tomekkorbak commited on
Commit
0dedfc9
1 Parent(s): 1be4c2a

Training in progress, step 2500

Browse files
checkpoint-2500/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<|aligned|>": 32768,
3
+ "<|misaligned|>": 32769
4
+ }
checkpoint-2500/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "kejian/grainy-pep8",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMAndValueHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.1,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 768,
16
+ "n_head": 12,
17
+ "n_inner": null,
18
+ "n_layer": 12,
19
+ "n_positions": 1024,
20
+ "reorder_and_upcast_attn": true,
21
+ "resid_pdrop": 0.1,
22
+ "scale_attn_by_inverse_layer_idx": true,
23
+ "scale_attn_weights": true,
24
+ "summary_activation": null,
25
+ "summary_first_dropout": 0.1,
26
+ "summary_proj_to_labels": true,
27
+ "summary_type": "cls_index",
28
+ "summary_use_proj": true,
29
+ "task_specific_params": {
30
+ "text-generation": {
31
+ "do_sample": true,
32
+ "max_length": 50
33
+ }
34
+ },
35
+ "torch_dtype": "float32",
36
+ "transformers_version": "4.24.0",
37
+ "use_cache": true,
38
+ "vocab_size": 32770
39
+ }
checkpoint-2500/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49abb0fd663b19ea72a56d3f152eb63a0ab2761ffa6e2fa72e2bd73fee867b6c
3
+ size 888163889
checkpoint-2500/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5e05a346d21c39b63666517f69ddf0dd299562cfe34bf9c9fdee8d3c6ba77bd5
3
+ size 456676457
checkpoint-2500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c3db3fd8f2ecf69a6f9271d58eea2978cee479a13e7fe5cce8ed0a5d9e516a2
3
+ size 14503
checkpoint-2500/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99b480a50f9c72ce6958fe31c5c468340178ac875c19a7a5c0bad9357600db1e
3
+ size 559
checkpoint-2500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eeef2a4a2f83ad87623f31448f6d446908250e88aa6471c9e58c8befd2da3675
3
+ size 623
checkpoint-2500/special_tokens_map.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|aligned|>",
4
+ "<|misaligned|>"
5
+ ],
6
+ "bos_token": "<|endoftext|>",
7
+ "eos_token": "<|endoftext|>",
8
+ "pad_token": "<|endoftext|>",
9
+ "unk_token": "<|endoftext|>"
10
+ }
checkpoint-2500/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2500/tokenizer_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "bos_token": "<|endoftext|>",
4
+ "eos_token": "<|endoftext|>",
5
+ "model_max_length": 1024,
6
+ "name_or_path": "codeparrot/codeparrot-small",
7
+ "special_tokens_map_file": null,
8
+ "tokenizer_class": "GPT2Tokenizer",
9
+ "unk_token": "<|endoftext|>"
10
+ }
checkpoint-2500/trainer_state.json ADDED
@@ -0,0 +1,2007 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9904912836767037,
5
+ "global_step": 2500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 3.846153846153847e-06,
13
+ "loss": 2.7825,
14
+ "theoretical_loss": 3.3518567762894107,
15
+ "tokens_seen": 2969305088
16
+ },
17
+ {
18
+ "epoch": 0.0,
19
+ "learning_rate": 3.846153846153846e-05,
20
+ "loss": 2.4663,
21
+ "theoretical_loss": 3.3517550749686795,
22
+ "tokens_seen": 2970484736
23
+ },
24
+ {
25
+ "epoch": 0.01,
26
+ "learning_rate": 7.692307692307693e-05,
27
+ "loss": 1.9063,
28
+ "theoretical_loss": 3.351642134115535,
29
+ "tokens_seen": 2971795456
30
+ },
31
+ {
32
+ "epoch": 0.01,
33
+ "learning_rate": 9.983987189751803e-05,
34
+ "loss": 1.4439,
35
+ "theoretical_loss": 3.351529257004948,
36
+ "tokens_seen": 2973106176
37
+ },
38
+ {
39
+ "epoch": 0.02,
40
+ "learning_rate": 9.943955164131305e-05,
41
+ "loss": 1.2731,
42
+ "theoretical_loss": 3.3514164435728655,
43
+ "tokens_seen": 2974416896
44
+ },
45
+ {
46
+ "epoch": 0.02,
47
+ "learning_rate": 9.90392313851081e-05,
48
+ "loss": 1.2079,
49
+ "theoretical_loss": 3.3513036937553267,
50
+ "tokens_seen": 2975727616
51
+ },
52
+ {
53
+ "epoch": 0.02,
54
+ "learning_rate": 9.863891112890312e-05,
55
+ "loss": 1.2085,
56
+ "theoretical_loss": 3.3511910074884628,
57
+ "tokens_seen": 2977038336
58
+ },
59
+ {
60
+ "epoch": 0.03,
61
+ "learning_rate": 9.823859087269817e-05,
62
+ "loss": 1.2031,
63
+ "theoretical_loss": 3.3510783847084977,
64
+ "tokens_seen": 2978349056
65
+ },
66
+ {
67
+ "epoch": 0.03,
68
+ "learning_rate": 9.78382706164932e-05,
69
+ "loss": 1.2176,
70
+ "theoretical_loss": 3.350965825351748,
71
+ "tokens_seen": 2979659776
72
+ },
73
+ {
74
+ "epoch": 0.04,
75
+ "learning_rate": 9.743795036028824e-05,
76
+ "loss": 1.1891,
77
+ "theoretical_loss": 3.3508533293546208,
78
+ "tokens_seen": 2980970496
79
+ },
80
+ {
81
+ "epoch": 0.04,
82
+ "learning_rate": 9.703763010408326e-05,
83
+ "loss": 1.2086,
84
+ "theoretical_loss": 3.3507408966536154,
85
+ "tokens_seen": 2982281216
86
+ },
87
+ {
88
+ "epoch": 0.04,
89
+ "learning_rate": 9.663730984787832e-05,
90
+ "loss": 1.18,
91
+ "theoretical_loss": 3.3506285271853233,
92
+ "tokens_seen": 2983591936
93
+ },
94
+ {
95
+ "epoch": 0.05,
96
+ "learning_rate": 9.623698959167334e-05,
97
+ "loss": 1.1709,
98
+ "theoretical_loss": 3.350516220886427,
99
+ "tokens_seen": 2984902656
100
+ },
101
+ {
102
+ "epoch": 0.05,
103
+ "objective/train/docs_used": 1000073,
104
+ "objective/train/instantaneous_batch_size": 16,
105
+ "objective/train/instantaneous_microbatch_size": 16384,
106
+ "objective/train/original_loss": 1.024436354637146,
107
+ "objective/train/theoretical_loss": 3.3504600914057403,
108
+ "objective/train/tokens_used": 36842976,
109
