ArthurZ HF staff commited on
Commit
ca6f404
1 Parent(s): 55d7dfd
Files changed (4) hide show
  1. README.md +47 -0
  2. config.json +27 -0
  3. tf_model-00001-of-00013.h5 +3 -0
  4. tf_model.h5.index.json +777 -0
README.md ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_keras_callback
4
+ model-index:
5
+ - name: opt-30b-sharded
6
+ results: []
7
+ ---
8
+
9
+ <!-- This model card has been generated automatically according to the information Keras had access to. You should
10
+ probably proofread and complete it, then remove this comment. -->
11
+
12
+ # opt-30b-sharded
13
+
14
+ This model was trained from scratch on an unknown dataset.
15
+ It achieves the following results on the evaluation set:
16
+
17
+
18
+ ## Model description
19
+
20
+ More information needed
21
+
22
+ ## Intended uses & limitations
23
+
24
+ More information needed
25
+
26
+ ## Training and evaluation data
27
+
28
+ More information needed
29
+
30
+ ## Training procedure
31
+
32
+ ### Training hyperparameters
33
+
34
+ The following hyperparameters were used during training:
35
+ - optimizer: None
36
+ - training_precision: float32
37
+
38
+ ### Training results
39
+
40
+
41
+
42
+ ### Framework versions
43
+
44
+ - Transformers 4.20.0.dev0
45
+ - TensorFlow 2.9.1
46
+ - Datasets 2.2.2
47
+ - Tokenizers 0.12.1
config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "opt-30b",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "relu",
5
+ "architectures": [
6
+ "OPTForCausalLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 2,
10
+ "do_layer_norm_before": true,
11
+ "dropout": 0.1,
12
+ "eos_token_id": 2,
13
+ "ffn_dim": 28672,
14
+ "hidden_size": 7168,
15
+ "init_std": 0.02,
16
+ "layerdrop": 0.0,
17
+ "max_position_embeddings": 2048,
18
+ "model_type": "opt",
19
+ "num_attention_heads": 56,
20
+ "num_hidden_layers": 48,
21
+ "pad_token_id": 1,
22
+ "torch_dtype": "float16",
23
+ "transformers_version": "4.20.0.dev0",
24
+ "use_cache": true,
25
+ "vocab_size": 50272,
26
+ "word_embed_proj_dim": 7168
27
+ }
tf_model-00001-of-00013.h5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:39dbe33709fb7a4995e27a6ee4d603742cd05320c3d4f1e9899ae49ca44187a0
3
+ size 9722369432
tf_model.h5.index.json ADDED
@@ -0,0 +1,777 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 119898103808
4
+ },
5
+ "weight_map": {
6
+ "decoder.embed_tokens/model.decoder.embed_tokens/weight:0": "tf_model-00001-of-00013.h5",
7
+ "tfopt_for_causal_lm/model/decoder/embed_positions/weight:0": "tf_model-00001-of-00013.h5",
8
+ "tfopt_for_causal_lm/model/decoder/layers.0/fc1/bias:0": "tf_model-00001-of-00013.h5",
9
+ "tfopt_for_causal_lm/model/decoder/layers.0/fc1/kernel:0": "tf_model-00001-of-00013.h5",
10
+ "tfopt_for_causal_lm/model/decoder/layers.0/fc2/bias:0": "tf_model-00001-of-00013.h5",
11
+ "tfopt_for_causal_lm/model/decoder/layers.0/fc2/kernel:0": "tf_model-00001-of-00013.h5",
12
+ "tfopt_for_causal_lm/model/decoder/layers.0/final_layer_norm/beta:0": "tf_model-00001-of-00013.h5",
13
+ "tfopt_for_causal_lm/model/decoder/layers.0/final_layer_norm/gamma:0": "tf_model-00001-of-00013.h5",
14
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn/k_proj/bias:0": "tf_model-00001-of-00013.h5",
15
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn/k_proj/kernel:0": "tf_model-00001-of-00013.h5",
16
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn/out_proj/bias:0": "tf_model-00001-of-00013.h5",
17
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn/out_proj/kernel:0": "tf_model-00001-of-00013.h5",
18
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn/q_proj/bias:0": "tf_model-00001-of-00013.h5",
19
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn/q_proj/kernel:0": "tf_model-00001-of-00013.h5",
20
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn/v_proj/bias:0": "tf_model-00001-of-00013.h5",
21
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn/v_proj/kernel:0": "tf_model-00001-of-00013.h5",
22
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn_layer_norm/beta:0": "tf_model-00001-of-00013.h5",
23
+ "tfopt_for_causal_lm/model/decoder/layers.0/self_attn_layer_norm/gamma:0": "tf_model-00001-of-00013.h5",
24
+ "tfopt_for_causal_lm/model/decoder/layers.1/fc1/bias:0": "tf_model-00001-of-00013.h5",
25
+ "tfopt_for_causal_lm/model/decoder/layers.1/fc1/kernel:0": "tf_model-00001-of-00013.h5",
26
+ "tfopt_for_causal_lm/model/decoder/layers.1/fc2/bias:0": "tf_model-00001-of-00013.h5",
27
+ "tfopt_for_causal_lm/model/decoder/layers.1/fc2/kernel:0": "tf_model-00001-of-00013.h5",
28
+ "tfopt_for_causal_lm/model/decoder/layers.1/final_layer_norm/beta:0": "tf_model-00001-of-00013.h5",
29
+ "tfopt_for_causal_lm/model/decoder/layers.1/final_layer_norm/gamma:0": "tf_model-00001-of-00013.h5",
30
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn/k_proj/bias:0": "tf_model-00001-of-00013.h5",
31
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn/k_proj/kernel:0": "tf_model-00001-of-00013.h5",
32
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn/out_proj/bias:0": "tf_model-00001-of-00013.h5",
33
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn/out_proj/kernel:0": "tf_model-00001-of-00013.h5",
34
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn/q_proj/bias:0": "tf_model-00001-of-00013.h5",
35
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn/q_proj/kernel:0": "tf_model-00001-of-00013.h5",
36
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn/v_proj/bias:0": "tf_model-00001-of-00013.h5",
37
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn/v_proj/kernel:0": "tf_model-00001-of-00013.h5",
38
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn_layer_norm/beta:0": "tf_model-00001-of-00013.h5",
39
+ "tfopt_for_causal_lm/model/decoder/layers.1/self_attn_layer_norm/gamma:0": "tf_model-00001-of-00013.h5",
40
+ "tfopt_for_causal_lm/model/decoder/layers.10/fc1/bias:0": "tf_model-00003-of-00013.h5",
41
+ "tfopt_for_causal_lm/model/decoder/layers.10/fc1/kernel:0": "tf_model-00003-of-00013.h5",
42
+ "tfopt_for_causal_lm/model/decoder/layers.10/fc2/bias:0": "tf_model-00003-of-00013.h5",
43
+ "tfopt_for_causal_lm/model/decoder/layers.10/fc2/kernel:0": "tf_model-00003-of-00013.h5",
44
+ "tfopt_for_causal_lm/model/decoder/layers.10/final_layer_norm/beta:0": "tf_model-00003-of-00013.h5",
45
+ "tfopt_for_causal_lm/model/decoder/layers.10/final_layer_norm/gamma:0": "tf_model-00003-of-00013.h5",
46
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn/k_proj/bias:0": "tf_model-00003-of-00013.h5",
47
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn/k_proj/kernel:0": "tf_model-00003-of-00013.h5",
48
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn/out_proj/bias:0": "tf_model-00003-of-00013.h5",
49
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn/out_proj/kernel:0": "tf_model-00003-of-00013.h5",
50
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn/q_proj/bias:0": "tf_model-00003-of-00013.h5",
51
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn/q_proj/kernel:0": "tf_model-00003-of-00013.h5",
52
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn/v_proj/bias:0": "tf_model-00003-of-00013.h5",
53
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn/v_proj/kernel:0": "tf_model-00003-of-00013.h5",
54
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn_layer_norm/beta:0": "tf_model-00003-of-00013.h5",
55
+ "tfopt_for_causal_lm/model/decoder/layers.10/self_attn_layer_norm/gamma:0": "tf_model-00003-of-00013.h5",
56
+ "tfopt_for_causal_lm/model/decoder/layers.11/fc1/bias:0": "tf_model-00004-of-00013.h5",
57
+ "tfopt_for_causal_lm/model/decoder/layers.11/fc1/kernel:0": "tf_model-00004-of-00013.h5",
58
+ "tfopt_for_causal_lm/model/decoder/layers.11/fc2/bias:0": "tf_model-00004-of-00013.h5",
59
+ "tfopt_for_causal_lm/model/decoder/layers.11/fc2/kernel:0": "tf_model-00004-of-00013.h5",
60
+ "tfopt_for_causal_lm/model/decoder/layers.11/final_layer_norm/beta:0": "tf_model-00004-of-00013.h5",
61
+ "tfopt_for_causal_lm/model/decoder/layers.11/final_layer_norm/gamma:0": "tf_model-00004-of-00013.h5",
62
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn/k_proj/bias:0": "tf_model-00003-of-00013.h5",
63
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn/k_proj/kernel:0": "tf_model-00003-of-00013.h5",
64
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn/out_proj/bias:0": "tf_model-00003-of-00013.h5",
65
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn/out_proj/kernel:0": "tf_model-00003-of-00013.h5",
66
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn/q_proj/bias:0": "tf_model-00003-of-00013.h5",
67
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn/q_proj/kernel:0": "tf_model-00003-of-00013.h5",
68
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn/v_proj/bias:0": "tf_model-00003-of-00013.h5",
69
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn/v_proj/kernel:0": "tf_model-00003-of-00013.h5",
70
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn_layer_norm/beta:0": "tf_model-00003-of-00013.h5",
71
+ "tfopt_for_causal_lm/model/decoder/layers.11/self_attn_layer_norm/gamma:0": "tf_model-00003-of-00013.h5",
72
+ "tfopt_for_causal_lm/model/decoder/layers.12/fc1/bias:0": "tf_model-00004-of-00013.h5",
73
+ "tfopt_for_causal_lm/model/decoder/layers.12/fc1/kernel:0": "tf_model-00004-of-00013.h5",
74
+ "tfopt_for_causal_lm/model/decoder/layers.12/fc2/bias:0": "tf_model-00004-of-00013.h5",
75
+ "tfopt_for_causal_lm/model/decoder/layers.12/fc2/kernel:0": "tf_model-00004-of-00013.h5",
76
+ "tfopt_for_causal_lm/model/decoder/layers.12/final_layer_norm/beta:0": "tf_model-00004-of-00013.h5",
77
+ "tfopt_for_causal_lm/model/decoder/layers.12/final_layer_norm/gamma:0": "tf_model-00004-of-00013.h5",
78
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn/k_proj/bias:0": "tf_model-00004-of-00013.h5",
79
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn/k_proj/kernel:0": "tf_model-00004-of-00013.h5",
80
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn/out_proj/bias:0": "tf_model-00004-of-00013.h5",
81
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn/out_proj/kernel:0": "tf_model-00004-of-00013.h5",
82
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn/q_proj/bias:0": "tf_model-00004-of-00013.h5",
83
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn/q_proj/kernel:0": "tf_model-00004-of-00013.h5",
