weight": "pytorch_model-00002-of-00014.bin",
582
+ "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00002-of-00014.bin",
583
+ "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00002-of-00014.bin",
584
+ "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00002-of-00014.bin",
585
+ "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00002-of-00014.bin",
586
+ "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00014.bin",
587
+ "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00002-of-00014.bin",
588
+ "model.layers.8.input_layernorm.weight": "pytorch_model-00003-of-00014.bin",
589
+ "model.layers.8.mlp.down_proj.weight": "pytorch_model-00002-of-00014.bin",
590
+ "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00002-of-00014.bin",
591
+ "model.layers.8.mlp.up_proj.weight": "pytorch_model-00003-of-00014.bin",
592
+ "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00003-of-00014.bin",
593
+ "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00002-of-00014.bin",
594
+ "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00002-of-00014.bin",
595
+ "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00002-of-00014.bin",
596
+ "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00014.bin",
597
+ "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00002-of-00014.bin",
598
+ "model.layers.9.input_layernorm.weight": "pytorch_model-00003-of-00014.bin",
599
+ "model.layers.9.mlp.down_proj.weight": "pytorch_model-00003-of-00014.bin",
600
+ "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00003-of-00014.bin",
601
+ "model.layers.9.mlp.up_proj.weight": "pytorch_model-00003-of-00014.bin",
602
+ "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00003-of-00014.bin",
603
+ "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00003-of-00014.bin",
604
+ "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00003-of-00014.bin",
605
+ "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00003-of-00014.bin",
606
+ "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00014.bin",
607
+ "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00003-of-00014.bin",
608
+ "model.norm.weight": "pytorch_model-00014-of-00014.bin"
609
+ }
610
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<unk>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "bos_token": {
5
+ "__type": "AddedToken",
6
+ "content": "<s>",
7
+ "lstrip": false,
8
+ "normalized": true,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "clean_up_tokenization_spaces": false,
13
+ "eos_token": {
14
+ "__type": "AddedToken",
15
+ "content": "</s>",
16
+ "lstrip": false,
17
+ "normalized": true,
18
+ "rstrip": false,
19
+ "single_word": false
20
+ },
21
+ "model_max_length": 2048,
22
+ "pad_token": null,
23
+ "padding_side": "right",
24
+ "sp_model_kwargs": {},
25
+ "tokenizer_class": "LlamaTokenizer",
26
+ "unk_token": {
27
+ "__type": "AddedToken",
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": true,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }
trainer_state.json ADDED
@@ -0,0 +1,826 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.967032967032967,
5
+ "global_step": 135,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 3.3333333333333333e-06,
13
+ "loss": 1.5703,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.04,
18
+ "learning_rate": 6.666666666666667e-06,
19
+ "loss": 1.5859,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.07,
24
+ "learning_rate": 1e-05,
25
+ "loss": 1.5566,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.09,
30
+ "learning_rate": 1.3333333333333333e-05,
31
+ "loss": 1.3877,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.11,
36
+ "learning_rate": 1.6666666666666667e-05,
37
+ "loss": 1.2617,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.13,
42
+ "learning_rate": 2e-05,
43
+ "loss": 1.1963,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.15,
48
+ "learning_rate": 1.9997034698451396e-05,
49
+ "loss": 1.1924,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.18,
54
+ "learning_rate": 1.998814055240823e-05,
55
+ "loss": 1.126,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.2,
60
+ "learning_rate": 1.9973322836635517e-05,
61
+ "loss": 1.123,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.22,
66
+ "learning_rate": 1.995259033893236e-05,
67
+ "loss": 1.0996,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.24,
72
+ "learning_rate": 1.9925955354920265e-05,
73
+ "loss": 1.0957,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.26,
78
+ "learning_rate": 1.9893433680751105e-05,
79
+ "loss": 1.0791,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.29,
84
+ "learning_rate": 1.985504460373903e-05,
85
+ "loss": 1.0791,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.31,
90
+ "learning_rate": 1.9810810890921943e-05,
91
+ "loss": 1.0762,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.33,
96
+ "learning_rate": 1.9760758775559275e-05,
97
+ "loss": 1.0347,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.35,
102
+ "learning_rate": 1.9704917941574053e-05,
103
+ "loss": 1.021,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.37,
108
+ "learning_rate": 1.9643321505948588e-05,