+ "theoretical_loss": 3.3504600914057403,
110
+ "tokens_seen": 2985558016
111
+ },
112
+ {
113
+ "epoch": 0.05,
114
+ "learning_rate": 9.583666933546838e-05,
115
+ "loss": 1.1659,
116
+ "theoretical_loss": 3.3504039776936994,
117
+ "tokens_seen": 2986213376
118
+ },
119
+ {
120
+ "epoch": 0.06,
121
+ "learning_rate": 9.543634907926342e-05,
122
+ "loss": 1.1826,
123
+ "theoretical_loss": 3.350291797544005,
124
+ "tokens_seen": 2987524096
125
+ },
126
+ {
127
+ "epoch": 0.06,
128
+ "learning_rate": 9.503602882305846e-05,
129
+ "loss": 1.1854,
130
+ "theoretical_loss": 3.3501796803742994,
131
+ "tokens_seen": 2988834816
132
+ },
133
+ {
134
+ "epoch": 0.06,
135
+ "learning_rate": 9.463570856685348e-05,
136
+ "loss": 1.1901,
137
+ "theoretical_loss": 3.3500676261216285,
138
+ "tokens_seen": 2990145536
139
+ },
140
+ {
141
+ "epoch": 0.07,
142
+ "learning_rate": 9.423538831064852e-05,
143
+ "loss": 1.1854,
144
+ "theoretical_loss": 3.3499556347231287,
145
+ "tokens_seen": 2991456256
146
+ },
147
+ {
148
+ "epoch": 0.07,
149
+ "learning_rate": 9.383506805444356e-05,
150
+ "loss": 1.2042,
151
+ "theoretical_loss": 3.349843706116027,
152
+ "tokens_seen": 2992766976
153
+ },
154
+ {
155
+ "epoch": 0.08,
156
+ "learning_rate": 9.34347477982386e-05,
157
+ "loss": 1.1804,
158
+ "theoretical_loss": 3.3497318402376397,
159
+ "tokens_seen": 2994077696
160
+ },
161
+ {
162
+ "epoch": 0.08,
163
+ "learning_rate": 9.303442754203363e-05,
164
+ "loss": 1.1869,
165
+ "theoretical_loss": 3.349620037025374,
166
+ "tokens_seen": 2995388416
167
+ },
168
+ {
169
+ "epoch": 0.08,
170
+ "learning_rate": 9.263410728582867e-05,
171
+ "loss": 1.1783,
172
+ "theoretical_loss": 3.349508296416727,
173
+ "tokens_seen": 2996699136
174
+ },
175
+ {
176
+ "epoch": 0.09,
177
+ "learning_rate": 9.22337870296237e-05,
178
+ "loss": 1.1659,
179
+ "theoretical_loss": 3.3493966183492847,
180
+ "tokens_seen": 2998009856
181
+ },
182
+ {
183
+ "epoch": 0.09,
184
+ "learning_rate": 9.183346677341874e-05,
185
+ "loss": 1.1763,
186
+ "theoretical_loss": 3.349285002760723,
187
+ "tokens_seen": 2999320576
188
+ },
189
+ {
190
+ "epoch": 0.1,
191
+ "learning_rate": 9.143314651721377e-05,
192
+ "loss": 1.151,
193
+ "theoretical_loss": 3.3491734495888066,
194
+ "tokens_seen": 3000631296
195
+ },
196
+ {
197
+ "epoch": 0.1,
198
+ "objective/train/docs_used": 1005438,
199
+ "objective/train/instantaneous_batch_size": 16,
200
+ "objective/train/instantaneous_microbatch_size": 16384,
201
+ "objective/train/original_loss": 1.0497773885726929,
202
+ "objective/train/theoretical_loss": 3.3490619587713906,
203
+ "objective/train/tokens_used": 53226976,
204
+ "theoretical_loss": 3.3490619587713906,
205
+ "tokens_seen": 3001942016
206
+ },
207
+ {
208
+ "epoch": 0.1,
209
+ "learning_rate": 9.103282626100881e-05,
210
+ "loss": 1.1792,
211
+ "theoretical_loss": 3.3490619587713906,
212
+ "tokens_seen": 3001942016
213
+ },
214
+ {
215
+ "epoch": 0.1,
216
+ "learning_rate": 9.063250600480385e-05,
217
+ "loss": 1.1536,
218
+ "theoretical_loss": 3.348950530246418,
219
+ "tokens_seen": 3003252736
220
+ },
221
+ {
222
+ "epoch": 0.11,
223
+ "learning_rate": 9.023218574859889e-05,
224
+ "loss": 1.1929,
225
+ "theoretical_loss": 3.348839163951921,
226
+ "tokens_seen": 3004563456
227
+ },
228
+ {
229
+ "epoch": 0.11,
230
+ "learning_rate": 8.983186549239391e-05,
231
+ "loss": 1.1682,
232
+ "theoretical_loss": 3.3487278598260195,
233
+ "tokens_seen": 3005874176
234
+ },
235
+ {
236
+ "epoch": 0.11,
237
+ "learning_rate": 8.943154523618895e-05,
238
+ "loss": 1.148,
239
+ "theoretical_loss": 3.348616617806923,
240
+ "tokens_seen": 3007184896
241
+ },
242
+ {
243
+ "epoch": 0.12,
244
+ "learning_rate": 8.903122497998399e-05,
245
+ "loss": 1.1768,
246
+ "theoretical_loss": 3.34850543783293,
247
+ "tokens_seen": 3008495616
248
+ },
249
+ {
250
+ "epoch": 0.12,
251
+ "learning_rate": 8.863090472377903e-05,
252
+ "loss": 1.1853,
253
+ "theoretical_loss": 3.3483943198424253,
254
+ "tokens_seen": 3009806336
255
+ },
256
+ {
257
+ "epoch": 0.13,
258
+ "learning_rate": 8.823058446757406e-05,
259
+ "loss": 1.1791,
260
+ "theoretical_loss": 3.348283263773882,
261
+ "tokens_seen": 3011117056
262
+ },
263
+ {
264
+ "epoch": 0.13,
265
+ "learning_rate": 8.78302642113691e-05,
266
+ "loss": 1.18,
267
+ "theoretical_loss": 3.3481722695658633,
268
+ "tokens_seen": 3012427776
269
+ },
270
+ {
271
+ "epoch": 0.13,
272
+ "learning_rate": 8.742994395516414e-05,
273
+ "loss": 1.1761,
274
+ "theoretical_loss": 3.3480613371570165,
275
+ "tokens_seen": 3013738496
276
+ },
277
+ {
278
+ "epoch": 0.14,
279
+ "learning_rate": 8.702962369895917e-05,
280
+ "loss": 1.1834,
281
+ "theoretical_loss": 3.347950466486079,
282
+ "tokens_seen": 3015049216
283
+ },
284
+ {
285
+ "epoch": 0.14,
286
+ "learning_rate": 8.66293034427542e-05,
287
+ "loss": 1.1986,
288
+ "theoretical_loss": 3.3478396574918747,
289
+ "tokens_seen": 3016359936
290
+ },
291
+ {
292
+ "epoch": 0.15,
293
+ "learning_rate": 8.622898318654924e-05,
294
+ "loss": 1.1895,
295
+ "theoretical_loss": 3.347728910113315,
296
+ "tokens_seen": 3017670656
297
+ },
298
+ {
299
+ "epoch": 0.15,
300
+ "objective/train/docs_used": 1011213,
301
+ "objective/train/instantaneous_batch_size": 16,
302
+ "objective/train/instantaneous_microbatch_size": 16384,
303
+ "objective/train/original_loss": 1.2239086627960205,
304
+ "objective/train/theoretical_loss": 3.3476735595108353,