84
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn/v_proj/bias:0": "tf_model-00004-of-00013.h5",
85
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn/v_proj/kernel:0": "tf_model-00004-of-00013.h5",
86
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn_layer_norm/beta:0": "tf_model-00004-of-00013.h5",
87
+ "tfopt_for_causal_lm/model/decoder/layers.12/self_attn_layer_norm/gamma:0": "tf_model-00004-of-00013.h5",
88
+ "tfopt_for_causal_lm/model/decoder/layers.13/fc1/bias:0": "tf_model-00004-of-00013.h5",
89
+ "tfopt_for_causal_lm/model/decoder/layers.13/fc1/kernel:0": "tf_model-00004-of-00013.h5",
90
+ "tfopt_for_causal_lm/model/decoder/layers.13/fc2/bias:0": "tf_model-00004-of-00013.h5",
91
+ "tfopt_for_causal_lm/model/decoder/layers.13/fc2/kernel:0": "tf_model-00004-of-00013.h5",
92
+ "tfopt_for_causal_lm/model/decoder/layers.13/final_layer_norm/beta:0": "tf_model-00004-of-00013.h5",
93
+ "tfopt_for_causal_lm/model/decoder/layers.13/final_layer_norm/gamma:0": "tf_model-00004-of-00013.h5",
94
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn/k_proj/bias:0": "tf_model-00004-of-00013.h5",
95
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn/k_proj/kernel:0": "tf_model-00004-of-00013.h5",
96
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn/out_proj/bias:0": "tf_model-00004-of-00013.h5",
97
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn/out_proj/kernel:0": "tf_model-00004-of-00013.h5",
98
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn/q_proj/bias:0": "tf_model-00004-of-00013.h5",
99
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn/q_proj/kernel:0": "tf_model-00004-of-00013.h5",
100
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn/v_proj/bias:0": "tf_model-00004-of-00013.h5",
101
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn/v_proj/kernel:0": "tf_model-00004-of-00013.h5",
102
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn_layer_norm/beta:0": "tf_model-00004-of-00013.h5",
103
+ "tfopt_for_causal_lm/model/decoder/layers.13/self_attn_layer_norm/gamma:0": "tf_model-00004-of-00013.h5",
104
+ "tfopt_for_causal_lm/model/decoder/layers.14/fc1/bias:0": "tf_model-00004-of-00013.h5",
105
+ "tfopt_for_causal_lm/model/decoder/layers.14/fc1/kernel:0": "tf_model-00004-of-00013.h5",
106
+ "tfopt_for_causal_lm/model/decoder/layers.14/fc2/bias:0": "tf_model-00004-of-00013.h5",
107
+ "tfopt_for_causal_lm/model/decoder/layers.14/fc2/kernel:0": "tf_model-00004-of-00013.h5",
108
+ "tfopt_for_causal_lm/model/decoder/layers.14/final_layer_norm/beta:0": "tf_model-00004-of-00013.h5",
109
+ "tfopt_for_causal_lm/model/decoder/layers.14/final_layer_norm/gamma:0": "tf_model-00004-of-00013.h5",
110
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn/k_proj/bias:0": "tf_model-00004-of-00013.h5",
111
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn/k_proj/kernel:0": "tf_model-00004-of-00013.h5",
112
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn/out_proj/bias:0": "tf_model-00004-of-00013.h5",
113
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn/out_proj/kernel:0": "tf_model-00004-of-00013.h5",
114
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn/q_proj/bias:0": "tf_model-00004-of-00013.h5",
115
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn/q_proj/kernel:0": "tf_model-00004-of-00013.h5",
116
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn/v_proj/bias:0": "tf_model-00004-of-00013.h5",
117
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn/v_proj/kernel:0": "tf_model-00004-of-00013.h5",
118
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn_layer_norm/beta:0": "tf_model-00004-of-00013.h5",
119
+ "tfopt_for_causal_lm/model/decoder/layers.14/self_attn_layer_norm/gamma:0": "tf_model-00004-of-00013.h5",
120
+ "tfopt_for_causal_lm/model/decoder/layers.15/fc1/bias:0": "tf_model-00005-of-00013.h5",
121
+ "tfopt_for_causal_lm/model/decoder/layers.15/fc1/kernel:0": "tf_model-00005-of-00013.h5",
122
+ "tfopt_for_causal_lm/model/decoder/layers.15/fc2/bias:0": "tf_model-00005-of-00013.h5",
123
+ "tfopt_for_causal_lm/model/decoder/layers.15/fc2/kernel:0": "tf_model-00005-of-00013.h5",
124
+ "tfopt_for_causal_lm/model/decoder/layers.15/final_layer_norm/beta:0": "tf_model-00005-of-00013.h5",
125
+ "tfopt_for_causal_lm/model/decoder/layers.15/final_layer_norm/gamma:0": "tf_model-00005-of-00013.h5",
126
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn/k_proj/bias:0": "tf_model-00004-of-00013.h5",
127
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn/k_proj/kernel:0": "tf_model-00004-of-00013.h5",
128
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn/out_proj/bias:0": "tf_model-00004-of-00013.h5",
129
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn/out_proj/kernel:0": "tf_model-00004-of-00013.h5",
130
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn/q_proj/bias:0": "tf_model-00004-of-00013.h5",
131
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn/q_proj/kernel:0": "tf_model-00004-of-00013.h5",
132
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn/v_proj/bias:0": "tf_model-00004-of-00013.h5",
133
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn/v_proj/kernel:0": "tf_model-00004-of-00013.h5",
134
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn_layer_norm/beta:0": "tf_model-00004-of-00013.h5",
135
+ "tfopt_for_causal_lm/model/decoder/layers.15/self_attn_layer_norm/gamma:0": "tf_model-00004-of-00013.h5",
136
+ "tfopt_for_causal_lm/model/decoder/layers.16/fc1/bias:0": "tf_model-00005-of-00013.h5",
137
+ "tfopt_for_causal_lm/model/decoder/layers.16/fc1/kernel:0": "tf_model-00005-of-00013.h5",
138
+ "tfopt_for_causal_lm/model/decoder/layers.16/fc2/bias:0": "tf_model-00005-of-00013.h5",
139
+ "tfopt_for_causal_lm/model/decoder/layers.16/fc2/kernel:0": "tf_model-00005-of-00013.h5",
140
+ "tfopt_for_causal_lm/model/decoder/layers.16/final_layer_norm/beta:0": "tf_model-00005-of-00013.h5",
141
+ "tfopt_for_causal_lm/model/decoder/layers.16/final_layer_norm/gamma:0": "tf_model-00005-of-00013.h5",
142
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn/k_proj/bias:0": "tf_model-00005-of-00013.h5",
143
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn/k_proj/kernel:0": "tf_model-00005-of-00013.h5",
144
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn/out_proj/bias:0": "tf_model-00005-of-00013.h5",
145
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn/out_proj/kernel:0": "tf_model-00005-of-00013.h5",
146
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn/q_proj/bias:0": "tf_model-00005-of-00013.h5",
147
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn/q_proj/kernel:0": "tf_model-00005-of-00013.h5",
148
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn/v_proj/bias:0": "tf_model-00005-of-00013.h5",
149
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn/v_proj/kernel:0": "tf_model-00005-of-00013.h5",
150
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn_layer_norm/beta:0": "tf_model-00005-of-00013.h5",
151
+ "tfopt_for_causal_lm/model/decoder/layers.16/self_attn_layer_norm/gamma:0": "tf_model-00005-of-00013.h5",
152
+ "tfopt_for_causal_lm/model/decoder/layers.17/fc1/bias:0": "tf_model-00005-of-00013.h5",
153
+ "tfopt_for_causal_lm/model/decoder/layers.17/fc1/kernel:0": "tf_model-00005-of-00013.h5",
154
+ "tfopt_for_causal_lm/model/decoder/layers.17/fc2/bias:0": "tf_model-00005-of-00013.h5",
155
+ "tfopt_for_causal_lm/model/decoder/layers.17/fc2/kernel:0": "tf_model-00005-of-00013.h5",
156
+ "tfopt_for_causal_lm/model/decoder/layers.17/final_layer_norm/beta:0": "tf_model-00005-of-00013.h5",
157
+ "tfopt_for_causal_lm/model/decoder/layers.17/final_layer_norm/gamma:0": "tf_model-00005-of-00013.h5",
158
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn/k_proj/bias:0": "tf_model-00005-of-00013.h5",
159
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn/k_proj/kernel:0": "tf_model-00005-of-00013.h5",
160
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn/out_proj/bias:0": "tf_model-00005-of-00013.h5",
161
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn/out_proj/kernel:0": "tf_model-00005-of-00013.h5",
162
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn/q_proj/bias:0": "tf_model-00005-of-00013.h5",
163
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn/q_proj/kernel:0": "tf_model-00005-of-00013.h5",
164
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn/v_proj/bias:0": "tf_model-00005-of-00013.h5",
165
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn/v_proj/kernel:0": "tf_model-00005-of-00013.h5",
166
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn_layer_norm/beta:0": "tf_model-00005-of-00013.h5",
167
+ "tfopt_for_causal_lm/model/decoder/layers.17/self_attn_layer_norm/gamma:0": "tf_model-00005-of-00013.h5",
168
+ "tfopt_for_causal_lm/model/decoder/layers.18/fc1/bias:0": "tf_model-00005-of-00013.h5",
169
+ "tfopt_for_causal_lm/model/decoder/layers.18/fc1/kernel:0": "tf_model-00005-of-00013.h5",
170
+ "tfopt_for_causal_lm/model/decoder/layers.18/fc2/bias:0": "tf_model-00005-of-00013.h5",
171
+ "tfopt_for_causal_lm/model/decoder/layers.18/fc2/kernel:0": "tf_model-00005-of-00013.h5",
172
+ "tfopt_for_causal_lm/model/decoder/layers.18/final_layer_norm/beta:0": "tf_model-00005-of-00013.h5",
173
+ "tfopt_for_causal_lm/model/decoder/layers.18/final_layer_norm/gamma:0": "tf_model-00005-of-00013.h5",
174
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn/k_proj/bias:0": "tf_model-00005-of-00013.h5",
175
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn/k_proj/kernel:0": "tf_model-00005-of-00013.h5",
176
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn/out_proj/bias:0": "tf_model-00005-of-00013.h5",
177
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn/out_proj/kernel:0": "tf_model-00005-of-00013.h5",
178
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn/q_proj/bias:0": "tf_model-00005-of-00013.h5",
179
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn/q_proj/kernel:0": "tf_model-00005-of-00013.h5",
180
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn/v_proj/bias:0": "tf_model-00005-of-00013.h5",
181
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn/v_proj/kernel:0": "tf_model-00005-of-00013.h5",
182
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn_layer_norm/beta:0": "tf_model-00005-of-00013.h5",