109
+ "loss": 1.0259,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.4,
114
+ "learning_rate": 1.957600599908406e-05,
115
+ "loss": 1.0317,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.42,
120
+ "learning_rate": 1.9503011343135828e-05,
121
+ "loss": 1.0078,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.44,
126
+ "learning_rate": 1.9424380828337146e-05,
127
+ "loss": 0.9849,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.46,
132
+ "learning_rate": 1.9340161087325483e-05,
133
+ "loss": 1.0127,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.48,
138
+ "learning_rate": 1.9250402067486523e-05,
139
+ "loss": 1.0103,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.51,
144
+ "learning_rate": 1.9155157001332374e-05,
145
+ "loss": 1.0024,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.53,
150
+ "learning_rate": 1.905448237493147e-05,
151
+ "loss": 0.9863,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.55,
156
+ "learning_rate": 1.894843789440892e-05,
157
+ "loss": 1.0117,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.57,
162
+ "learning_rate": 1.8837086450537195e-05,
163
+ "loss": 0.9775,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.59,
168
+ "learning_rate": 1.872049408143808e-05,
169
+ "loss": 0.9653,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.62,
174
+ "learning_rate": 1.8598729933418102e-05,
175
+ "loss": 0.9487,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.64,
180
+ "learning_rate": 1.8471866219960604e-05,
181
+ "loss": 0.9282,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.66,
186
+ "learning_rate": 1.833997817889878e-05,
187
+ "loss": 0.9424,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.68,
192
+ "learning_rate": 1.820314402779511e-05,
193
+ "loss": 0.9258,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.7,
198
+ "learning_rate": 1.806144491755363e-05,
199
+ "loss": 0.9292,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.73,
204
+ "learning_rate": 1.7914964884292543e-05,
205
+ "loss": 0.9473,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.75,
210
+ "learning_rate": 1.7763790799505746e-05,
211
+ "loss": 0.9453,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.77,
216
+ "learning_rate": 1.760801231854278e-05,
217
+ "loss": 0.9355,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.79,
222
+ "learning_rate": 1.744772182743782e-05,
223
+ "loss": 0.917,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.81,
228
+ "learning_rate": 1.728301438811916e-05,
229
+ "loss": 0.8838,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.84,
234
+ "learning_rate": 1.711398768203178e-05,
235
+ "loss": 0.897,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.86,
240
+ "learning_rate": 1.6940741952206342e-05,
241
+ "loss": 0.8926,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.88,
246
+ "learning_rate": 1.676337994380903e-05,
247
+ "loss": 0.8936,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.9,
252
+ "learning_rate": 1.658200684320748e-05,
253
+ "loss": 0.894,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.92,
258
+ "learning_rate": 1.6396730215588913e-05,
259
+ "loss": 0.8799,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.95,
264
+ "learning_rate": 1.6207659941167485e-05,
265
+ "loss": 0.8945,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.97,
270
+ "learning_rate": 1.6014908150018703e-05,
271
+ "loss": 0.8838,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.99,
276
+ "learning_rate": 1.581858915557953e-05,
277
+ "loss": 0.8745,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 1.01,
282
+ "learning_rate": 1.5618819386853607e-05,
283
+ "loss": 0.9004,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 1.03,
288
+ "learning_rate": 1.541571731936185e-05,
289
+ "loss": 0.7832,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 1.05,
294
+ "learning_rate": 1.5209403404879305e-05,
295
+ "loss": 0.8062,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 1.08,
300
+ "learning_rate": 1.5000000000000002e-05,
301
+ "loss": 0.8164,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 1.1,
306
+ "learning_rate": 1.4787631293572094e-05,
307
+ "loss": 0.812,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 1.12,
312
+ "learning_rate": 1.4572423233046386e-05,
313
+ "loss": 0.8169,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 1.14,
318
+ "learning_rate": 1.4354503449781914e-05,
319
+ "loss": 0.8101,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 1.16,
324
+ "learning_rate": 1.4134001183352833e-05,
325
+ "loss": 0.8052,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 1.19,
330
+ "learning_rate": 1.391104720490156e-05,
331
+ "loss": 0.7568,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 1.21,
336
+ "learning_rate": 1.368577373958362e-05,
337
+ "loss": 0.7974,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 1.23,
342
+ "learning_rate": 1.3458314388150115e-05,
343
+ "loss": 0.7812,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 1.25,
348
+ "learning_rate": 1.3228804047714462e-05,
349
+ "loss": 0.7603,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 1.27,
354