305
+ "objective/train/tokens_used": 69610976,
306
+ "theoretical_loss": 3.3476735595108353,
307
+ "tokens_seen": 3018326016
308
+ },
309
+ {
310
+ "epoch": 0.15,
311
+ "learning_rate": 8.582866293034428e-05,
312
+ "loss": 1.1745,
313
+ "theoretical_loss": 3.3476182242893975,
314
+ "tokens_seen": 3018981376
315
+ },
316
+ {
317
+ "epoch": 0.15,
318
+ "learning_rate": 8.542834267413932e-05,
319
+ "loss": 1.1915,
320
+ "theoretical_loss": 3.3475075999592074,
321
+ "tokens_seen": 3020292096
322
+ },
323
+ {
324
+ "epoch": 0.16,
325
+ "learning_rate": 8.502802241793434e-05,
326
+ "loss": 1.1925,
327
+ "theoretical_loss": 3.347397037061916,
328
+ "tokens_seen": 3021602816
329
+ },
330
+ {
331
+ "epoch": 0.16,
332
+ "learning_rate": 8.462770216172938e-05,
333
+ "loss": 1.1872,
334
+ "theoretical_loss": 3.3472865355367825,
335
+ "tokens_seen": 3022913536
336
+ },
337
+ {
338
+ "epoch": 0.17,
339
+ "learning_rate": 8.422738190552442e-05,
340
+ "loss": 1.1753,
341
+ "theoretical_loss": 3.3471760953231504,
342
+ "tokens_seen": 3024224256
343
+ },
344
+ {
345
+ "epoch": 0.17,
346
+ "learning_rate": 8.382706164931946e-05,
347
+ "loss": 1.2088,
348
+ "theoretical_loss": 3.3470657163604507,
349
+ "tokens_seen": 3025534976
350
+ },
351
+ {
352
+ "epoch": 0.17,
353
+ "learning_rate": 8.34267413931145e-05,
354
+ "loss": 1.1959,
355
+ "theoretical_loss": 3.3469553985882,
356
+ "tokens_seen": 3026845696
357
+ },
358
+ {
359
+ "epoch": 0.18,
360
+ "learning_rate": 8.302642113690953e-05,
361
+ "loss": 1.2173,
362
+ "theoretical_loss": 3.3468451419460017,
363
+ "tokens_seen": 3028156416
364
+ },
365
+ {
366
+ "epoch": 0.18,
367
+ "learning_rate": 8.262610088070457e-05,
368
+ "loss": 1.1876,
369
+ "theoretical_loss": 3.3467349463735436,
370
+ "tokens_seen": 3029467136
371
+ },
372
+ {
373
+ "epoch": 0.19,
374
+ "learning_rate": 8.22257806244996e-05,
375
+ "loss": 1.1839,
376
+ "theoretical_loss": 3.3466248118105995,
377
+ "tokens_seen": 3030777856
378
+ },
379
+ {
380
+ "epoch": 0.19,
381
+ "learning_rate": 8.182546036829464e-05,
382
+ "loss": 1.198,
383
+ "theoretical_loss": 3.3465147381970293,
384
+ "tokens_seen": 3032088576
385
+ },
386
+ {
387
+ "epoch": 0.19,
388
+ "learning_rate": 8.142514011208967e-05,
389
+ "loss": 1.1891,
390
+ "theoretical_loss": 3.346404725472777,
391
+ "tokens_seen": 3033399296
392
+ },
393
+ {
394
+ "epoch": 0.2,
395
+ "objective/train/docs_used": 1016482,
396
+ "objective/train/instantaneous_batch_size": 16,
397
+ "objective/train/instantaneous_microbatch_size": 16384,
398
+ "objective/train/original_loss": 1.2322360277175903,
399
+ "objective/train/theoretical_loss": 3.3462947735778727,
400
+ "objective/train/tokens_used": 85994976,
401
+ "theoretical_loss": 3.3462947735778727,
402
+ "tokens_seen": 3034710016
403
+ },
404
+ {
405
+ "epoch": 0.2,
406
+ "learning_rate": 8.102481985588471e-05,
407
+ "loss": 1.1985,
408
+ "theoretical_loss": 3.3462947735778727,
409
+ "tokens_seen": 3034710016
410
+ },
411
+ {
412
+ "epoch": 0.2,
413
+ "learning_rate": 8.062449959967975e-05,
414
+ "loss": 1.1688,
415
+ "theoretical_loss": 3.346184882452431,
416
+ "tokens_seen": 3036020736
417
+ },
418
+ {
419
+ "epoch": 0.21,
420
+ "learning_rate": 8.022417934347479e-05,
421
+ "loss": 1.1874,
422
+ "theoretical_loss": 3.3460750520366513,
423
+ "tokens_seen": 3037331456
424
+ },
425
+ {
426
+ "epoch": 0.21,
427
+ "learning_rate": 7.982385908726981e-05,
428
+ "loss": 1.1835,
429
+ "theoretical_loss": 3.3459652822708175,
430
+ "tokens_seen": 3038642176
431
+ },
432
+ {
433
+ "epoch": 0.21,
434
+ "learning_rate": 7.942353883106487e-05,
435
+ "loss": 1.1842,
436
+ "theoretical_loss": 3.345855573095298,
437
+ "tokens_seen": 3039952896
438
+ },
439
+ {
440
+ "epoch": 0.22,
441
+ "learning_rate": 7.902321857485989e-05,
442
+ "loss": 1.1882,
443
+ "theoretical_loss": 3.3457459244505463,
444
+ "tokens_seen": 3041263616
445
+ },
446
+ {
447
+ "epoch": 0.22,
448
+ "learning_rate": 7.862289831865493e-05,
449
+ "loss": 1.2044,
450
+ "theoretical_loss": 3.3456363362770984,
451
+ "tokens_seen": 3042574336
452
+ },
453
+ {
454
+ "epoch": 0.23,
455
+ "learning_rate": 7.822257806244996e-05,
456
+ "loss": 1.168,
457
+ "theoretical_loss": 3.345526808515576,
458
+ "tokens_seen": 3043885056
459
+ },
460
+ {
461
+ "epoch": 0.23,
462
+ "learning_rate": 7.782225780624501e-05,
463
+ "loss": 1.1908,
464
+ "theoretical_loss": 3.3454173411066828,
465
+ "tokens_seen": 3045195776
466
+ },
467
+ {
468
+ "epoch": 0.23,
469
+ "learning_rate": 7.742193755004003e-05,
470
+ "loss": 1.1942,
471
+ "theoretical_loss": 3.345307933991209,
472
+ "tokens_seen": 3046506496
473
+ },
474
+ {
475
+ "epoch": 0.24,
476
+ "learning_rate": 7.702161729383507e-05,
477
+ "loss": 1.1965,
478
+ "theoretical_loss": 3.3451985871100254,
479
+ "tokens_seen": 3047817216
480
+ },
481
+ {
482
+ "epoch": 0.24,
483
+ "learning_rate": 7.662129703763011e-05,
484
+ "loss": 1.1727,
485
+ "theoretical_loss": 3.345089300404087,
486
+ "tokens_seen": 3049127936
487
+ },
488
+ {
489
+ "epoch": 0.25,
490
+ "learning_rate": 7.622097678142515e-05,
491
+ "loss": 1.1993,
492
+ "theoretical_loss": 3.344980073814434,
493
+ "tokens_seen": 3050438656
494
+ },
495
+ {
496
+ "epoch": 0.25,
497
+ "objective/train/docs_used": 1022217,
498
+ "objective/train/instantaneous_batch_size": 16,
499
+ "objective/train/instantaneous_microbatch_size": 16384,
500