183
+ "tfopt_for_causal_lm/model/decoder/layers.18/self_attn_layer_norm/gamma:0": "tf_model-00005-of-00013.h5",
184
+ "tfopt_for_causal_lm/model/decoder/layers.19/fc1/bias:0": "tf_model-00006-of-00013.h5",
185
+ "tfopt_for_causal_lm/model/decoder/layers.19/fc1/kernel:0": "tf_model-00006-of-00013.h5",
186
+ "tfopt_for_causal_lm/model/decoder/layers.19/fc2/bias:0": "tf_model-00006-of-00013.h5",
187
+ "tfopt_for_causal_lm/model/decoder/layers.19/fc2/kernel:0": "tf_model-00006-of-00013.h5",
188
+ "tfopt_for_causal_lm/model/decoder/layers.19/final_layer_norm/beta:0": "tf_model-00006-of-00013.h5",
189
+ "tfopt_for_causal_lm/model/decoder/layers.19/final_layer_norm/gamma:0": "tf_model-00006-of-00013.h5",
190
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn/k_proj/bias:0": "tf_model-00005-of-00013.h5",
191
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn/k_proj/kernel:0": "tf_model-00005-of-00013.h5",
192
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn/out_proj/bias:0": "tf_model-00005-of-00013.h5",
193
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn/out_proj/kernel:0": "tf_model-00005-of-00013.h5",
194
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn/q_proj/bias:0": "tf_model-00005-of-00013.h5",
195
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn/q_proj/kernel:0": "tf_model-00005-of-00013.h5",
196
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn/v_proj/bias:0": "tf_model-00005-of-00013.h5",
197
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn/v_proj/kernel:0": "tf_model-00005-of-00013.h5",
198
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn_layer_norm/beta:0": "tf_model-00005-of-00013.h5",
199
+ "tfopt_for_causal_lm/model/decoder/layers.19/self_attn_layer_norm/gamma:0": "tf_model-00005-of-00013.h5",
200
+ "tfopt_for_causal_lm/model/decoder/layers.2/fc1/bias:0": "tf_model-00001-of-00013.h5",
201
+ "tfopt_for_causal_lm/model/decoder/layers.2/fc1/kernel:0": "tf_model-00001-of-00013.h5",
202
+ "tfopt_for_causal_lm/model/decoder/layers.2/fc2/bias:0": "tf_model-00001-of-00013.h5",
203
+ "tfopt_for_causal_lm/model/decoder/layers.2/fc2/kernel:0": "tf_model-00001-of-00013.h5",
204
+ "tfopt_for_causal_lm/model/decoder/layers.2/final_layer_norm/beta:0": "tf_model-00001-of-00013.h5",
205
+ "tfopt_for_causal_lm/model/decoder/layers.2/final_layer_norm/gamma:0": "tf_model-00001-of-00013.h5",
206
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn/k_proj/bias:0": "tf_model-00001-of-00013.h5",
207
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn/k_proj/kernel:0": "tf_model-00001-of-00013.h5",
208
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn/out_proj/bias:0": "tf_model-00001-of-00013.h5",
209
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn/out_proj/kernel:0": "tf_model-00001-of-00013.h5",
210
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn/q_proj/bias:0": "tf_model-00001-of-00013.h5",
211
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn/q_proj/kernel:0": "tf_model-00001-of-00013.h5",
212
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn/v_proj/bias:0": "tf_model-00001-of-00013.h5",
213
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn/v_proj/kernel:0": "tf_model-00001-of-00013.h5",
214
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn_layer_norm/beta:0": "tf_model-00001-of-00013.h5",
215
+ "tfopt_for_causal_lm/model/decoder/layers.2/self_attn_layer_norm/gamma:0": "tf_model-00001-of-00013.h5",
216
+ "tfopt_for_causal_lm/model/decoder/layers.20/fc1/bias:0": "tf_model-00006-of-00013.h5",
217
+ "tfopt_for_causal_lm/model/decoder/layers.20/fc1/kernel:0": "tf_model-00006-of-00013.h5",
218
+ "tfopt_for_causal_lm/model/decoder/layers.20/fc2/bias:0": "tf_model-00006-of-00013.h5",
219
+ "tfopt_for_causal_lm/model/decoder/layers.20/fc2/kernel:0": "tf_model-00006-of-00013.h5",
220
+ "tfopt_for_causal_lm/model/decoder/layers.20/final_layer_norm/beta:0": "tf_model-00006-of-00013.h5",
221
+ "tfopt_for_causal_lm/model/decoder/layers.20/final_layer_norm/gamma:0": "tf_model-00006-of-00013.h5",
222
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn/k_proj/bias:0": "tf_model-00006-of-00013.h5",
223
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn/k_proj/kernel:0": "tf_model-00006-of-00013.h5",
224
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn/out_proj/bias:0": "tf_model-00006-of-00013.h5",
225
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn/out_proj/kernel:0": "tf_model-00006-of-00013.h5",
226
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn/q_proj/bias:0": "tf_model-00006-of-00013.h5",
227
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn/q_proj/kernel:0": "tf_model-00006-of-00013.h5",
228
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn/v_proj/bias:0": "tf_model-00006-of-00013.h5",
229
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn/v_proj/kernel:0": "tf_model-00006-of-00013.h5",
230
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn_layer_norm/beta:0": "tf_model-00006-of-00013.h5",
231
+ "tfopt_for_causal_lm/model/decoder/layers.20/self_attn_layer_norm/gamma:0": "tf_model-00006-of-00013.h5",
232
+ "tfopt_for_causal_lm/model/decoder/layers.21/fc1/bias:0": "tf_model-00006-of-00013.h5",
233
+ "tfopt_for_causal_lm/model/decoder/layers.21/fc1/kernel:0": "tf_model-00006-of-00013.h5",
234
+ "tfopt_for_causal_lm/model/decoder/layers.21/fc2/bias:0": "tf_model-00006-of-00013.h5",
235
+ "tfopt_for_causal_lm/model/decoder/layers.21/fc2/kernel:0": "tf_model-00006-of-00013.h5",
236
+ "tfopt_for_causal_lm/model/decoder/layers.21/final_layer_norm/beta:0": "tf_model-00006-of-00013.h5",
237
+ "tfopt_for_causal_lm/model/decoder/layers.21/final_layer_norm/gamma:0": "tf_model-00006-of-00013.h5",
238
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn/k_proj/bias:0": "tf_model-00006-of-00013.h5",
239
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn/k_proj/kernel:0": "tf_model-00006-of-00013.h5",
240
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn/out_proj/bias:0": "tf_model-00006-of-00013.h5",
241
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn/out_proj/kernel:0": "tf_model-00006-of-00013.h5",
242
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn/q_proj/bias:0": "tf_model-00006-of-00013.h5",
243
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn/q_proj/kernel:0": "tf_model-00006-of-00013.h5",
244
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn/v_proj/bias:0": "tf_model-00006-of-00013.h5",
245
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn/v_proj/kernel:0": "tf_model-00006-of-00013.h5",
246
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn_layer_norm/beta:0": "tf_model-00006-of-00013.h5",
247
+ "tfopt_for_causal_lm/model/decoder/layers.21/self_attn_layer_norm/gamma:0": "tf_model-00006-of-00013.h5",
248
+ "tfopt_for_causal_lm/model/decoder/layers.22/fc1/bias:0": "tf_model-00006-of-00013.h5",
249
+ "tfopt_for_causal_lm/model/decoder/layers.22/fc1/kernel:0": "tf_model-00006-of-00013.h5",
250
+ "tfopt_for_causal_lm/model/decoder/layers.22/fc2/bias:0": "tf_model-00006-of-00013.h5",
251
+ "tfopt_for_causal_lm/model/decoder/layers.22/fc2/kernel:0": "tf_model-00006-of-00013.h5",
252
+ "tfopt_for_causal_lm/model/decoder/layers.22/final_layer_norm/beta:0": "tf_model-00006-of-00013.h5",
253
+ "tfopt_for_causal_lm/model/decoder/layers.22/final_layer_norm/gamma:0": "tf_model-00006-of-00013.h5",
254
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn/k_proj/bias:0": "tf_model-00006-of-00013.h5",
255
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn/k_proj/kernel:0": "tf_model-00006-of-00013.h5",
256
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn/out_proj/bias:0": "tf_model-00006-of-00013.h5",
257
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn/out_proj/kernel:0": "tf_model-00006-of-00013.h5",
258
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn/q_proj/bias:0": "tf_model-00006-of-00013.h5",
259
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn/q_proj/kernel:0": "tf_model-00006-of-00013.h5",
260
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn/v_proj/bias:0": "tf_model-00006-of-00013.h5",
261
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn/v_proj/kernel:0": "tf_model-00006-of-00013.h5",
262
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn_layer_norm/beta:0": "tf_model-00006-of-00013.h5",
263
+ "tfopt_for_causal_lm/model/decoder/layers.22/self_attn_layer_norm/gamma:0": "tf_model-00006-of-00013.h5",
264
+ "tfopt_for_causal_lm/model/decoder/layers.23/fc1/bias:0": "tf_model-00007-of-00013.h5",
265
+ "tfopt_for_causal_lm/model/decoder/layers.23/fc1/kernel:0": "tf_model-00007-of-00013.h5",
266
+ "tfopt_for_causal_lm/model/decoder/layers.23/fc2/bias:0": "tf_model-00007-of-00013.h5",
267
+ "tfopt_for_causal_lm/model/decoder/layers.23/fc2/kernel:0": "tf_model-00007-of-00013.h5",
268
+ "tfopt_for_causal_lm/model/decoder/layers.23/final_layer_norm/beta:0": "tf_model-00007-of-00013.h5",
269
+ "tfopt_for_causal_lm/model/decoder/layers.23/final_layer_norm/gamma:0": "tf_model-00007-of-00013.h5",
270
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn/k_proj/bias:0": "tf_model-00006-of-00013.h5",
271
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn/k_proj/kernel:0": "tf_model-00006-of-00013.h5",
272
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn/out_proj/bias:0": "tf_model-00006-of-00013.h5",
273
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn/out_proj/kernel:0": "tf_model-00006-of-00013.h5",
274
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn/q_proj/bias:0": "tf_model-00006-of-00013.h5",
275
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn/q_proj/kernel:0": "tf_model-00006-of-00013.h5",
276
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn/v_proj/bias:0": "tf_model-00006-of-00013.h5",
277
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn/v_proj/kernel:0": "tf_model-00006-of-00013.h5",
278
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn_layer_norm/beta:0": "tf_model-00006-of-00013.h5",
279
+ "tfopt_for_causal_lm/model/decoder/layers.23/self_attn_layer_norm/gamma:0": "tf_model-00006-of-00013.h5",
280
+ "tfopt_for_causal_lm/model/decoder/layers.24/fc1/bias:0": "tf_model-00007-of-00013.h5",
281
+ "tfopt_for_causal_lm/model/decoder/layers.24/fc1/kernel:0": "tf_model-00007-of-00013.h5",