+ "learning_rate": 1.2997378831750242e-05,
355
+ "loss": 0.7432,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 1.3,
360
+ "learning_rate": 1.2764175989367717e-05,
361
+ "loss": 0.7617,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 1.32,
366
+ "learning_rate": 1.2529333823916807e-05,
367
+ "loss": 0.7676,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 1.34,
372
+ "learning_rate": 1.2292991610964902e-05,
373
+ "loss": 0.751,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 1.36,
378
+ "learning_rate": 1.2055289515698008e-05,
379
+ "loss": 0.7607,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 1.38,
384
+ "learning_rate": 1.1816368509794365e-05,
385
+ "loss": 0.7427,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 1.41,
390
+ "learning_rate": 1.1576370287819737e-05,
391
+ "loss": 0.772,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 1.43,
396
+ "learning_rate": 1.133543718319398e-05,
397
+ "loss": 0.7065,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 1.45,
402
+ "learning_rate": 1.1093712083778748e-05,
403
+ "loss": 0.75,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 1.47,
408
+ "learning_rate": 1.0851338347136358e-05,
409
+ "loss": 0.7437,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 1.49,
414
+ "learning_rate": 1.060845971551014e-05,
415
+ "loss": 0.6865,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 1.52,
420
+ "learning_rate": 1.0365220230576592e-05,
421
+ "loss": 0.7139,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 1.54,
426
+ "learning_rate": 1.0121764148019977e-05,
427
+ "loss": 0.7056,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 1.56,
432
+ "learning_rate": 9.878235851980027e-06,
433
+ "loss": 0.6992,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 1.58,
438
+ "learning_rate": 9.634779769423412e-06,
439
+ "loss": 0.7368,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 1.6,
444
+ "learning_rate": 9.391540284489862e-06,
445
+ "loss": 0.6738,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 1.63,
450
+ "learning_rate": 9.148661652863644e-06,
451
+ "loss": 0.6826,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 1.65,
456
+ "learning_rate": 8.906287916221259e-06,
457
+ "loss": 0.7212,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 1.67,
462
+ "learning_rate": 8.664562816806022e-06,
463
+ "loss": 0.6875,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 1.69,
468
+ "learning_rate": 8.423629712180265e-06,
469
+ "loss": 0.7065,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 1.71,
474
+ "learning_rate": 8.183631490205636e-06,
475
+ "loss": 0.6846,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 1.74,
480
+ "learning_rate": 7.944710484301995e-06,
481
+ "loss": 0.666,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 1.76,
486
+ "learning_rate": 7.707008389035102e-06,
487
+ "loss": 0.6987,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 1.78,
492
+ "learning_rate": 7.470666176083193e-06,
493
+ "loss": 0.6719,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 1.8,
498
+ "learning_rate": 7.235824010632284e-06,
499
+ "loss": 0.6743,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 1.82,
504
+ "learning_rate": 7.002621168249759e-06,
505
+ "loss": 0.6567,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 1.85,
510
+ "learning_rate": 6.771195952285541e-06,
511
+ "loss": 0.6733,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 1.87,
516
+ "learning_rate": 6.5416856118498874e-06,
517
+ "loss": 0.6538,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 1.89,
522
+ "learning_rate": 6.314226260416383e-06,
523
+ "loss": 0.7173,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 1.91,
528
+ "learning_rate": 6.088952795098442e-06,
529
+ "loss": 0.6758,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 1.93,
534
+ "learning_rate": 5.8659988166471715e-06,
535
+ "loss": 0.6123,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 1.96,
540
+ "learning_rate": 5.645496550218089e-06,
541
+ "loss": 0.6304,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 1.98,
546
+ "learning_rate": 5.427576766953615e-06,
547
+ "loss": 0.667,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 2.0,
552
+ "learning_rate": 5.212368706427913e-06,
553
+ "loss": 0.6431,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 2.02,
558
+ "learning_rate": 5.000000000000003e-06,
559
+ "loss": 0.583,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 2.04,
564
+ "learning_rate": 4.790596595120699e-06,
565
+ "loss": 0.6089,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 2.07,
570
+ "learning_rate": 4.584282680638155e-06,
571
+ "loss": 0.5725,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 2.09,
576
+ "learning_rate": 4.381180613146396e-06,
577
+ "loss": 0.5759,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 2.11,
582
+ "learning_rate": 4.181410844420473e-06,
583
+ "loss": 0.5928,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 2.13,
588
+ "learning_rate": 3.9850918499812976e-06,
589
+ "loss": 0.6167,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 2.15,
594
+ "learning_rate": 3.7923400588325156e-06,
595