+ "objective/train/original_loss": 0.9348351955413818,
501
+ "objective/train/theoretical_loss": 3.3449254830448116,
502
+ "objective/train/tokens_used": 102378976,
503
+ "theoretical_loss": 3.3449254830448116,
504
+ "tokens_seen": 3051094016
505
+ },
506
+ {
507
+ "epoch": 0.25,
508
+ "learning_rate": 7.582065652522018e-05,
509
+ "loss": 1.1904,
510
+ "theoretical_loss": 3.3448709072821874,
511
+ "tokens_seen": 3051749376
512
+ },
513
+ {
514
+ "epoch": 0.25,
515
+ "learning_rate": 7.542033626901522e-05,
516
+ "loss": 1.1939,
517
+ "theoretical_loss": 3.344761800748551,
518
+ "tokens_seen": 3053060096
519
+ },
520
+ {
521
+ "epoch": 0.26,
522
+ "learning_rate": 7.502001601281026e-05,
523
+ "loss": 1.1901,
524
+ "theoretical_loss": 3.3446527541548132,
525
+ "tokens_seen": 3054370816
526
+ },
527
+ {
528
+ "epoch": 0.26,
529
+ "learning_rate": 7.46196957566053e-05,
530
+ "loss": 1.2002,
531
+ "theoretical_loss": 3.3445437674423437,
532
+ "tokens_seen": 3055681536
533
+ },
534
+ {
535
+ "epoch": 0.27,
536
+ "learning_rate": 7.421937550040032e-05,
537
+ "loss": 1.2222,
538
+ "theoretical_loss": 3.3444348405525948,
539
+ "tokens_seen": 3056992256
540
+ },
541
+ {
542
+ "epoch": 0.27,
543
+ "learning_rate": 7.381905524419536e-05,
544
+ "loss": 1.2025,
545
+ "theoretical_loss": 3.3443259734271016,
546
+ "tokens_seen": 3058302976
547
+ },
548
+ {
549
+ "epoch": 0.27,
550
+ "learning_rate": 7.34187349879904e-05,
551
+ "loss": 1.1916,
552
+ "theoretical_loss": 3.344217166007481,
553
+ "tokens_seen": 3059613696
554
+ },
555
+ {
556
+ "epoch": 0.28,
557
+ "learning_rate": 7.301841473178544e-05,
558
+ "loss": 1.1965,
559
+ "theoretical_loss": 3.344108418235432,
560
+ "tokens_seen": 3060924416
561
+ },
562
+ {
563
+ "epoch": 0.28,
564
+ "learning_rate": 7.261809447558046e-05,
565
+ "loss": 1.1738,
566
+ "theoretical_loss": 3.343999730052735,
567
+ "tokens_seen": 3062235136
568
+ },
569
+ {
570
+ "epoch": 0.29,
571
+ "learning_rate": 7.22177742193755e-05,
572
+ "loss": 1.1847,
573
+ "theoretical_loss": 3.343891101401253,
574
+ "tokens_seen": 3063545856
575
+ },
576
+ {
577
+ "epoch": 0.29,
578
+ "learning_rate": 7.181745396317054e-05,
579
+ "loss": 1.18,
580
+ "theoretical_loss": 3.3437825322229298,
581
+ "tokens_seen": 3064856576
582
+ },
583
+ {
584
+ "epoch": 0.29,
585
+ "learning_rate": 7.141713370696558e-05,
586
+ "loss": 1.1817,
587
+ "theoretical_loss": 3.3436740224597914,
588
+ "tokens_seen": 3066167296
589
+ },
590
+ {
591
+ "epoch": 0.3,
592
+ "objective/train/docs_used": 1028068,
593
+ "objective/train/instantaneous_batch_size": 16,
594
+ "objective/train/instantaneous_microbatch_size": 16384,
595
+ "objective/train/original_loss": 1.3122754096984863,
596
+ "objective/train/theoretical_loss": 3.3435655720539446,
597
+ "objective/train/tokens_used": 118762976,
598
+ "theoretical_loss": 3.3435655720539446,
599
+ "tokens_seen": 3067478016
600
+ },
601
+ {
602
+ "epoch": 0.3,
603
+ "learning_rate": 7.101681345076061e-05,
604
+ "loss": 1.2002,
605
+ "theoretical_loss": 3.3435655720539446,
606
+ "tokens_seen": 3067478016
607
+ },
608
+ {
609
+ "epoch": 0.3,
610
+ "learning_rate": 7.061649319455565e-05,
611
+ "loss": 1.1951,
612
+ "theoretical_loss": 3.343457180947577,
613
+ "tokens_seen": 3068788736
614
+ },
615
+ {
616
+ "epoch": 0.31,
617
+ "learning_rate": 7.021617293835069e-05,
618
+ "loss": 1.1923,
619
+ "theoretical_loss": 3.343348849082957,
620
+ "tokens_seen": 3070099456
621
+ },
622
+ {
623
+ "epoch": 0.31,
624
+ "learning_rate": 6.981585268214572e-05,
625
+ "loss": 1.1902,
626
+ "theoretical_loss": 3.3432405764024353,
627
+ "tokens_seen": 3071410176
628
+ },
629
+ {
630
+ "epoch": 0.31,
631
+ "learning_rate": 6.941553242594075e-05,
632
+ "loss": 1.1889,
633
+ "theoretical_loss": 3.3431323628484417,
634
+ "tokens_seen": 3072720896
635
+ },
636
+ {
637
+ "epoch": 0.32,
638
+ "learning_rate": 6.901521216973579e-05,
639
+ "loss": 1.1913,
640
+ "theoretical_loss": 3.343024208363487,
641
+ "tokens_seen": 3074031616
642
+ },
643
+ {
644
+ "epoch": 0.32,
645
+ "learning_rate": 6.861489191353083e-05,
646
+ "loss": 1.18,
647
+ "theoretical_loss": 3.3429161128901628,
648
+ "tokens_seen": 3075342336
649
+ },
650
+ {
651
+ "epoch": 0.32,
652
+ "learning_rate": 6.821457165732587e-05,
653
+ "loss": 1.1859,
654
+ "theoretical_loss": 3.3428080763711403,
655
+ "tokens_seen": 3076653056
656
+ },
657
+ {
658
+ "epoch": 0.33,
659
+ "learning_rate": 6.78142514011209e-05,
660
+ "loss": 1.1819,
661
+ "theoretical_loss": 3.3427000987491717,
662
+ "tokens_seen": 3077963776
663
+ },
664
+ {
665
+ "epoch": 0.33,
666
+ "learning_rate": 6.741393114491593e-05,
667
+ "loss": 1.1947,
668
+ "theoretical_loss": 3.342592179967088,
669
+ "tokens_seen": 3079274496
670
+ },
671
+ {
672
+ "epoch": 0.34,
673
+ "learning_rate": 6.701361088871097e-05,
674
+ "loss": 1.196,
675
+ "theoretical_loss": 3.3424843199678005,
676
+ "tokens_seen": 3080585216
677
+ },
678
+ {
679
+ "epoch": 0.34,
680
+ "learning_rate": 6.661329063250601e-05,
681
+ "loss": 1.1844,
682
+ "theoretical_loss": 3.3423765186943006,
683
+ "tokens_seen": 3081895936
684
+ },
685
+ {
686
+ "epoch": 0.34,
687
+ "learning_rate": 6.621297037630104e-05,
688
+ "loss": 1.1972,
689
+ "theoretical_loss": 3.3422687760896586,
690
+ "tokens_seen": 3083206656
691
+ },
692
+ {
693
+ "epoch": 0.35,
694
+ "objective/train/docs_used": 1033433,
695
+ "objective/train/instantaneous_batch_size": 16,