282
+ "tfopt_for_causal_lm/model/decoder/layers.24/fc2/bias:0": "tf_model-00007-of-00013.h5",
283
+ "tfopt_for_causal_lm/model/decoder/layers.24/fc2/kernel:0": "tf_model-00007-of-00013.h5",
284
+ "tfopt_for_causal_lm/model/decoder/layers.24/final_layer_norm/beta:0": "tf_model-00007-of-00013.h5",
285
+ "tfopt_for_causal_lm/model/decoder/layers.24/final_layer_norm/gamma:0": "tf_model-00007-of-00013.h5",
286
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn/k_proj/bias:0": "tf_model-00007-of-00013.h5",
287
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn/k_proj/kernel:0": "tf_model-00007-of-00013.h5",
288
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn/out_proj/bias:0": "tf_model-00007-of-00013.h5",
289
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn/out_proj/kernel:0": "tf_model-00007-of-00013.h5",
290
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn/q_proj/bias:0": "tf_model-00007-of-00013.h5",
291
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn/q_proj/kernel:0": "tf_model-00007-of-00013.h5",
292
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn/v_proj/bias:0": "tf_model-00007-of-00013.h5",
293
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn/v_proj/kernel:0": "tf_model-00007-of-00013.h5",
294
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn_layer_norm/beta:0": "tf_model-00007-of-00013.h5",
295
+ "tfopt_for_causal_lm/model/decoder/layers.24/self_attn_layer_norm/gamma:0": "tf_model-00007-of-00013.h5",
296
+ "tfopt_for_causal_lm/model/decoder/layers.25/fc1/bias:0": "tf_model-00007-of-00013.h5",
297
+ "tfopt_for_causal_lm/model/decoder/layers.25/fc1/kernel:0": "tf_model-00007-of-00013.h5",
298
+ "tfopt_for_causal_lm/model/decoder/layers.25/fc2/bias:0": "tf_model-00007-of-00013.h5",
299
+ "tfopt_for_causal_lm/model/decoder/layers.25/fc2/kernel:0": "tf_model-00007-of-00013.h5",
300
+ "tfopt_for_causal_lm/model/decoder/layers.25/final_layer_norm/beta:0": "tf_model-00007-of-00013.h5",
301
+ "tfopt_for_causal_lm/model/decoder/layers.25/final_layer_norm/gamma:0": "tf_model-00007-of-00013.h5",
302
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn/k_proj/bias:0": "tf_model-00007-of-00013.h5",
303
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn/k_proj/kernel:0": "tf_model-00007-of-00013.h5",
304
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn/out_proj/bias:0": "tf_model-00007-of-00013.h5",
305
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn/out_proj/kernel:0": "tf_model-00007-of-00013.h5",
306
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn/q_proj/bias:0": "tf_model-00007-of-00013.h5",
307
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn/q_proj/kernel:0": "tf_model-00007-of-00013.h5",
308
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn/v_proj/bias:0": "tf_model-00007-of-00013.h5",
309
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn/v_proj/kernel:0": "tf_model-00007-of-00013.h5",
310
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn_layer_norm/beta:0": "tf_model-00007-of-00013.h5",
311
+ "tfopt_for_causal_lm/model/decoder/layers.25/self_attn_layer_norm/gamma:0": "tf_model-00007-of-00013.h5",
312
+ "tfopt_for_causal_lm/model/decoder/layers.26/fc1/bias:0": "tf_model-00007-of-00013.h5",
313
+ "tfopt_for_causal_lm/model/decoder/layers.26/fc1/kernel:0": "tf_model-00007-of-00013.h5",
314
+ "tfopt_for_causal_lm/model/decoder/layers.26/fc2/bias:0": "tf_model-00007-of-00013.h5",
315
+ "tfopt_for_causal_lm/model/decoder/layers.26/fc2/kernel:0": "tf_model-00007-of-00013.h5",
316
+ "tfopt_for_causal_lm/model/decoder/layers.26/final_layer_norm/beta:0": "tf_model-00007-of-00013.h5",
317
+ "tfopt_for_causal_lm/model/decoder/layers.26/final_layer_norm/gamma:0": "tf_model-00007-of-00013.h5",
318
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn/k_proj/bias:0": "tf_model-00007-of-00013.h5",
319
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn/k_proj/kernel:0": "tf_model-00007-of-00013.h5",
320
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn/out_proj/bias:0": "tf_model-00007-of-00013.h5",
321
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn/out_proj/kernel:0": "tf_model-00007-of-00013.h5",
322
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn/q_proj/bias:0": "tf_model-00007-of-00013.h5",
323
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn/q_proj/kernel:0": "tf_model-00007-of-00013.h5",
324
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn/v_proj/bias:0": "tf_model-00007-of-00013.h5",
325
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn/v_proj/kernel:0": "tf_model-00007-of-00013.h5",
326
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn_layer_norm/beta:0": "tf_model-00007-of-00013.h5",
327
+ "tfopt_for_causal_lm/model/decoder/layers.26/self_attn_layer_norm/gamma:0": "tf_model-00007-of-00013.h5",
328
+ "tfopt_for_causal_lm/model/decoder/layers.27/fc1/bias:0": "tf_model-00008-of-00013.h5",
329
+ "tfopt_for_causal_lm/model/decoder/layers.27/fc1/kernel:0": "tf_model-00008-of-00013.h5",
330
+ "tfopt_for_causal_lm/model/decoder/layers.27/fc2/bias:0": "tf_model-00008-of-00013.h5",
331
+ "tfopt_for_causal_lm/model/decoder/layers.27/fc2/kernel:0": "tf_model-00008-of-00013.h5",
332
+ "tfopt_for_causal_lm/model/decoder/layers.27/final_layer_norm/beta:0": "tf_model-00008-of-00013.h5",
333
+ "tfopt_for_causal_lm/model/decoder/layers.27/final_layer_norm/gamma:0": "tf_model-00008-of-00013.h5",
334
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn/k_proj/bias:0": "tf_model-00007-of-00013.h5",
335
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn/k_proj/kernel:0": "tf_model-00007-of-00013.h5",
336
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn/out_proj/bias:0": "tf_model-00007-of-00013.h5",
337
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn/out_proj/kernel:0": "tf_model-00007-of-00013.h5",
338
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn/q_proj/bias:0": "tf_model-00007-of-00013.h5",
339
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn/q_proj/kernel:0": "tf_model-00007-of-00013.h5",
340
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn/v_proj/bias:0": "tf_model-00007-of-00013.h5",
341
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn/v_proj/kernel:0": "tf_model-00007-of-00013.h5",
342
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn_layer_norm/beta:0": "tf_model-00007-of-00013.h5",
343
+ "tfopt_for_causal_lm/model/decoder/layers.27/self_attn_layer_norm/gamma:0": "tf_model-00007-of-00013.h5",
344
+ "tfopt_for_causal_lm/model/decoder/layers.28/fc1/bias:0": "tf_model-00008-of-00013.h5",
345
+ "tfopt_for_causal_lm/model/decoder/layers.28/fc1/kernel:0": "tf_model-00008-of-00013.h5",
346
+ "tfopt_for_causal_lm/model/decoder/layers.28/fc2/bias:0": "tf_model-00008-of-00013.h5",
347
+ "tfopt_for_causal_lm/model/decoder/layers.28/fc2/kernel:0": "tf_model-00008-of-00013.h5",
348
+ "tfopt_for_causal_lm/model/decoder/layers.28/final_layer_norm/beta:0": "tf_model-00008-of-00013.h5",
349
+ "tfopt_for_causal_lm/model/decoder/layers.28/final_layer_norm/gamma:0": "tf_model-00008-of-00013.h5",
350
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn/k_proj/bias:0": "tf_model-00008-of-00013.h5",
351
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn/k_proj/kernel:0": "tf_model-00008-of-00013.h5",
352
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn/out_proj/bias:0": "tf_model-00008-of-00013.h5",
353
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn/out_proj/kernel:0": "tf_model-00008-of-00013.h5",
354
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn/q_proj/bias:0": "tf_model-00008-of-00013.h5",
355
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn/q_proj/kernel:0": "tf_model-00008-of-00013.h5",
356
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn/v_proj/bias:0": "tf_model-00008-of-00013.h5",
357
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn/v_proj/kernel:0": "tf_model-00008-of-00013.h5",
358
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn_layer_norm/beta:0": "tf_model-00008-of-00013.h5",
359
+ "tfopt_for_causal_lm/model/decoder/layers.28/self_attn_layer_norm/gamma:0": "tf_model-00008-of-00013.h5",
360
+ "tfopt_for_causal_lm/model/decoder/layers.29/fc1/bias:0": "tf_model-00008-of-00013.h5",
361
+ "tfopt_for_causal_lm/model/decoder/layers.29/fc1/kernel:0": "tf_model-00008-of-00013.h5",
362
+ "tfopt_for_causal_lm/model/decoder/layers.29/fc2/bias:0": "tf_model-00008-of-00013.h5",
363
+ "tfopt_for_causal_lm/model/decoder/layers.29/fc2/kernel:0": "tf_model-00008-of-00013.h5",
364
+ "tfopt_for_causal_lm/model/decoder/layers.29/final_layer_norm/beta:0": "tf_model-00008-of-00013.h5",
365
+ "tfopt_for_causal_lm/model/decoder/layers.29/final_layer_norm/gamma:0": "tf_model-00008-of-00013.h5",
366
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn/k_proj/bias:0": "tf_model-00008-of-00013.h5",
367
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn/k_proj/kernel:0": "tf_model-00008-of-00013.h5",
368
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn/out_proj/bias:0": "tf_model-00008-of-00013.h5",
369
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn/out_proj/kernel:0": "tf_model-00008-of-00013.h5",
370
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn/q_proj/bias:0": "tf_model-00008-of-00013.h5",
371
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn/q_proj/kernel:0": "tf_model-00008-of-00013.h5",
372
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn/v_proj/bias:0": "tf_model-00008-of-00013.h5",
373
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn/v_proj/kernel:0": "tf_model-00008-of-00013.h5",
374
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn_layer_norm/beta:0": "tf_model-00008-of-00013.h5",
375
+ "tfopt_for_causal_lm/model/decoder/layers.29/self_attn_layer_norm/gamma:0": "tf_model-00008-of-00013.h5",
376
+ "tfopt_for_causal_lm/model/decoder/layers.3/fc1/bias:0": "tf_model-00002-of-00013.h5",
377
+ "tfopt_for_causal_lm/model/decoder/layers.3/fc1/kernel:0": "tf_model-00002-of-00013.h5",
378
+ "tfopt_for_causal_lm/model/decoder/layers.3/fc2/bias:0": "tf_model-00002-of-00013.h5",
379
+ "tfopt_for_causal_lm/model/decoder/layers.3/fc2/kernel:0": "tf_model-00002-of-00013.h5",
380
+ "tfopt_for_causal_lm/model/decoder/layers.3/final_layer_norm/beta:0": "tf_model-00002-of-00013.h5",