+ "loss": 0.5425,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 2.18,
600
+ "learning_rate": 3.6032697844110896e-06,
601
+ "loss": 0.5405,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 2.2,
606
+ "learning_rate": 3.4179931567925216e-06,
607
+ "loss": 0.5547,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 2.22,
612
+ "learning_rate": 3.236620056190972e-06,
613
+ "loss": 0.5686,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 2.24,
618
+ "learning_rate": 3.0592580477936606e-06,
619
+ "loss": 0.5874,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 2.26,
624
+ "learning_rate": 2.8860123179682244e-06,
625
+ "loss": 0.5217,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 2.29,
630
+ "learning_rate": 2.7169856118808414e-06,
631
+ "loss": 0.5024,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 2.31,
636
+ "learning_rate": 2.5522781725621814e-06,
637
+ "loss": 0.5518,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 2.33,
642
+ "learning_rate": 2.3919876814572197e-06,
643
+ "loss": 0.5415,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 2.35,
648
+ "learning_rate": 2.2362092004942583e-06,
649
+ "loss": 0.5635,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 2.37,
654
+ "learning_rate": 2.08503511570746e-06,
655
+ "loss": 0.5464,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 2.4,
660
+ "learning_rate": 1.9385550824463727e-06,
661
+ "loss": 0.5669,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 2.42,
666
+ "learning_rate": 1.7968559722048906e-06,
667
+ "loss": 0.4912,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 2.44,
672
+ "learning_rate": 1.660021821101222e-06,
673
+ "loss": 0.533,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 2.46,
678
+ "learning_rate": 1.528133780039397e-06,
679
+ "loss": 0.5229,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 2.48,
684
+ "learning_rate": 1.401270066581899e-06,
685
+ "loss": 0.5544,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 2.51,
690
+ "learning_rate": 1.279505918561923e-06,
691
+ "loss": 0.54,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 2.53,
696
+ "learning_rate": 1.1629135494628097e-06,
697
+ "loss": 0.5322,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 2.55,
702
+ "learning_rate": 1.051562105591082e-06,
703
+ "loss": 0.5723,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 2.57,
708
+ "learning_rate": 9.455176250685338e-07,
709
+ "loss": 0.5669,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 2.59,
714
+ "learning_rate": 8.448429986676298e-07,
715
+ "loss": 0.5312,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 2.62,
720
+ "learning_rate": 7.495979325134806e-07,
721
+ "loss": 0.563,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 2.64,
726
+ "learning_rate": 6.598389126745209e-07,
727
+ "loss": 0.5959,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 2.66,
732
+ "learning_rate": 5.756191716628556e-07,
733
+ "loss": 0.5435,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 2.68,
738
+ "learning_rate": 4.969886568641757e-07,
739
+ "loss": 0.5308,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 2.7,
744
+ "learning_rate": 4.2399400091594154e-07,
745
+ "loss": 0.5337,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 2.73,
750
+ "learning_rate": 3.566784940514145e-07,
751
+ "loss": 0.5068,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 2.75,
756
+ "learning_rate": 2.9508205842594727e-07,
757
+ "loss": 0.5024,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 2.77,
762
+ "learning_rate": 2.392412244407294e-07,
763
+ "loss": 0.554,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 2.79,
768
+ "learning_rate": 1.8918910907805733e-07,
769
+ "loss": 0.5505,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 2.81,
774
+ "learning_rate": 1.4495539626097289e-07,
775
+ "loss": 0.5974,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 2.84,
780
+ "learning_rate": 1.0656631924889749e-07,
781
+ "loss": 0.4895,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 2.86,
786
+ "learning_rate": 7.404464507973608e-08,
787
+ "loss": 0.5193,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 2.88,
792
+ "learning_rate": 4.740966106764222e-08,
793
+ "loss": 0.5391,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 2.9,
798
+ "learning_rate": 2.667716336448356e-08,
799
+ "loss": 0.5173,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 2.92,
804
+ "learning_rate": 1.1859447591769934e-08,
805
+ "loss": 0.5071,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 2.95,
810
+ "learning_rate": 2.9653015486064143e-09,
811
+ "loss": 0.4993,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 2.97,
816
+ "learning_rate": 0.0,
817
+ "loss": 0.5388,
818
+ "step": 135
819
+ }
820
+ ],
821
+ "max_steps": 135,
822
+ "num_train_epochs": 3,
823
+ "total_flos": 169585743495168.0,
824
+ "trial_name": null,
825
+ "trial_params": null
826
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b70d9cbc45eaff0bc4aca899a5f9a3c611a3d4605c926b478fe2d76f267d0b0
3
+ size 4731
zero_to_fp32.py ADDED
@@ -0,0 +1,578 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python
2
+
3
+ # Copyright (c) Microsoft Corporation.