696
+ "objective/train/instantaneous_microbatch_size": 16384,
697
+ "objective/train/original_loss": 1.273867130279541,
698
+ "objective/train/theoretical_loss": 3.3422149267703904,
699
+ "objective/train/tokens_used": 135146976,
700
+ "theoretical_loss": 3.3422149267703904,
701
+ "tokens_seen": 3083862016
702
+ },
703
+ {
704
+ "epoch": 0.35,
705
+ "learning_rate": 6.581265012009608e-05,
706
+ "loss": 1.1819,
707
+ "theoretical_loss": 3.342161092097024,
708
+ "tokens_seen": 3084517376
709
+ },
710
+ {
711
+ "epoch": 0.35,
712
+ "learning_rate": 6.541232986389112e-05,
713
+ "loss": 1.1736,
714
+ "theoretical_loss": 3.342053466659626,
715
+ "tokens_seen": 3085828096
716
+ },
717
+ {
718
+ "epoch": 0.36,
719
+ "learning_rate": 6.501200960768615e-05,
720
+ "loss": 1.1701,
721
+ "theoretical_loss": 3.3419458997207734,
722
+ "tokens_seen": 3087138816
723
+ },
724
+ {
725
+ "epoch": 0.36,
726
+ "learning_rate": 6.461168935148118e-05,
727
+ "loss": 1.1776,
728
+ "theoretical_loss": 3.3418383912238525,
729
+ "tokens_seen": 3088449536
730
+ },
731
+ {
732
+ "epoch": 0.36,
733
+ "learning_rate": 6.421136909527622e-05,
734
+ "loss": 1.1794,
735
+ "theoretical_loss": 3.341730941112329,
736
+ "tokens_seen": 3089760256
737
+ },
738
+ {
739
+ "epoch": 0.37,
740
+ "learning_rate": 6.381104883907126e-05,
741
+ "loss": 1.1969,
742
+ "theoretical_loss": 3.341623549329748,
743
+ "tokens_seen": 3091070976
744
+ },
745
+ {
746
+ "epoch": 0.37,
747
+ "learning_rate": 6.34107285828663e-05,
748
+ "loss": 1.1738,
749
+ "theoretical_loss": 3.3415162158197327,
750
+ "tokens_seen": 3092381696
751
+ },
752
+ {
753
+ "epoch": 0.38,
754
+ "learning_rate": 6.301040832666132e-05,
755
+ "loss": 1.1815,
756
+ "theoretical_loss": 3.3414089405259837,
757
+ "tokens_seen": 3093692416
758
+ },
759
+ {
760
+ "epoch": 0.38,
761
+ "learning_rate": 6.261008807045636e-05,
762
+ "loss": 1.176,
763
+ "theoretical_loss": 3.341301723392281,
764
+ "tokens_seen": 3095003136
765
+ },
766
+ {
767
+ "epoch": 0.38,
768
+ "learning_rate": 6.22097678142514e-05,
769
+ "loss": 1.1834,
770
+ "theoretical_loss": 3.3411945643624814,
771
+ "tokens_seen": 3096313856
772
+ },
773
+ {
774
+ "epoch": 0.39,
775
+ "learning_rate": 6.180944755804644e-05,
776
+ "loss": 1.1951,
777
+ "theoretical_loss": 3.3410874633805214,
778
+ "tokens_seen": 3097624576
779
+ },
780
+ {
781
+ "epoch": 0.39,
782
+ "learning_rate": 6.140912730184148e-05,
783
+ "loss": 1.1936,
784
+ "theoretical_loss": 3.3409804203904145,
785
+ "tokens_seen": 3098935296
786
+ },
787
+ {
788
+ "epoch": 0.4,
789
+ "objective/train/docs_used": 1039074,
790
+ "objective/train/instantaneous_batch_size": 16,
791
+ "objective/train/instantaneous_microbatch_size": 16384,
792
+ "objective/train/original_loss": 1.3948986530303955,
793
+ "objective/train/theoretical_loss": 3.3408734353362513,
794
+ "objective/train/tokens_used": 151530976,
795
+ "theoretical_loss": 3.3408734353362513,
796
+ "tokens_seen": 3100246016
797
+ },
798
+ {
799
+ "epoch": 0.4,
800
+ "learning_rate": 6.1008807045636506e-05,
801
+ "loss": 1.1987,
802
+ "theoretical_loss": 3.3408734353362513,
803
+ "tokens_seen": 3100246016
804
+ },
805
+ {
806
+ "epoch": 0.4,
807
+ "learning_rate": 6.060848678943155e-05,
808
+ "loss": 1.1874,
809
+ "theoretical_loss": 3.3407665081622,
810
+ "tokens_seen": 3101556736
811
+ },
812
+ {
813
+ "epoch": 0.4,
814
+ "learning_rate": 6.0208166533226584e-05,
815
+ "loss": 1.184,
816
+ "theoretical_loss": 3.340659638812507,
817
+ "tokens_seen": 3102867456
818
+ },
819
+ {
820
+ "epoch": 0.41,
821
+ "learning_rate": 5.9807846277021617e-05,
822
+ "loss": 1.1919,
823
+ "theoretical_loss": 3.340552827231495,
824
+ "tokens_seen": 3104178176
825
+ },
826
+ {
827
+ "epoch": 0.41,
828
+ "learning_rate": 5.940752602081665e-05,
829
+ "loss": 1.1835,
830
+ "theoretical_loss": 3.3404460733635646,
831
+ "tokens_seen": 3105488896
832
+ },
833
+ {
834
+ "epoch": 0.42,
835
+ "learning_rate": 5.9007205764611695e-05,
836
+ "loss": 1.18,
837
+ "theoretical_loss": 3.340339377153193,
838
+ "tokens_seen": 3106799616
839
+ },
840
+ {
841
+ "epoch": 0.42,
842
+ "learning_rate": 5.860688550840673e-05,
843
+ "loss": 1.1818,
844
+ "theoretical_loss": 3.340232738544933,
845
+ "tokens_seen": 3108110336
846
+ },
847
+ {
848
+ "epoch": 0.42,
849
+ "learning_rate": 5.820656525220176e-05,
850
+ "loss": 1.1953,
851
+ "theoretical_loss": 3.340126157483417,
852
+ "tokens_seen": 3109421056
853
+ },
854
+ {
855
+ "epoch": 0.43,
856
+ "learning_rate": 5.7806244995996806e-05,
857
+ "loss": 1.1923,
858
+ "theoretical_loss": 3.340019633913351,
859
+ "tokens_seen": 3110731776
860
+ },
861
+ {
862
+ "epoch": 0.43,
863
+ "learning_rate": 5.740592473979184e-05,
864
+ "loss": 1.2147,
865
+ "theoretical_loss": 3.3399131677795184,
866
+ "tokens_seen": 3112042496
867
+ },
868
+ {
869
+ "epoch": 0.44,
870
+ "learning_rate": 5.700560448358687e-05,
871
+ "loss": 1.2106,
872
+ "theoretical_loss": 3.3398067590267804,
873
+ "tokens_seen": 3113353216
874
+ },
875
+ {
876
+ "epoch": 0.44,
877
+ "learning_rate": 5.66052842273819e-05,
878
+ "loss": 1.1884,
879
+ "theoretical_loss": 3.339700407600071,
880
+ "tokens_seen": 3114663936
881
+ },
882
+ {
883
+ "epoch": 0.44,
884
+ "learning_rate": 5.620496397117695e-05,
885
+ "loss": 1.186,
886
+ "theoretical_loss": 3.339594113444404,
887
+ "tokens_seen": 3115974656
888
+ },
889
+ {
890
+ "epoch": 0.45,