381
+ "tfopt_for_causal_lm/model/decoder/layers.3/final_layer_norm/gamma:0": "tf_model-00002-of-00013.h5",
382
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn/k_proj/bias:0": "tf_model-00001-of-00013.h5",
383
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn/k_proj/kernel:0": "tf_model-00001-of-00013.h5",
384
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn/out_proj/bias:0": "tf_model-00001-of-00013.h5",
385
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn/out_proj/kernel:0": "tf_model-00001-of-00013.h5",
386
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn/q_proj/bias:0": "tf_model-00001-of-00013.h5",
387
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn/q_proj/kernel:0": "tf_model-00001-of-00013.h5",
388
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn/v_proj/bias:0": "tf_model-00001-of-00013.h5",
389
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn/v_proj/kernel:0": "tf_model-00001-of-00013.h5",
390
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn_layer_norm/beta:0": "tf_model-00001-of-00013.h5",
391
+ "tfopt_for_causal_lm/model/decoder/layers.3/self_attn_layer_norm/gamma:0": "tf_model-00001-of-00013.h5",
392
+ "tfopt_for_causal_lm/model/decoder/layers.30/fc1/bias:0": "tf_model-00008-of-00013.h5",
393
+ "tfopt_for_causal_lm/model/decoder/layers.30/fc1/kernel:0": "tf_model-00008-of-00013.h5",
394
+ "tfopt_for_causal_lm/model/decoder/layers.30/fc2/bias:0": "tf_model-00008-of-00013.h5",
395
+ "tfopt_for_causal_lm/model/decoder/layers.30/fc2/kernel:0": "tf_model-00008-of-00013.h5",
396
+ "tfopt_for_causal_lm/model/decoder/layers.30/final_layer_norm/beta:0": "tf_model-00008-of-00013.h5",
397
+ "tfopt_for_causal_lm/model/decoder/layers.30/final_layer_norm/gamma:0": "tf_model-00008-of-00013.h5",
398
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn/k_proj/bias:0": "tf_model-00008-of-00013.h5",
399
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn/k_proj/kernel:0": "tf_model-00008-of-00013.h5",
400
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn/out_proj/bias:0": "tf_model-00008-of-00013.h5",
401
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn/out_proj/kernel:0": "tf_model-00008-of-00013.h5",
402
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn/q_proj/bias:0": "tf_model-00008-of-00013.h5",
403
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn/q_proj/kernel:0": "tf_model-00008-of-00013.h5",
404
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn/v_proj/bias:0": "tf_model-00008-of-00013.h5",
405
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn/v_proj/kernel:0": "tf_model-00008-of-00013.h5",
406
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn_layer_norm/beta:0": "tf_model-00008-of-00013.h5",
407
+ "tfopt_for_causal_lm/model/decoder/layers.30/self_attn_layer_norm/gamma:0": "tf_model-00008-of-00013.h5",
408
+ "tfopt_for_causal_lm/model/decoder/layers.31/fc1/bias:0": "tf_model-00009-of-00013.h5",
409
+ "tfopt_for_causal_lm/model/decoder/layers.31/fc1/kernel:0": "tf_model-00009-of-00013.h5",
410
+ "tfopt_for_causal_lm/model/decoder/layers.31/fc2/bias:0": "tf_model-00009-of-00013.h5",
411
+ "tfopt_for_causal_lm/model/decoder/layers.31/fc2/kernel:0": "tf_model-00009-of-00013.h5",
412
+ "tfopt_for_causal_lm/model/decoder/layers.31/final_layer_norm/beta:0": "tf_model-00009-of-00013.h5",
413
+ "tfopt_for_causal_lm/model/decoder/layers.31/final_layer_norm/gamma:0": "tf_model-00009-of-00013.h5",
414
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn/k_proj/bias:0": "tf_model-00008-of-00013.h5",
415
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn/k_proj/kernel:0": "tf_model-00008-of-00013.h5",
416
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn/out_proj/bias:0": "tf_model-00008-of-00013.h5",
417
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn/out_proj/kernel:0": "tf_model-00008-of-00013.h5",
418
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn/q_proj/bias:0": "tf_model-00008-of-00013.h5",
419
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn/q_proj/kernel:0": "tf_model-00008-of-00013.h5",
420
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn/v_proj/bias:0": "tf_model-00008-of-00013.h5",
421
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn/v_proj/kernel:0": "tf_model-00008-of-00013.h5",
422
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn_layer_norm/beta:0": "tf_model-00008-of-00013.h5",
423
+ "tfopt_for_causal_lm/model/decoder/layers.31/self_attn_layer_norm/gamma:0": "tf_model-00008-of-00013.h5",
424
+ "tfopt_for_causal_lm/model/decoder/layers.32/fc1/bias:0": "tf_model-00009-of-00013.h5",
425
+ "tfopt_for_causal_lm/model/decoder/layers.32/fc1/kernel:0": "tf_model-00009-of-00013.h5",
426
+ "tfopt_for_causal_lm/model/decoder/layers.32/fc2/bias:0": "tf_model-00009-of-00013.h5",
427
+ "tfopt_for_causal_lm/model/decoder/layers.32/fc2/kernel:0": "tf_model-00009-of-00013.h5",
428
+ "tfopt_for_causal_lm/model/decoder/layers.32/final_layer_norm/beta:0": "tf_model-00009-of-00013.h5",
429
+ "tfopt_for_causal_lm/model/decoder/layers.32/final_layer_norm/gamma:0": "tf_model-00009-of-00013.h5",
430
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn/k_proj/bias:0": "tf_model-00009-of-00013.h5",
431
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn/k_proj/kernel:0": "tf_model-00009-of-00013.h5",
432
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn/out_proj/bias:0": "tf_model-00009-of-00013.h5",
433
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn/out_proj/kernel:0": "tf_model-00009-of-00013.h5",
434
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn/q_proj/bias:0": "tf_model-00009-of-00013.h5",
435
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn/q_proj/kernel:0": "tf_model-00009-of-00013.h5",
436
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn/v_proj/bias:0": "tf_model-00009-of-00013.h5",
437
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn/v_proj/kernel:0": "tf_model-00009-of-00013.h5",
438
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn_layer_norm/beta:0": "tf_model-00009-of-00013.h5",
439
+ "tfopt_for_causal_lm/model/decoder/layers.32/self_attn_layer_norm/gamma:0": "tf_model-00009-of-00013.h5",
440
+ "tfopt_for_causal_lm/model/decoder/layers.33/fc1/bias:0": "tf_model-00009-of-00013.h5",
441
+ "tfopt_for_causal_lm/model/decoder/layers.33/fc1/kernel:0": "tf_model-00009-of-00013.h5",
442
+ "tfopt_for_causal_lm/model/decoder/layers.33/fc2/bias:0": "tf_model-00009-of-00013.h5",
443
+ "tfopt_for_causal_lm/model/decoder/layers.33/fc2/kernel:0": "tf_model-00009-of-00013.h5",
444
+ "tfopt_for_causal_lm/model/decoder/layers.33/final_layer_norm/beta:0": "tf_model-00009-of-00013.h5",
445
+ "tfopt_for_causal_lm/model/decoder/layers.33/final_layer_norm/gamma:0": "tf_model-00009-of-00013.h5",
446
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn/k_proj/bias:0": "tf_model-00009-of-00013.h5",
447
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn/k_proj/kernel:0": "tf_model-00009-of-00013.h5",
448
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn/out_proj/bias:0": "tf_model-00009-of-00013.h5",
449
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn/out_proj/kernel:0": "tf_model-00009-of-00013.h5",
450
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn/q_proj/bias:0": "tf_model-00009-of-00013.h5",
451
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn/q_proj/kernel:0": "tf_model-00009-of-00013.h5",
452
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn/v_proj/bias:0": "tf_model-00009-of-00013.h5",
453
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn/v_proj/kernel:0": "tf_model-00009-of-00013.h5",
454
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn_layer_norm/beta:0": "tf_model-00009-of-00013.h5",
455
+ "tfopt_for_causal_lm/model/decoder/layers.33/self_attn_layer_norm/gamma:0": "tf_model-00009-of-00013.h5",
456
+ "tfopt_for_causal_lm/model/decoder/layers.34/fc1/bias:0": "tf_model-00009-of-00013.h5",
457
+ "tfopt_for_causal_lm/model/decoder/layers.34/fc1/kernel:0": "tf_model-00009-of-00013.h5",
458
+ "tfopt_for_causal_lm/model/decoder/layers.34/fc2/bias:0": "tf_model-00009-of-00013.h5",
459
+ "tfopt_for_causal_lm/model/decoder/layers.34/fc2/kernel:0": "tf_model-00009-of-00013.h5",
460
+ "tfopt_for_causal_lm/model/decoder/layers.34/final_layer_norm/beta:0": "tf_model-00009-of-00013.h5",
461
+ "tfopt_for_causal_lm/model/decoder/layers.34/final_layer_norm/gamma:0": "tf_model-00009-of-00013.h5",
462
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn/k_proj/bias:0": "tf_model-00009-of-00013.h5",
463
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn/k_proj/kernel:0": "tf_model-00009-of-00013.h5",
464
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn/out_proj/bias:0": "tf_model-00009-of-00013.h5",
465
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn/out_proj/kernel:0": "tf_model-00009-of-00013.h5",
466
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn/q_proj/bias:0": "tf_model-00009-of-00013.h5",
467
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn/q_proj/kernel:0": "tf_model-00009-of-00013.h5",
468
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn/v_proj/bias:0": "tf_model-00009-of-00013.h5",
469
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn/v_proj/kernel:0": "tf_model-00009-of-00013.h5",
470
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn_layer_norm/beta:0": "tf_model-00009-of-00013.h5",
471
+ "tfopt_for_causal_lm/model/decoder/layers.34/self_attn_layer_norm/gamma:0": "tf_model-00009-of-00013.h5",
472
+ "tfopt_for_causal_lm/model/decoder/layers.35/fc1/bias:0": "tf_model-00010-of-00013.h5",
473
+ "tfopt_for_causal_lm/model/decoder/layers.35/fc1/kernel:0": "tf_model-00010-of-00013.h5",
474
+ "tfopt_for_causal_lm/model/decoder/layers.35/fc2/bias:0": "tf_model-00010-of-00013.h5",
475
+ "tfopt_for_causal_lm/model/decoder/layers.35/fc2/kernel:0": "tf_model-00010-of-00013.h5",
476
+ "tfopt_for_causal_lm/model/decoder/layers.35/final_layer_norm/beta:0": "tf_model-00010-of-00013.h5",
477
+ "tfopt_for_causal_lm/model/decoder/layers.35/final_layer_norm/gamma:0": "tf_model-00010-of-00013.h5",
478
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn/k_proj/bias:0": "tf_model-00009-of-00013.h5",
479