4
+ # SPDX-License-Identifier: Apache-2.0
5
+
6
+ # DeepSpeed Team
7
+
8
+ # This script extracts fp32 consolidated weights from a zero 2 and 3 DeepSpeed checkpoints. It gets
9
+ # copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
10
+ # the future. Once extracted, the weights don't require DeepSpeed and can be used in any
11
+ # application.
12
+ #
13
+ # example: python zero_to_fp32.py . pytorch_model.bin
14
+
15
+ import argparse
16
+ import torch
17
+ import glob
18
+ import math
19
+ import os
20
+ import re
21
+ from collections import OrderedDict
22
+ from dataclasses import dataclass
23
+
24
+ # while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
25
+ # DeepSpeed data structures it has to be available in the current python environment.
26
+ from deepspeed.utils import logger
27
+ from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
28
+ FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
29
+ FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
30
+
31
+
32
+ @dataclass
33
+ class zero_model_state:
34
+ buffers: dict()
35
+ param_shapes: dict()
36
+ shared_params: list
37
+ ds_version: int
38
+ frozen_param_shapes: dict()
39
+ frozen_param_fragments: dict()
40
+
41
+
42
+ debug = 0
43
+
44
+ # load to cpu
45
+ device = torch.device('cpu')
46
+
47
+
48
+ def atoi(text):
49
+ return int(text) if text.isdigit() else text
50
+
51
+
52
+ def natural_keys(text):
53
+ '''
54
+ alist.sort(key=natural_keys) sorts in human order
55
+ http://nedbatchelder.com/blog/200712/human_sorting.html
56
+ (See Toothy's implementation in the comments)
57
+ '''
58
+ return [atoi(c) for c in re.split(r'(\d+)', text)]
59
+
60
+
61
+ def get_model_state_file(checkpoint_dir, zero_stage):
62
+ if not os.path.isdir(checkpoint_dir):
63
+ raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
64
+
65
+ # there should be only one file
66
+ if zero_stage == 2:
67
+ file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
68
+ elif zero_stage == 3:
69
+ file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
70
+
71
+ if not os.path.exists(file):
72
+ raise FileNotFoundError(f"can't find model states file at '{file}'")
73
+
74
+ return file
75
+
76
+
77
+ def get_checkpoint_files(checkpoint_dir, glob_pattern):
78
+ # XXX: need to test that this simple glob rule works for multi-node setup too
79
+ ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
80
+
81
+ if len(ckpt_files) == 0:
82
+ raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
83
+
84
+ return ckpt_files
85
+
86
+
87
+ def get_optim_files(checkpoint_dir):
88
+ return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
89
+
90
+
91
+ def get_model_state_files(checkpoint_dir):
92
+ return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
93
+
94
+
95
+ def parse_model_states(files):
96
+ zero_model_states = []
97
+ for file in files:
98
+ state_dict = torch.load(file, map_location=device)
99
+
100
+ if BUFFER_NAMES not in state_dict:
101
+ raise ValueError(f"{file} is not a model state checkpoint")
102
+ buffer_names = state_dict[BUFFER_NAMES]
103
+ if debug:
104
+ print("Found buffers:", buffer_names)
105
+
106
+ # recover just the buffers while restoring them to fp32 if they were saved in fp16
107
+ buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
108
+ param_shapes = state_dict[PARAM_SHAPES]
109
+
110
+ # collect parameters that are included in param_shapes
111
+ param_names = []
112
+ for s in param_shapes:
113
+ for name in s.keys():
114
+ param_names.append(name)
115
+
116
+ # update with frozen parameters
117
+ frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
118
+ if frozen_param_shapes is not None:
119
+ if debug:
120
+ print(f"Found frozen_param_shapes: {frozen_param_shapes}")
121
+ param_names += list(frozen_param_shapes.keys())
122
+
123
+ # handle shared params
124
+ shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
125
+
126
+ ds_version = state_dict.get(DS_VERSION, None)
127
+
128
+ frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
129
+
130
+ z_model_state = zero_model_state(buffers=buffers,
131
+ param_shapes=param_shapes,
132
+ shared_params=shared_params,
133
+ ds_version=ds_version,
134
+ frozen_param_shapes=frozen_param_shapes,
135
+ frozen_param_fragments=frozen_param_fragments)
136
+ zero_model_states.append(z_model_state)
137
+
138
+ return zero_model_states
139
+
140
+
141
+ def parse_optim_states(files, ds_checkpoint_dir):
142
+
143
+ total_files = len(files)
144
+ state_dicts = []
145
+ for f in files:
146
+ state_dicts.append(torch.load(f, map_location=device))
147
+
148
+ if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
149
+ raise ValueError(f"{files[0]} is not a zero checkpoint")
150
+ zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
151
+ world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
152
+
153
+ # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
154
+ # parameters can be different from data parallelism for non-expert parameters. So we can just
155
+ # use the max of the partition_count to get the dp world_size.