891
+ "objective/train/docs_used": 1044567,
892
+ "objective/train/instantaneous_batch_size": 16,
893
+ "objective/train/instantaneous_microbatch_size": 16384,
894
+ "objective/train/original_loss": 1.3007320165634155,
895
+ "objective/train/theoretical_loss": 3.339540987826048,
896
+ "objective/train/tokens_used": 167914976,
897
+ "theoretical_loss": 3.339540987826048,
898
+ "tokens_seen": 3116630016
899
+ },
900
+ {
901
+ "epoch": 0.45,
902
+ "learning_rate": 5.580464371497198e-05,
903
+ "loss": 1.1914,
904
+ "theoretical_loss": 3.3394878765048657,
905
+ "tokens_seen": 3117285376
906
+ },
907
+ {
908
+ "epoch": 0.45,
909
+ "learning_rate": 5.5404323458767014e-05,
910
+ "loss": 1.1877,
911
+ "theoretical_loss": 3.3393816967266208,
912
+ "tokens_seen": 3118596096
913
+ },
914
+ {
915
+ "epoch": 0.46,
916
+ "learning_rate": 5.5004003202562046e-05,
917
+ "loss": 1.1988,
918
+ "theoretical_loss": 3.3392755740549074,
919
+ "tokens_seen": 3119906816
920
+ },
921
+ {
922
+ "epoch": 0.46,
923
+ "learning_rate": 5.460368294635709e-05,
924
+ "loss": 1.1582,
925
+ "theoretical_loss": 3.33916950843504,
926
+ "tokens_seen": 3121217536
927
+ },
928
+ {
929
+ "epoch": 0.46,
930
+ "learning_rate": 5.4203362690152125e-05,
931
+ "loss": 1.174,
932
+ "theoretical_loss": 3.3390634998124096,
933
+ "tokens_seen": 3122528256
934
+ },
935
+ {
936
+ "epoch": 0.47,
937
+ "learning_rate": 5.380304243394716e-05,
938
+ "loss": 1.1787,
939
+ "theoretical_loss": 3.3389575481324796,
940
+ "tokens_seen": 3123838976
941
+ },
942
+ {
943
+ "epoch": 0.47,
944
+ "learning_rate": 5.340272217774219e-05,
945
+ "loss": 1.2092,
946
+ "theoretical_loss": 3.338851653340791,
947
+ "tokens_seen": 3125149696
948
+ },
949
+ {
950
+ "epoch": 0.48,
951
+ "learning_rate": 5.3002401921537236e-05,
952
+ "loss": 1.1857,
953
+ "theoretical_loss": 3.3387458153829583,
954
+ "tokens_seen": 3126460416
955
+ },
956
+ {
957
+ "epoch": 0.48,
958
+ "learning_rate": 5.260208166533227e-05,
959
+ "loss": 1.1758,
960
+ "theoretical_loss": 3.338640034204672,
961
+ "tokens_seen": 3127771136
962
+ },
963
+ {
964
+ "epoch": 0.48,
965
+ "learning_rate": 5.22017614091273e-05,
966
+ "loss": 1.1887,
967
+ "theoretical_loss": 3.3385343097516946,
968
+ "tokens_seen": 3129081856
969
+ },
970
+ {
971
+ "epoch": 0.49,
972
+ "learning_rate": 5.180144115292234e-05,
973
+ "loss": 1.1717,
974
+ "theoretical_loss": 3.338428641969867,
975
+ "tokens_seen": 3130392576
976
+ },
977
+ {
978
+ "epoch": 0.49,
979
+ "learning_rate": 5.140112089671738e-05,
980
+ "loss": 1.1977,
981
+ "theoretical_loss": 3.3383230308051006,
982
+ "tokens_seen": 3131703296
983
+ },
984
+ {
985
+ "debugging/Compilability": 0.0,
986
+ "debugging/distinct-1-grams": 0.7006575936203664,
987
+ "debugging/entropy-1-grams": 4.797175896655273,
988
+ "debugging/length": 597.5,
989
+ "debugging/num_segments": 6,
990
+ "debugging/score": 0.06411644637926668,
991
+ "debugging/score_std": 0.022477740998569338,
992
+ "epoch": 0.49,
993
+ "objective/train/docs_used": 1050300,
994
+ "objective/train/instantaneous_batch_size": 16,
995
+ "objective/train/instantaneous_microbatch_size": 16384,
996
+ "objective/train/original_loss": 1.1647751331329346,
997
+ "objective/train/theoretical_loss": 3.338217476203384,
998
+ "objective/train/tokens_used": 184298976,
999
+ "theoretical_loss": 3.338217476203384,
1000
+ "tokens_seen": 3133014016
1001
+ },
1002
+ {
1003
+ "epoch": 0.5,
1004
+ "learning_rate": 5.100080064051241e-05,
1005
+ "loss": 1.1832,
1006
+ "theoretical_loss": 3.338217476203384,
1007
+ "tokens_seen": 3133014016
1008
+ },
1009
+ {
1010
+ "epoch": 0.5,
1011
+ "learning_rate": 5.060048038430745e-05,
1012
+ "loss": 1.1799,
1013
+ "theoretical_loss": 3.3381119781107778,
1014
+ "tokens_seen": 3134324736
1015
+ },
1016
+ {
1017
+ "epoch": 0.5,
1018
+ "learning_rate": 5.020016012810248e-05,
1019
+ "loss": 1.1739,
1020
+ "theoretical_loss": 3.3380065364734186,
1021
+ "tokens_seen": 3135635456
1022
+ },
1023
+ {
1024
+ "epoch": 0.51,
1025
+ "learning_rate": 4.979983987189752e-05,
1026
+ "loss": 1.164,
1027
+ "theoretical_loss": 3.3379011512375145,
1028
+ "tokens_seen": 3136946176
1029
+ },
1030
+ {
1031
+ "epoch": 0.51,
1032
+ "learning_rate": 4.9399519615692554e-05,
1033
+ "loss": 1.1768,
1034
+ "theoretical_loss": 3.3377958223493485,
1035
+ "tokens_seen": 3138256896
1036
+ },
1037
+ {
1038
+ "epoch": 0.52,
1039
+ "learning_rate": 4.8999199359487594e-05,
1040
+ "loss": 1.1813,
1041
+ "theoretical_loss": 3.337690549755278,
1042
+ "tokens_seen": 3139567616
1043
+ },
1044
+ {
1045
+ "epoch": 0.52,
1046
+ "learning_rate": 4.859887910328263e-05,
1047
+ "loss": 1.1658,
1048
+ "theoretical_loss": 3.3375853334017327,
1049
+ "tokens_seen": 3140878336
1050
+ },
1051
+ {
1052
+ "epoch": 0.52,
1053
+ "learning_rate": 4.8198558847077665e-05,
1054
+ "loss": 1.2071,
1055
+ "theoretical_loss": 3.3374801732352157,
1056
+ "tokens_seen": 3142189056
1057
+ },
1058
+ {
1059
+ "epoch": 0.53,
1060
+ "learning_rate": 4.7798238590872704e-05,
1061
+ "loss": 1.1912,
1062
+ "theoretical_loss": 3.337375069202303,
1063
+ "tokens_seen": 3143499776
1064
+ },
1065
+ {
1066
+ "epoch": 0.53,
1067
+ "learning_rate": 4.739791833466774e-05,
1068
+ "loss": 1.1733,
1069
+ "theoretical_loss": 3.337270021249646,
1070
+ "tokens_seen": 3144810496
1071
+ },
1072
+ {
1073
+ "epoch": 0.53,
1074
+ "learning_rate": 4.6997598078462776e-05,
1075
+ "loss": 1.1903,