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn/k_proj/kernel:0": "tf_model-00009-of-00013.h5",
480
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn/out_proj/bias:0": "tf_model-00009-of-00013.h5",
481
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn/out_proj/kernel:0": "tf_model-00009-of-00013.h5",
482
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn/q_proj/bias:0": "tf_model-00009-of-00013.h5",
483
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn/q_proj/kernel:0": "tf_model-00009-of-00013.h5",
484
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn/v_proj/bias:0": "tf_model-00009-of-00013.h5",
485
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn/v_proj/kernel:0": "tf_model-00009-of-00013.h5",
486
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn_layer_norm/beta:0": "tf_model-00009-of-00013.h5",
487
+ "tfopt_for_causal_lm/model/decoder/layers.35/self_attn_layer_norm/gamma:0": "tf_model-00009-of-00013.h5",
488
+ "tfopt_for_causal_lm/model/decoder/layers.36/fc1/bias:0": "tf_model-00010-of-00013.h5",
489
+ "tfopt_for_causal_lm/model/decoder/layers.36/fc1/kernel:0": "tf_model-00010-of-00013.h5",
490
+ "tfopt_for_causal_lm/model/decoder/layers.36/fc2/bias:0": "tf_model-00010-of-00013.h5",
491
+ "tfopt_for_causal_lm/model/decoder/layers.36/fc2/kernel:0": "tf_model-00010-of-00013.h5",
492
+ "tfopt_for_causal_lm/model/decoder/layers.36/final_layer_norm/beta:0": "tf_model-00010-of-00013.h5",
493
+ "tfopt_for_causal_lm/model/decoder/layers.36/final_layer_norm/gamma:0": "tf_model-00010-of-00013.h5",
494
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn/k_proj/bias:0": "tf_model-00010-of-00013.h5",
495
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn/k_proj/kernel:0": "tf_model-00010-of-00013.h5",
496
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn/out_proj/bias:0": "tf_model-00010-of-00013.h5",
497
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn/out_proj/kernel:0": "tf_model-00010-of-00013.h5",
498
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn/q_proj/bias:0": "tf_model-00010-of-00013.h5",
499
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn/q_proj/kernel:0": "tf_model-00010-of-00013.h5",
500
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn/v_proj/bias:0": "tf_model-00010-of-00013.h5",
501
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn/v_proj/kernel:0": "tf_model-00010-of-00013.h5",
502
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn_layer_norm/beta:0": "tf_model-00010-of-00013.h5",
503
+ "tfopt_for_causal_lm/model/decoder/layers.36/self_attn_layer_norm/gamma:0": "tf_model-00010-of-00013.h5",
504
+ "tfopt_for_causal_lm/model/decoder/layers.37/fc1/bias:0": "tf_model-00010-of-00013.h5",
505
+ "tfopt_for_causal_lm/model/decoder/layers.37/fc1/kernel:0": "tf_model-00010-of-00013.h5",
506
+ "tfopt_for_causal_lm/model/decoder/layers.37/fc2/bias:0": "tf_model-00010-of-00013.h5",
507
+ "tfopt_for_causal_lm/model/decoder/layers.37/fc2/kernel:0": "tf_model-00010-of-00013.h5",
508
+ "tfopt_for_causal_lm/model/decoder/layers.37/final_layer_norm/beta:0": "tf_model-00010-of-00013.h5",
509
+ "tfopt_for_causal_lm/model/decoder/layers.37/final_layer_norm/gamma:0": "tf_model-00010-of-00013.h5",
510
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn/k_proj/bias:0": "tf_model-00010-of-00013.h5",
511
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn/k_proj/kernel:0": "tf_model-00010-of-00013.h5",
512
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn/out_proj/bias:0": "tf_model-00010-of-00013.h5",
513
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn/out_proj/kernel:0": "tf_model-00010-of-00013.h5",
514
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn/q_proj/bias:0": "tf_model-00010-of-00013.h5",
515
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn/q_proj/kernel:0": "tf_model-00010-of-00013.h5",
516
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn/v_proj/bias:0": "tf_model-00010-of-00013.h5",
517
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn/v_proj/kernel:0": "tf_model-00010-of-00013.h5",
518
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn_layer_norm/beta:0": "tf_model-00010-of-00013.h5",
519
+ "tfopt_for_causal_lm/model/decoder/layers.37/self_attn_layer_norm/gamma:0": "tf_model-00010-of-00013.h5",
520
+ "tfopt_for_causal_lm/model/decoder/layers.38/fc1/bias:0": "tf_model-00010-of-00013.h5",
521
+ "tfopt_for_causal_lm/model/decoder/layers.38/fc1/kernel:0": "tf_model-00010-of-00013.h5",
522
+ "tfopt_for_causal_lm/model/decoder/layers.38/fc2/bias:0": "tf_model-00010-of-00013.h5",
523
+ "tfopt_for_causal_lm/model/decoder/layers.38/fc2/kernel:0": "tf_model-00010-of-00013.h5",
524
+ "tfopt_for_causal_lm/model/decoder/layers.38/final_layer_norm/beta:0": "tf_model-00010-of-00013.h5",
525
+ "tfopt_for_causal_lm/model/decoder/layers.38/final_layer_norm/gamma:0": "tf_model-00010-of-00013.h5",
526
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn/k_proj/bias:0": "tf_model-00010-of-00013.h5",
527
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn/k_proj/kernel:0": "tf_model-00010-of-00013.h5",
528
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn/out_proj/bias:0": "tf_model-00010-of-00013.h5",
529
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn/out_proj/kernel:0": "tf_model-00010-of-00013.h5",
530
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn/q_proj/bias:0": "tf_model-00010-of-00013.h5",
531
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn/q_proj/kernel:0": "tf_model-00010-of-00013.h5",
532
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn/v_proj/bias:0": "tf_model-00010-of-00013.h5",
533
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn/v_proj/kernel:0": "tf_model-00010-of-00013.h5",
534
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn_layer_norm/beta:0": "tf_model-00010-of-00013.h5",
535
+ "tfopt_for_causal_lm/model/decoder/layers.38/self_attn_layer_norm/gamma:0": "tf_model-00010-of-00013.h5",
536
+ "tfopt_for_causal_lm/model/decoder/layers.39/fc1/bias:0": "tf_model-00011-of-00013.h5",
537
+ "tfopt_for_causal_lm/model/decoder/layers.39/fc1/kernel:0": "tf_model-00011-of-00013.h5",
538
+ "tfopt_for_causal_lm/model/decoder/layers.39/fc2/bias:0": "tf_model-00011-of-00013.h5",
539
+ "tfopt_for_causal_lm/model/decoder/layers.39/fc2/kernel:0": "tf_model-00011-of-00013.h5",
540
+ "tfopt_for_causal_lm/model/decoder/layers.39/final_layer_norm/beta:0": "tf_model-00011-of-00013.h5",
541
+ "tfopt_for_causal_lm/model/decoder/layers.39/final_layer_norm/gamma:0": "tf_model-00011-of-00013.h5",
542
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn/k_proj/bias:0": "tf_model-00010-of-00013.h5",
543
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn/k_proj/kernel:0": "tf_model-00010-of-00013.h5",
544
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn/out_proj/bias:0": "tf_model-00010-of-00013.h5",
545
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn/out_proj/kernel:0": "tf_model-00010-of-00013.h5",
546
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn/q_proj/bias:0": "tf_model-00010-of-00013.h5",
547
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn/q_proj/kernel:0": "tf_model-00010-of-00013.h5",
548
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn/v_proj/bias:0": "tf_model-00010-of-00013.h5",
549
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn/v_proj/kernel:0": "tf_model-00010-of-00013.h5",
550
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn_layer_norm/beta:0": "tf_model-00010-of-00013.h5",
551
+ "tfopt_for_causal_lm/model/decoder/layers.39/self_attn_layer_norm/gamma:0": "tf_model-00010-of-00013.h5",
552
+ "tfopt_for_causal_lm/model/decoder/layers.4/fc1/bias:0": "tf_model-00002-of-00013.h5",
553
+ "tfopt_for_causal_lm/model/decoder/layers.4/fc1/kernel:0": "tf_model-00002-of-00013.h5",
554
+ "tfopt_for_causal_lm/model/decoder/layers.4/fc2/bias:0": "tf_model-00002-of-00013.h5",
555
+ "tfopt_for_causal_lm/model/decoder/layers.4/fc2/kernel:0": "tf_model-00002-of-00013.h5",
556
+ "tfopt_for_causal_lm/model/decoder/layers.4/final_layer_norm/beta:0": "tf_model-00002-of-00013.h5",
557
+ "tfopt_for_causal_lm/model/decoder/layers.4/final_layer_norm/gamma:0": "tf_model-00002-of-00013.h5",
558
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn/k_proj/bias:0": "tf_model-00002-of-00013.h5",
559
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn/k_proj/kernel:0": "tf_model-00002-of-00013.h5",
560
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn/out_proj/bias:0": "tf_model-00002-of-00013.h5",
561
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn/out_proj/kernel:0": "tf_model-00002-of-00013.h5",
562
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn/q_proj/bias:0": "tf_model-00002-of-00013.h5",
563
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn/q_proj/kernel:0": "tf_model-00002-of-00013.h5",
564
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn/v_proj/bias:0": "tf_model-00002-of-00013.h5",
565
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn/v_proj/kernel:0": "tf_model-00002-of-00013.h5",
566
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn_layer_norm/beta:0": "tf_model-00002-of-00013.h5",
567
+ "tfopt_for_causal_lm/model/decoder/layers.4/self_attn_layer_norm/gamma:0": "tf_model-00002-of-00013.h5",
568
+ "tfopt_for_causal_lm/model/decoder/layers.40/fc1/bias:0": "tf_model-00011-of-00013.h5",
569
+ "tfopt_for_causal_lm/model/decoder/layers.40/fc1/kernel:0": "tf_model-00011-of-00013.h5",
570
+ "tfopt_for_causal_lm/model/decoder/layers.40/fc2/bias:0": "tf_model-00011-of-00013.h5",
571
+ "tfopt_for_causal_lm/model/decoder/layers.40/fc2/kernel:0": "tf_model-00011-of-00013.h5",
572
+ "tfopt_for_causal_lm/model/decoder/layers.40/final_layer_norm/beta:0": "tf_model-00011-of-00013.h5",
573
+ "tfopt_for_causal_lm/model/decoder/layers.40/final_layer_norm/gamma:0": "tf_model-00011-of-00013.h5",
574
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn/k_proj/bias:0": "tf_model-00011-of-00013.h5",
575
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn/k_proj/kernel:0": "tf_model-00011-of-00013.h5",
576
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn/out_proj/bias:0": "tf_model-00011-of-00013.h5",
577
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn/out_proj/kernel:0": "tf_model-00011-of-00013.h5",
578