156
+
157
+ if type(world_size) is list:
158
+ world_size = max(world_size)
159
+
160
+ if world_size != total_files:
161
+ raise ValueError(
162
+ f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
163
+ "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
164
+ )
165
+
166
+ # the groups are named differently in each stage
167
+ if zero_stage == 2:
168
+ fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
169
+ elif zero_stage == 3:
170
+ fp32_groups_key = FP32_FLAT_GROUPS
171
+ else:
172
+ raise ValueError(f"unknown zero stage {zero_stage}")
173
+
174
+ if zero_stage == 2:
175
+ fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
176
+ elif zero_stage == 3:
177
+ # if there is more than one param group, there will be multiple flattened tensors - one
178
+ # flattened tensor per group - for simplicity merge them into a single tensor
179
+ #
180
+ # XXX: could make the script more memory efficient for when there are multiple groups - it
181
+ # will require matching the sub-lists of param_shapes for each param group flattened tensor
182
+
183
+ fp32_flat_groups = [
184
+ torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
185
+ ]
186
+
187
+ return zero_stage, world_size, fp32_flat_groups
188
+
189
+
190
+ def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
191
+ """
192
+ Returns fp32 state_dict reconstructed from ds checkpoint
193
+
194
+ Args:
195
+ - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
196
+
197
+ """
198
+ print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
199
+
200
+ optim_files = get_optim_files(ds_checkpoint_dir)
201
+ zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
202
+ print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
203
+
204
+ model_files = get_model_state_files(ds_checkpoint_dir)
205
+
206
+ zero_model_states = parse_model_states(model_files)
207
+ print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
208
+
209
+ if zero_stage == 2:
210
+ return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
211
+ elif zero_stage == 3:
212
+ return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
213
+
214
+
215
+ def _zero2_merge_frozen_params(state_dict, zero_model_states):
216
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
217
+ return
218
+
219
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
220
+ frozen_param_fragments = zero_model_states[0].frozen_param_fragments
221
+
222
+ if debug:
223
+ num_elem = sum(s.numel() for s in frozen_param_shapes.values())
224
+ print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
225
+
226
+ wanted_params = len(frozen_param_shapes)
227
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
228
+ avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
229
+ print(f'Frozen params: Have {avail_numel} numels to process.')
230
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
231
+
232
+ total_params = 0
233
+ total_numel = 0
234
+ for name, shape in frozen_param_shapes.items():
235
+ total_params += 1
236
+ unpartitioned_numel = shape.numel()
237
+ total_numel += unpartitioned_numel
238
+
239
+ state_dict[name] = frozen_param_fragments[name]
240
+
241
+ if debug:
242
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
243
+
244
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
245
+
246
+
247
+ def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
248
+ param_shapes = zero_model_states[0].param_shapes
249
+
250
+ # Reconstruction protocol:
251
+ #
252
+ # XXX: document this
253
+
254
+ if debug:
255
+ for i in range(world_size):
256
+ for j in range(len(fp32_flat_groups[0])):
257
+ print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
258
+
259
+ # XXX: memory usage doubles here (zero2)
260
+ num_param_groups = len(fp32_flat_groups[0])
261
+ merged_single_partition_of_fp32_groups = []
262
+ for i in range(num_param_groups):
263
+ merged_partitions = [sd[i] for sd in fp32_flat_groups]
264
+ full_single_fp32_vector = torch.cat(merged_partitions, 0)
265
+ merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
266
+ avail_numel = sum(
267
+ [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
268
+
269
+ if debug:
270
+ wanted_params = sum([len(shapes) for shapes in param_shapes])
271
+ wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
272
+ # not asserting if there is a mismatch due to possible padding
273
+ print(f"Have {avail_numel} numels to process.")