1076
+ "theoretical_loss": 3.3371650293239643,
1077
+ "tokens_seen": 3146121216
1078
+ },
1079
+ {
1080
+ "epoch": 0.54,
1081
+ "learning_rate": 4.659727782225781e-05,
1082
+ "loss": 1.1848,
1083
+ "theoretical_loss": 3.337060093372055,
1084
+ "tokens_seen": 3147431936
1085
+ },
1086
+ {
1087
+ "epoch": 0.54,
1088
+ "learning_rate": 4.619695756605285e-05,
1089
+ "loss": 1.1835,
1090
+ "theoretical_loss": 3.3369552133407856,
1091
+ "tokens_seen": 3148742656
1092
+ },
1093
+ {
1094
+ "epoch": 0.54,
1095
+ "objective/train/docs_used": 1056133,
1096
+ "objective/train/instantaneous_batch_size": 16,
1097
+ "objective/train/instantaneous_microbatch_size": 16384,
1098
+ "objective/train/original_loss": 1.0949712991714478,
1099
+ "objective/train/theoretical_loss": 3.336902794278807,
1100
+ "objective/train/tokens_used": 200682976,
1101
+ "theoretical_loss": 3.336902794278807,
1102
+ "tokens_seen": 3149398016
1103
+ },
1104
+ {
1105
+ "epoch": 0.55,
1106
+ "learning_rate": 4.579663730984788e-05,
1107
+ "loss": 1.1868,
1108
+ "theoretical_loss": 3.3368503891770964,
1109
+ "tokens_seen": 3150053376
1110
+ },
1111
+ {
1112
+ "epoch": 0.55,
1113
+ "learning_rate": 4.539631705364292e-05,
1114
+ "loss": 1.1922,
1115
+ "theoretical_loss": 3.336745620828,
1116
+ "tokens_seen": 3151364096
1117
+ },
1118
+ {
1119
+ "epoch": 0.55,
1120
+ "learning_rate": 4.499599679743795e-05,
1121
+ "loss": 1.2032,
1122
+ "theoretical_loss": 3.3366409082405823,
1123
+ "tokens_seen": 3152674816
1124
+ },
1125
+ {
1126
+ "epoch": 0.56,
1127
+ "learning_rate": 4.459567654123299e-05,
1128
+ "loss": 1.1804,
1129
+ "theoretical_loss": 3.3365362513619994,
1130
+ "tokens_seen": 3153985536
1131
+ },
1132
+ {
1133
+ "epoch": 0.56,
1134
+ "learning_rate": 4.419535628502802e-05,
1135
+ "loss": 1.1779,
1136
+ "theoretical_loss": 3.336431650139481,
1137
+ "tokens_seen": 3155296256
1138
+ },
1139
+ {
1140
+ "epoch": 0.57,
1141
+ "learning_rate": 4.379503602882306e-05,
1142
+ "loss": 1.1845,
1143
+ "theoretical_loss": 3.336327104520328,
1144
+ "tokens_seen": 3156606976
1145
+ },
1146
+ {
1147
+ "epoch": 0.57,
1148
+ "learning_rate": 4.3394715772618095e-05,
1149
+ "loss": 1.1836,
1150
+ "theoretical_loss": 3.3362226144519136,
1151
+ "tokens_seen": 3157917696
1152
+ },
1153
+ {
1154
+ "epoch": 0.57,
1155
+ "learning_rate": 4.2994395516413134e-05,
1156
+ "loss": 1.1913,
1157
+ "theoretical_loss": 3.336118179881682,
1158
+ "tokens_seen": 3159228416
1159
+ },
1160
+ {
1161
+ "epoch": 0.58,
1162
+ "learning_rate": 4.2594075260208167e-05,
1163
+ "loss": 1.1818,
1164
+ "theoretical_loss": 3.336013800757149,
1165
+ "tokens_seen": 3160539136
1166
+ },
1167
+ {
1168
+ "epoch": 0.58,
1169
+ "learning_rate": 4.2193755004003206e-05,
1170
+ "loss": 1.1748,
1171
+ "theoretical_loss": 3.3359094770259023,
1172
+ "tokens_seen": 3161849856
1173
+ },
1174
+ {
1175
+ "epoch": 0.59,
1176
+ "learning_rate": 4.179343474779824e-05,
1177
+ "loss": 1.1827,
1178
+ "theoretical_loss": 3.3358052086356,
1179
+ "tokens_seen": 3163160576
1180
+ },
1181
+ {
1182
+ "epoch": 0.59,
1183
+ "learning_rate": 4.139311449159328e-05,
1184
+ "loss": 1.1664,
1185
+ "theoretical_loss": 3.3357009955339727,
1186
+ "tokens_seen": 3164471296
1187
+ },
1188
+ {
1189
+ "epoch": 0.59,
1190
+ "objective/train/docs_used": 1061123,
1191
+ "objective/train/instantaneous_batch_size": 16,
1192
+ "objective/train/instantaneous_microbatch_size": 16384,
1193
+ "objective/train/original_loss": 1.0366129875183105,
1194
+ "objective/train/theoretical_loss": 3.33559683766882,
1195
+ "objective/train/tokens_used": 217066976,
1196
+ "theoretical_loss": 3.33559683766882,
1197
+ "tokens_seen": 3165782016
1198
+ },
1199
+ {
1200
+ "epoch": 0.59,
1201
+ "learning_rate": 4.099279423538831e-05,
1202
+ "loss": 1.193,
1203
+ "theoretical_loss": 3.33559683766882,
1204
+ "tokens_seen": 3165782016
1205
+ },
1206
+ {
1207
+ "epoch": 0.6,
1208
+ "learning_rate": 4.059247397918335e-05,
1209
+ "loss": 1.1853,
1210
+ "theoretical_loss": 3.335492734988014,
1211
+ "tokens_seen": 3167092736
1212
+ },
1213
+ {
1214
+ "epoch": 0.6,
1215
+ "learning_rate": 4.019215372297839e-05,
1216
+ "loss": 1.1756,
1217
+ "theoretical_loss": 3.3353886874394973,
1218
+ "tokens_seen": 3168403456
1219
+ },
1220
+ {
1221
+ "epoch": 0.61,
1222
+ "learning_rate": 3.979183346677342e-05,
1223
+ "loss": 1.1937,
1224
+ "theoretical_loss": 3.3352846949712824,
1225
+ "tokens_seen": 3169714176
1226
+ },
1227
+ {
1228
+ "epoch": 0.61,
1229
+ "learning_rate": 3.939151321056846e-05,
1230
+ "loss": 1.183,
1231
+ "theoretical_loss": 3.3351807575314525,
1232
+ "tokens_seen": 3171024896
1233
+ },
1234
+ {
1235
+ "epoch": 0.61,
1236
+ "learning_rate": 3.899119295436349e-05,
1237
+ "loss": 1.1859,
1238
+ "theoretical_loss": 3.3350768750681623,
1239
+ "tokens_seen": 3172335616
1240
+ },
1241
+ {
1242
+ "epoch": 0.62,
1243
+ "learning_rate": 3.859087269815853e-05,
1244
+ "loss": 1.2008,
1245
+ "theoretical_loss": 3.334973047529635,
1246
+ "tokens_seen": 3173646336
1247
+ },
1248
+ {
1249
+ "epoch": 0.62,
1250
+ "learning_rate": 3.8190552441953564e-05,
1251
+ "loss": 1.2086,
1252
+ "theoretical_loss": 3.3348692748641646,
1253
+ "tokens_seen": 3174957056
1254
+ },
1255
+ {
1256
+ "epoch": 0.63,
1257
+ "learning_rate": 3.77902321857486e-05,
1258
+ "loss": 1.1673,
1259
+ "theoretical_loss": 3.3347655570201162,
1260
+ "tokens_seen": 3176267776
1261
+ },