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn/q_proj/bias:0": "tf_model-00011-of-00013.h5",
579
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn/q_proj/kernel:0": "tf_model-00011-of-00013.h5",
580
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn/v_proj/bias:0": "tf_model-00011-of-00013.h5",
581
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn/v_proj/kernel:0": "tf_model-00011-of-00013.h5",
582
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn_layer_norm/beta:0": "tf_model-00011-of-00013.h5",
583
+ "tfopt_for_causal_lm/model/decoder/layers.40/self_attn_layer_norm/gamma:0": "tf_model-00011-of-00013.h5",
584
+ "tfopt_for_causal_lm/model/decoder/layers.41/fc1/bias:0": "tf_model-00011-of-00013.h5",
585
+ "tfopt_for_causal_lm/model/decoder/layers.41/fc1/kernel:0": "tf_model-00011-of-00013.h5",
586
+ "tfopt_for_causal_lm/model/decoder/layers.41/fc2/bias:0": "tf_model-00011-of-00013.h5",
587
+ "tfopt_for_causal_lm/model/decoder/layers.41/fc2/kernel:0": "tf_model-00011-of-00013.h5",
588
+ "tfopt_for_causal_lm/model/decoder/layers.41/final_layer_norm/beta:0": "tf_model-00011-of-00013.h5",
589
+ "tfopt_for_causal_lm/model/decoder/layers.41/final_layer_norm/gamma:0": "tf_model-00011-of-00013.h5",
590
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn/k_proj/bias:0": "tf_model-00011-of-00013.h5",
591
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn/k_proj/kernel:0": "tf_model-00011-of-00013.h5",
592
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn/out_proj/bias:0": "tf_model-00011-of-00013.h5",
593
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn/out_proj/kernel:0": "tf_model-00011-of-00013.h5",
594
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn/q_proj/bias:0": "tf_model-00011-of-00013.h5",
595
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn/q_proj/kernel:0": "tf_model-00011-of-00013.h5",
596
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn/v_proj/bias:0": "tf_model-00011-of-00013.h5",
597
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn/v_proj/kernel:0": "tf_model-00011-of-00013.h5",
598
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn_layer_norm/beta:0": "tf_model-00011-of-00013.h5",
599
+ "tfopt_for_causal_lm/model/decoder/layers.41/self_attn_layer_norm/gamma:0": "tf_model-00011-of-00013.h5",
600
+ "tfopt_for_causal_lm/model/decoder/layers.42/fc1/bias:0": "tf_model-00011-of-00013.h5",
601
+ "tfopt_for_causal_lm/model/decoder/layers.42/fc1/kernel:0": "tf_model-00011-of-00013.h5",
602
+ "tfopt_for_causal_lm/model/decoder/layers.42/fc2/bias:0": "tf_model-00011-of-00013.h5",
603
+ "tfopt_for_causal_lm/model/decoder/layers.42/fc2/kernel:0": "tf_model-00011-of-00013.h5",
604
+ "tfopt_for_causal_lm/model/decoder/layers.42/final_layer_norm/beta:0": "tf_model-00011-of-00013.h5",
605
+ "tfopt_for_causal_lm/model/decoder/layers.42/final_layer_norm/gamma:0": "tf_model-00011-of-00013.h5",
606
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn/k_proj/bias:0": "tf_model-00011-of-00013.h5",
607
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn/k_proj/kernel:0": "tf_model-00011-of-00013.h5",
608
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn/out_proj/bias:0": "tf_model-00011-of-00013.h5",
609
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn/out_proj/kernel:0": "tf_model-00011-of-00013.h5",
610
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn/q_proj/bias:0": "tf_model-00011-of-00013.h5",
611
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn/q_proj/kernel:0": "tf_model-00011-of-00013.h5",
612
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn/v_proj/bias:0": "tf_model-00011-of-00013.h5",
613
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn/v_proj/kernel:0": "tf_model-00011-of-00013.h5",
614
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn_layer_norm/beta:0": "tf_model-00011-of-00013.h5",
615
+ "tfopt_for_causal_lm/model/decoder/layers.42/self_attn_layer_norm/gamma:0": "tf_model-00011-of-00013.h5",
616
+ "tfopt_for_causal_lm/model/decoder/layers.43/fc1/bias:0": "tf_model-00012-of-00013.h5",
617
+ "tfopt_for_causal_lm/model/decoder/layers.43/fc1/kernel:0": "tf_model-00012-of-00013.h5",
618
+ "tfopt_for_causal_lm/model/decoder/layers.43/fc2/bias:0": "tf_model-00012-of-00013.h5",
619
+ "tfopt_for_causal_lm/model/decoder/layers.43/fc2/kernel:0": "tf_model-00012-of-00013.h5",
620
+ "tfopt_for_causal_lm/model/decoder/layers.43/final_layer_norm/beta:0": "tf_model-00012-of-00013.h5",
621
+ "tfopt_for_causal_lm/model/decoder/layers.43/final_layer_norm/gamma:0": "tf_model-00012-of-00013.h5",
622
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn/k_proj/bias:0": "tf_model-00011-of-00013.h5",
623
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn/k_proj/kernel:0": "tf_model-00011-of-00013.h5",
624
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn/out_proj/bias:0": "tf_model-00011-of-00013.h5",
625
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn/out_proj/kernel:0": "tf_model-00011-of-00013.h5",
626
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn/q_proj/bias:0": "tf_model-00011-of-00013.h5",
627
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn/q_proj/kernel:0": "tf_model-00011-of-00013.h5",
628
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn/v_proj/bias:0": "tf_model-00011-of-00013.h5",
629
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn/v_proj/kernel:0": "tf_model-00011-of-00013.h5",
630
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn_layer_norm/beta:0": "tf_model-00011-of-00013.h5",
631
+ "tfopt_for_causal_lm/model/decoder/layers.43/self_attn_layer_norm/gamma:0": "tf_model-00011-of-00013.h5",
632
+ "tfopt_for_causal_lm/model/decoder/layers.44/fc1/bias:0": "tf_model-00012-of-00013.h5",
633
+ "tfopt_for_causal_lm/model/decoder/layers.44/fc1/kernel:0": "tf_model-00012-of-00013.h5",
634
+ "tfopt_for_causal_lm/model/decoder/layers.44/fc2/bias:0": "tf_model-00012-of-00013.h5",
635
+ "tfopt_for_causal_lm/model/decoder/layers.44/fc2/kernel:0": "tf_model-00012-of-00013.h5",
636
+ "tfopt_for_causal_lm/model/decoder/layers.44/final_layer_norm/beta:0": "tf_model-00012-of-00013.h5",
637
+ "tfopt_for_causal_lm/model/decoder/layers.44/final_layer_norm/gamma:0": "tf_model-00012-of-00013.h5",
638
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn/k_proj/bias:0": "tf_model-00012-of-00013.h5",
639
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn/k_proj/kernel:0": "tf_model-00012-of-00013.h5",
640
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn/out_proj/bias:0": "tf_model-00012-of-00013.h5",
641
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn/out_proj/kernel:0": "tf_model-00012-of-00013.h5",
642
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn/q_proj/bias:0": "tf_model-00012-of-00013.h5",
643
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn/q_proj/kernel:0": "tf_model-00012-of-00013.h5",
644
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn/v_proj/bias:0": "tf_model-00012-of-00013.h5",
645
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn/v_proj/kernel:0": "tf_model-00012-of-00013.h5",
646
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn_layer_norm/beta:0": "tf_model-00012-of-00013.h5",
647
+ "tfopt_for_causal_lm/model/decoder/layers.44/self_attn_layer_norm/gamma:0": "tf_model-00012-of-00013.h5",
648
+ "tfopt_for_causal_lm/model/decoder/layers.45/fc1/bias:0": "tf_model-00012-of-00013.h5",
649
+ "tfopt_for_causal_lm/model/decoder/layers.45/fc1/kernel:0": "tf_model-00012-of-00013.h5",
650
+ "tfopt_for_causal_lm/model/decoder/layers.45/fc2/bias:0": "tf_model-00012-of-00013.h5",
651
+ "tfopt_for_causal_lm/model/decoder/layers.45/fc2/kernel:0": "tf_model-00012-of-00013.h5",
652
+ "tfopt_for_causal_lm/model/decoder/layers.45/final_layer_norm/beta:0": "tf_model-00012-of-00013.h5",
653
+ "tfopt_for_causal_lm/model/decoder/layers.45/final_layer_norm/gamma:0": "tf_model-00012-of-00013.h5",
654
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn/k_proj/bias:0": "tf_model-00012-of-00013.h5",
655
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn/k_proj/kernel:0": "tf_model-00012-of-00013.h5",
656
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn/out_proj/bias:0": "tf_model-00012-of-00013.h5",
657
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn/out_proj/kernel:0": "tf_model-00012-of-00013.h5",
658
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn/q_proj/bias:0": "tf_model-00012-of-00013.h5",
659
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn/q_proj/kernel:0": "tf_model-00012-of-00013.h5",
660
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn/v_proj/bias:0": "tf_model-00012-of-00013.h5",
661
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn/v_proj/kernel:0": "tf_model-00012-of-00013.h5",
662
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn_layer_norm/beta:0": "tf_model-00012-of-00013.h5",
663
+ "tfopt_for_causal_lm/model/decoder/layers.45/self_attn_layer_norm/gamma:0": "tf_model-00012-of-00013.h5",
664
+ "tfopt_for_causal_lm/model/decoder/layers.46/fc1/bias:0": "tf_model-00012-of-00013.h5",
665
+ "tfopt_for_causal_lm/model/decoder/layers.46/fc1/kernel:0": "tf_model-00012-of-00013.h5",
666
+ "tfopt_for_causal_lm/model/decoder/layers.46/fc2/bias:0": "tf_model-00012-of-00013.h5",
667
+ "tfopt_for_causal_lm/model/decoder/layers.46/fc2/kernel:0": "tf_model-00012-of-00013.h5",
668
+ "tfopt_for_causal_lm/model/decoder/layers.46/final_layer_norm/beta:0": "tf_model-00012-of-00013.h5",
669
+ "tfopt_for_causal_lm/model/decoder/layers.46/final_layer_norm/gamma:0": "tf_model-00012-of-00013.h5",
670
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn/k_proj/bias:0": "tf_model-00012-of-00013.h5",
671
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn/k_proj/kernel:0": "tf_model-00012-of-00013.h5",
672
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn/out_proj/bias:0": "tf_model-00012-of-00013.h5",
673
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn/out_proj/kernel:0": "tf_model-00012-of-00013.h5",
674
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn/q_proj/bias:0": "tf_model-00012-of-00013.h5",
675
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn/q_proj/kernel:0": "tf_model-00012-of-00013.h5",
676
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn/v_proj/bias:0": "tf_model-00012-of-00013.h5",
677