274
+ print(f"Need {wanted_numel} numels in {wanted_params} params.")
275
+
276
+ # params
277
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
278
+ # out-of-core computing solution
279
+ total_numel = 0
280
+ total_params = 0
281
+ for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
282
+ offset = 0
283
+ avail_numel = full_single_fp32_vector.numel()
284
+ for name, shape in shapes.items():
285
+
286
+ unpartitioned_numel = shape.numel()
287
+ total_numel += unpartitioned_numel
288
+ total_params += 1
289
+
290
+ if debug:
291
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
292
+ state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
293
+ offset += unpartitioned_numel
294
+
295
+ # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
296
+ # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
297
+ # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
298
+ # live optimizer object, so we are checking that the numbers are within the right range
299
+ align_to = 2 * world_size
300
+
301
+ def zero2_align(x):
302
+ return align_to * math.ceil(x / align_to)
303
+
304
+ if debug:
305
+ print(f"original offset={offset}, avail_numel={avail_numel}")
306
+
307
+ offset = zero2_align(offset)
308
+ avail_numel = zero2_align(avail_numel)
309
+
310
+ if debug:
311
+ print(f"aligned offset={offset}, avail_numel={avail_numel}")
312
+
313
+ # Sanity check
314
+ if offset != avail_numel:
315
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
316
+
317
+ print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
318
+
319
+
320
+ def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
321
+ state_dict = OrderedDict()
322
+
323
+ # buffers
324
+ buffers = zero_model_states[0].buffers
325
+ state_dict.update(buffers)
326
+ if debug:
327
+ print(f"added {len(buffers)} buffers")
328
+
329
+ _zero2_merge_frozen_params(state_dict, zero_model_states)
330
+
331
+ _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
332
+
333
+ # recover shared parameters
334
+ for pair in zero_model_states[0].shared_params:
335
+ if pair[1] in state_dict:
336
+ state_dict[pair[0]] = state_dict[pair[1]]
337
+
338
+ return state_dict
339
+
340
+
341
+ def zero3_partitioned_param_info(unpartitioned_numel, world_size):
342
+ remainder = unpartitioned_numel % world_size
343
+ padding_numel = (world_size - remainder) if remainder else 0
344
+ partitioned_numel = math.ceil(unpartitioned_numel / world_size)
345
+ return partitioned_numel, padding_numel
346
+
347
+
348
+ def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
349
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
350
+ return
351
+
352
+ if debug:
353
+ for i in range(world_size):
354
+ num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
355
+ print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
356
+
357
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
358
+ wanted_params = len(frozen_param_shapes)
359
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
360
+ avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
361
+ print(f'Frozen params: Have {avail_numel} numels to process.')
362
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
363
+
364
+ total_params = 0
365
+ total_numel = 0
366
+ for name, shape in zero_model_states[0].frozen_param_shapes.items():
367
+ total_params += 1
368
+ unpartitioned_numel = shape.numel()
369
+ total_numel += unpartitioned_numel
370
+
371
+ param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
372
+ state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
373
+
374
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
375
+
376
+ if debug:
377
+ print(
378
+ f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
379
+ )
380
+
381
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
382
+
383
+
384
+ def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
385
+ param_shapes = zero_model_states[0].param_shapes
386
+ avail_numel = fp32_flat_groups[0].numel() * world_size
387
+ # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
388
+ # param, re-consolidating each param, while dealing with padding if any
389
+
390
+ # merge list of dicts, preserving order
391
+ param_shapes = {k: v for d in param_shapes for k, v in d.items()}
392
+
393
+ if debug:
394
+ for i in range(world_size):
395
+ print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
396
+
397
+ wanted_params = len(param_shapes)
398
+ wanted_numel = sum(shape.numel() for shape in param_shapes.values())
399
+ # not asserting if there is a mismatch due to possible padding
400
+ avail_numel = fp32_flat_groups[0].numel() * world_size
401
+ print(f"Trainable params: Have {avail_numel} numels to process.")