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn/v_proj/kernel:0": "tf_model-00012-of-00013.h5",
678
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn_layer_norm/beta:0": "tf_model-00012-of-00013.h5",
679
+ "tfopt_for_causal_lm/model/decoder/layers.46/self_attn_layer_norm/gamma:0": "tf_model-00012-of-00013.h5",
680
+ "tfopt_for_causal_lm/model/decoder/layers.47/fc1/bias:0": "tf_model-00013-of-00013.h5",
681
+ "tfopt_for_causal_lm/model/decoder/layers.47/fc1/kernel:0": "tf_model-00013-of-00013.h5",
682
+ "tfopt_for_causal_lm/model/decoder/layers.47/fc2/bias:0": "tf_model-00013-of-00013.h5",
683
+ "tfopt_for_causal_lm/model/decoder/layers.47/fc2/kernel:0": "tf_model-00013-of-00013.h5",
684
+ "tfopt_for_causal_lm/model/decoder/layers.47/final_layer_norm/beta:0": "tf_model-00013-of-00013.h5",
685
+ "tfopt_for_causal_lm/model/decoder/layers.47/final_layer_norm/gamma:0": "tf_model-00013-of-00013.h5",
686
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn/k_proj/bias:0": "tf_model-00012-of-00013.h5",
687
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn/k_proj/kernel:0": "tf_model-00012-of-00013.h5",
688
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn/out_proj/bias:0": "tf_model-00012-of-00013.h5",
689
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn/out_proj/kernel:0": "tf_model-00012-of-00013.h5",
690
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn/q_proj/bias:0": "tf_model-00012-of-00013.h5",
691
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn/q_proj/kernel:0": "tf_model-00012-of-00013.h5",
692
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn/v_proj/bias:0": "tf_model-00012-of-00013.h5",
693
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn/v_proj/kernel:0": "tf_model-00012-of-00013.h5",
694
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn_layer_norm/beta:0": "tf_model-00012-of-00013.h5",
695
+ "tfopt_for_causal_lm/model/decoder/layers.47/self_attn_layer_norm/gamma:0": "tf_model-00012-of-00013.h5",
696
+ "tfopt_for_causal_lm/model/decoder/layers.5/fc1/bias:0": "tf_model-00002-of-00013.h5",
697
+ "tfopt_for_causal_lm/model/decoder/layers.5/fc1/kernel:0": "tf_model-00002-of-00013.h5",
698
+ "tfopt_for_causal_lm/model/decoder/layers.5/fc2/bias:0": "tf_model-00002-of-00013.h5",
699
+ "tfopt_for_causal_lm/model/decoder/layers.5/fc2/kernel:0": "tf_model-00002-of-00013.h5",
700
+ "tfopt_for_causal_lm/model/decoder/layers.5/final_layer_norm/beta:0": "tf_model-00002-of-00013.h5",
701
+ "tfopt_for_causal_lm/model/decoder/layers.5/final_layer_norm/gamma:0": "tf_model-00002-of-00013.h5",
702
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn/k_proj/bias:0": "tf_model-00002-of-00013.h5",
703
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn/k_proj/kernel:0": "tf_model-00002-of-00013.h5",
704
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn/out_proj/bias:0": "tf_model-00002-of-00013.h5",
705
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn/out_proj/kernel:0": "tf_model-00002-of-00013.h5",
706
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn/q_proj/bias:0": "tf_model-00002-of-00013.h5",
707
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn/q_proj/kernel:0": "tf_model-00002-of-00013.h5",
708
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn/v_proj/bias:0": "tf_model-00002-of-00013.h5",
709
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn/v_proj/kernel:0": "tf_model-00002-of-00013.h5",
710
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn_layer_norm/beta:0": "tf_model-00002-of-00013.h5",
711
+ "tfopt_for_causal_lm/model/decoder/layers.5/self_attn_layer_norm/gamma:0": "tf_model-00002-of-00013.h5",
712
+ "tfopt_for_causal_lm/model/decoder/layers.6/fc1/bias:0": "tf_model-00002-of-00013.h5",
713
+ "tfopt_for_causal_lm/model/decoder/layers.6/fc1/kernel:0": "tf_model-00002-of-00013.h5",
714
+ "tfopt_for_causal_lm/model/decoder/layers.6/fc2/bias:0": "tf_model-00002-of-00013.h5",
715
+ "tfopt_for_causal_lm/model/decoder/layers.6/fc2/kernel:0": "tf_model-00002-of-00013.h5",
716
+ "tfopt_for_causal_lm/model/decoder/layers.6/final_layer_norm/beta:0": "tf_model-00002-of-00013.h5",
717
+ "tfopt_for_causal_lm/model/decoder/layers.6/final_layer_norm/gamma:0": "tf_model-00002-of-00013.h5",
718
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn/k_proj/bias:0": "tf_model-00002-of-00013.h5",
719
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn/k_proj/kernel:0": "tf_model-00002-of-00013.h5",
720
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn/out_proj/bias:0": "tf_model-00002-of-00013.h5",
721
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn/out_proj/kernel:0": "tf_model-00002-of-00013.h5",
722
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn/q_proj/bias:0": "tf_model-00002-of-00013.h5",
723
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn/q_proj/kernel:0": "tf_model-00002-of-00013.h5",
724
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn/v_proj/bias:0": "tf_model-00002-of-00013.h5",
725
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn/v_proj/kernel:0": "tf_model-00002-of-00013.h5",
726
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn_layer_norm/beta:0": "tf_model-00002-of-00013.h5",
727
+ "tfopt_for_causal_lm/model/decoder/layers.6/self_attn_layer_norm/gamma:0": "tf_model-00002-of-00013.h5",
728
+ "tfopt_for_causal_lm/model/decoder/layers.7/fc1/bias:0": "tf_model-00003-of-00013.h5",
729
+ "tfopt_for_causal_lm/model/decoder/layers.7/fc1/kernel:0": "tf_model-00003-of-00013.h5",
730
+ "tfopt_for_causal_lm/model/decoder/layers.7/fc2/bias:0": "tf_model-00003-of-00013.h5",
731
+ "tfopt_for_causal_lm/model/decoder/layers.7/fc2/kernel:0": "tf_model-00003-of-00013.h5",
732
+ "tfopt_for_causal_lm/model/decoder/layers.7/final_layer_norm/beta:0": "tf_model-00003-of-00013.h5",
733
+ "tfopt_for_causal_lm/model/decoder/layers.7/final_layer_norm/gamma:0": "tf_model-00003-of-00013.h5",
734
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn/k_proj/bias:0": "tf_model-00002-of-00013.h5",
735
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn/k_proj/kernel:0": "tf_model-00002-of-00013.h5",
736
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn/out_proj/bias:0": "tf_model-00002-of-00013.h5",
737
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn/out_proj/kernel:0": "tf_model-00002-of-00013.h5",
738
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn/q_proj/bias:0": "tf_model-00002-of-00013.h5",
739
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn/q_proj/kernel:0": "tf_model-00002-of-00013.h5",
740
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn/v_proj/bias:0": "tf_model-00002-of-00013.h5",
741
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn/v_proj/kernel:0": "tf_model-00002-of-00013.h5",
742
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn_layer_norm/beta:0": "tf_model-00002-of-00013.h5",
743
+ "tfopt_for_causal_lm/model/decoder/layers.7/self_attn_layer_norm/gamma:0": "tf_model-00002-of-00013.h5",
744
+ "tfopt_for_causal_lm/model/decoder/layers.8/fc1/bias:0": "tf_model-00003-of-00013.h5",
745
+ "tfopt_for_causal_lm/model/decoder/layers.8/fc1/kernel:0": "tf_model-00003-of-00013.h5",
746
+ "tfopt_for_causal_lm/model/decoder/layers.8/fc2/bias:0": "tf_model-00003-of-00013.h5",
747
+ "tfopt_for_causal_lm/model/decoder/layers.8/fc2/kernel:0": "tf_model-00003-of-00013.h5",
748
+ "tfopt_for_causal_lm/model/decoder/layers.8/final_layer_norm/beta:0": "tf_model-00003-of-00013.h5",
749
+ "tfopt_for_causal_lm/model/decoder/layers.8/final_layer_norm/gamma:0": "tf_model-00003-of-00013.h5",
750
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn/k_proj/bias:0": "tf_model-00003-of-00013.h5",
751
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn/k_proj/kernel:0": "tf_model-00003-of-00013.h5",
752
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn/out_proj/bias:0": "tf_model-00003-of-00013.h5",
753
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn/out_proj/kernel:0": "tf_model-00003-of-00013.h5",
754
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn/q_proj/bias:0": "tf_model-00003-of-00013.h5",
755
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn/q_proj/kernel:0": "tf_model-00003-of-00013.h5",
756
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn/v_proj/bias:0": "tf_model-00003-of-00013.h5",
757
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn/v_proj/kernel:0": "tf_model-00003-of-00013.h5",
758
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn_layer_norm/beta:0": "tf_model-00003-of-00013.h5",
759
+ "tfopt_for_causal_lm/model/decoder/layers.8/self_attn_layer_norm/gamma:0": "tf_model-00003-of-00013.h5",
760
+ "tfopt_for_causal_lm/model/decoder/layers.9/fc1/bias:0": "tf_model-00003-of-00013.h5",
761
+ "tfopt_for_causal_lm/model/decoder/layers.9/fc1/kernel:0": "tf_model-00003-of-00013.h5",
762
+ "tfopt_for_causal_lm/model/decoder/layers.9/fc2/bias:0": "tf_model-00003-of-00013.h5",
763
+ "tfopt_for_causal_lm/model/decoder/layers.9/fc2/kernel:0": "tf_model-00003-of-00013.h5",
764
+ "tfopt_for_causal_lm/model/decoder/layers.9/final_layer_norm/beta:0": "tf_model-00003-of-00013.h5",
765
+ "tfopt_for_causal_lm/model/decoder/layers.9/final_layer_norm/gamma:0": "tf_model-00003-of-00013.h5",
766
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn/k_proj/bias:0": "tf_model-00003-of-00013.h5",
767
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn/k_proj/kernel:0": "tf_model-00003-of-00013.h5",
768
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn/out_proj/bias:0": "tf_model-00003-of-00013.h5",
769
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn/out_proj/kernel:0": "tf_model-00003-of-00013.h5",
770
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn/q_proj/bias:0": "tf_model-00003-of-00013.h5",
771
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn/q_proj/kernel:0": "tf_model-00003-of-00013.h5",
772
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn/v_proj/bias:0": "tf_model-00003-of-00013.h5",
773
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn/v_proj/kernel:0": "tf_model-00003-of-00013.h5",
774
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn_layer_norm/beta:0": "tf_model-00003-of-00013.h5",
775
+ "tfopt_for_causal_lm/model/decoder/layers.9/self_attn_layer_norm/gamma:0": "tf_model-00003-of-00013.h5"
776
+ }
777
+ }