402
+ print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
403
+
404
+ # params
405
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
406
+ # out-of-core computing solution
407
+ offset = 0
408
+ total_numel = 0
409
+ total_params = 0
410
+ for name, shape in param_shapes.items():
411
+
412
+ unpartitioned_numel = shape.numel()
413
+ total_numel += unpartitioned_numel
414
+ total_params += 1
415
+
416
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
417
+
418
+ if debug:
419
+ print(
420
+ f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
421
+ )
422
+
423
+ # XXX: memory usage doubles here
424
+ state_dict[name] = torch.cat(
425
+ tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
426
+ 0).narrow(0, 0, unpartitioned_numel).view(shape)
427
+ offset += partitioned_numel
428
+
429
+ offset *= world_size
430
+
431
+ # Sanity check
432
+ if offset != avail_numel:
433
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
434
+
435
+ print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
436
+
437
+
438
+ def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
439
+ state_dict = OrderedDict()
440
+
441
+ # buffers
442
+ buffers = zero_model_states[0].buffers
443
+ state_dict.update(buffers)
444
+ if debug:
445
+ print(f"added {len(buffers)} buffers")
446
+
447
+ _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
448
+
449
+ _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
450
+
451
+ # recover shared parameters
452
+ for pair in zero_model_states[0].shared_params:
453
+ if pair[1] in state_dict:
454
+ state_dict[pair[0]] = state_dict[pair[1]]
455
+
456
+ return state_dict
457
+
458
+
459
+ def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
460
+ """
461
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
462
+ ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
463
+ via a model hub.
464
+
465
+ Args:
466
+ - ``checkpoint_dir``: path to the desired checkpoint folder
467
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
468
+
469
+ Returns:
470
+ - pytorch ``state_dict``
471
+
472
+ Note: this approach may not work if your application doesn't have sufficient free CPU memory and
473
+ you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
474
+ the checkpoint.
475
+
476
+ A typical usage might be ::
477
+
478
+ from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
479
+ # do the training and checkpoint saving
480
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
481
+ model = model.cpu() # move to cpu
482
+ model.load_state_dict(state_dict)
483
+ # submit to model hub or save the model to share with others
484
+
485
+ In this example the ``model`` will no longer be usable in the deepspeed context of the same
486
+ application. i.e. you will need to re-initialize the deepspeed engine, since
487
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
488
+
489
+ If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
490
+
491
+ """
492
+ if tag is None:
493
+ latest_path = os.path.join(checkpoint_dir, 'latest')
494
+ if os.path.isfile(latest_path):
495
+ with open(latest_path, 'r') as fd:
496
+ tag = fd.read().strip()
497
+ else:
498
+ raise ValueError(f"Unable to find 'latest' file at {latest_path}")
499
+
500
+ ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
501
+
502
+ if not os.path.isdir(ds_checkpoint_dir):
503
+ raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
504
+
505
+ return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
506
+
507
+
508
+ def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
509
+ """
510
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
511
+ loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
512
+
513
+ Args:
514
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
515
+ - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
516
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
517
+ """
518
+
519
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
520
+ print(f"Saving fp32 state dict to {output_file}")
521
+ torch.save(state_dict, output_file)
522
+
523
+
524
+ def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
525
+ """
526
+ 1. Put the provided model to cpu
527
+ 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
528
+ 3. Load it into the provided model
529
+
530
+ Args:
531
+ - ``model``: the model object to update
532
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
533
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
534
+
535
+ Returns:
536
+ - ``model`: modified model
537
+
538
+ Make sure you have plenty of CPU memory available before you call this function. If you don't
539
+ have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
540
+ conveniently placed for you in the checkpoint folder.
541
+
542
+ A typical usage might be ::
543
+
544
+ from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
545
+ model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
546
+ # submit to model hub or save the model to share with others
547
+
548
+ Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
549
+ of the same application. i.e. you will need to re-initialize the deepspeed engine, since
550
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
551
+
552
+ """
553
+ logger.info(f"Extracting fp32 weights")
554
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
555
+
556
+ logger.info(f"Overwriting model with fp32 weights")
557
+ model = model.cpu()
558
+ model.load_state_dict(state_dict, strict=False)
559
+
560
+ return model
561
+
562
+
563
+ if __name__ == "__main__":
564
+
565
+ parser = argparse.ArgumentParser()
566
+ parser.add_argument("checkpoint_dir",
567
+ type=str,
568
+ help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
569
+ parser.add_argument(
570
+ "output_file",
571
+ type=str,
572
+ help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
573
+ parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
574
+ args = parser.parse_args()
575
+
576
+ debug = args.debug
577
+
578
+ convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file)