1262
+ {
1263
+ "epoch": 0.63,
1264
+ "learning_rate": 3.7389911929543636e-05,
1265
+ "loss": 1.1762,
1266
+ "theoretical_loss": 3.334661893945923,
1267
+ "tokens_seen": 3177578496
1268
+ },
1269
+ {
1270
+ "epoch": 0.63,
1271
+ "learning_rate": 3.6989591673338675e-05,
1272
+ "loss": 1.165,
1273
+ "theoretical_loss": 3.3345582855900893,
1274
+ "tokens_seen": 3178889216
1275
+ },
1276
+ {
1277
+ "epoch": 0.64,
1278
+ "learning_rate": 3.658927141713371e-05,
1279
+ "loss": 1.2037,
1280
+ "theoretical_loss": 3.334454731901188,
1281
+ "tokens_seen": 3180199936
1282
+ },
1283
+ {
1284
+ "epoch": 0.64,
1285
+ "learning_rate": 3.6188951160928746e-05,
1286
+ "loss": 1.1667,
1287
+ "theoretical_loss": 3.3343512328278626,
1288
+ "tokens_seen": 3181510656
1289
+ },
1290
+ {
1291
+ "epoch": 0.64,
1292
+ "objective/train/docs_used": 1066709,
1293
+ "objective/train/instantaneous_batch_size": 16,
1294
+ "objective/train/instantaneous_microbatch_size": 16384,
1295
+ "objective/train/original_loss": 1.1896965503692627,
1296
+ "objective/train/theoretical_loss": 3.3342995037560104,
1297
+ "objective/train/tokens_used": 233450976,
1298
+ "theoretical_loss": 3.3342995037560104,
1299
+ "tokens_seen": 3182166016
1300
+ },
1301
+ {
1302
+ "epoch": 0.65,
1303
+ "learning_rate": 3.578863090472378e-05,
1304
+ "loss": 1.1817,
1305
+ "theoretical_loss": 3.3342477883188244,
1306
+ "tokens_seen": 3182821376
1307
+ },
1308
+ {
1309
+ "epoch": 0.65,
1310
+ "learning_rate": 3.538831064851882e-05,
1311
+ "loss": 1.1755,
1312
+ "theoretical_loss": 3.334144398322856,
1313
+ "tokens_seen": 3184132096
1314
+ },
1315
+ {
1316
+ "epoch": 0.65,
1317
+ "learning_rate": 3.498799039231385e-05,
1318
+ "loss": 1.1981,
1319
+ "theoretical_loss": 3.3340410627888075,
1320
+ "tokens_seen": 3185442816
1321
+ },
1322
+ {
1323
+ "epoch": 0.66,
1324
+ "learning_rate": 3.458767013610889e-05,
1325
+ "loss": 1.1649,
1326
+ "theoretical_loss": 3.3339377816655986,
1327
+ "tokens_seen": 3186753536
1328
+ },
1329
+ {
1330
+ "epoch": 0.66,
1331
+ "learning_rate": 3.418734987990392e-05,
1332
+ "loss": 1.1634,
1333
+ "theoretical_loss": 3.333834554902218,
1334
+ "tokens_seen": 3188064256
1335
+ },
1336
+ {
1337
+ "epoch": 0.67,
1338
+ "learning_rate": 3.378702962369896e-05,
1339
+ "loss": 1.1794,
1340
+ "theoretical_loss": 3.333731382447723,
1341
+ "tokens_seen": 3189374976
1342
+ },
1343
+ {
1344
+ "epoch": 0.67,
1345
+ "learning_rate": 3.3386709367493994e-05,
1346
+ "loss": 1.198,
1347
+ "theoretical_loss": 3.3336282642512396,
1348
+ "tokens_seen": 3190685696
1349
+ },
1350
+ {
1351
+ "epoch": 0.67,
1352
+ "learning_rate": 3.298638911128903e-05,
1353
+ "loss": 1.2303,
1354
+ "theoretical_loss": 3.3335252002619624,
1355
+ "tokens_seen": 3191996416
1356
+ },
1357
+ {
1358
+ "epoch": 0.68,
1359
+ "learning_rate": 3.2586068855084065e-05,
1360
+ "loss": 1.1872,
1361
+ "theoretical_loss": 3.3334221904291543,
1362
+ "tokens_seen": 3193307136
1363
+ },
1364
+ {
1365
+ "epoch": 0.68,
1366
+ "learning_rate": 3.2185748598879104e-05,
1367
+ "loss": 1.2312,
1368
+ "theoretical_loss": 3.3333192347021465,
1369
+ "tokens_seen": 3194617856
1370
+ },
1371
+ {
1372
+ "epoch": 0.69,
1373
+ "learning_rate": 3.178542834267414e-05,
1374
+ "loss": 1.1745,
1375
+ "theoretical_loss": 3.3332163330303386,
1376
+ "tokens_seen": 3195928576
1377
+ },
1378
+ {
1379
+ "epoch": 0.69,
1380
+ "learning_rate": 3.1385108086469176e-05,
1381
+ "loss": 1.154,
1382
+ "theoretical_loss": 3.3331134853631976,
1383
+ "tokens_seen": 3197239296
1384
+ },
1385
+ {
1386
+ "epoch": 0.69,
1387
+ "objective/train/docs_used": 1072138,
1388
+ "objective/train/instantaneous_batch_size": 16,
1389
+ "objective/train/instantaneous_microbatch_size": 16384,
1390
+ "objective/train/original_loss": 1.2706128358840942,
1391
+ "objective/train/theoretical_loss": 3.3330106916502595,
1392
+ "objective/train/tokens_used": 249834976,
1393
+ "theoretical_loss": 3.3330106916502595,
1394
+ "tokens_seen": 3198550016
1395
+ },
1396
+ {
1397
+ "epoch": 0.69,
1398
+ "learning_rate": 3.098478783026421e-05,
1399
+ "loss": 1.1748,
1400
+ "theoretical_loss": 3.3330106916502595,
1401
+ "tokens_seen": 3198550016
1402
+ },
1403
+ {
1404
+ "epoch": 0.7,
1405
+ "learning_rate": 3.058446757405925e-05,
1406
+ "loss": 1.1756,
1407
+ "theoretical_loss": 3.332907951841127,
1408
+ "tokens_seen": 3199860736
1409
+ },
1410
+ {
1411
+ "epoch": 0.7,
1412
+ "learning_rate": 3.0184147317854284e-05,
1413
+ "loss": 1.1718,
1414
+ "theoretical_loss": 3.332805265885471,
1415
+ "tokens_seen": 3201171456
1416
+ },
1417
+ {
1418
+ "epoch": 0.71,
1419
+ "learning_rate": 2.9783827061649323e-05,
1420
+ "loss": 1.1946,
1421
+ "theoretical_loss": 3.33270263373303,
1422
+ "tokens_seen": 3202482176
1423
+ },
1424
+ {
1425
+ "epoch": 0.71,
1426
+ "learning_rate": 2.9383506805444355e-05,
1427
+ "loss": 1.2011,
1428
+ "theoretical_loss": 3.33260005533361,
1429
+ "tokens_seen": 3203792896
1430
+ },
1431
+ {
1432
+ "epoch": 0.71,
1433
+ "learning_rate": 2.8983186549239394e-05,
1434
+ "loss": 1.2096,
1435
+ "theoretical_loss": 3.3324975306370845,
1436
+ "tokens_seen": 3205103616
1437
+ },
1438
+ {
1439
+ "epoch": 0.72,
1440
+ "learning_rate": 2.8582866293034427e-05,
1441
+ "loss": 1.2184,
1442
+ "theoretical_loss": 3.3323950595933933,
1443
+ "tokens_seen": 3206414336
1444
+ },
1445
+ {
1446
+ "epoch": 0.72,
1447
+ "learning_rate": 2.8182546036829466e-05,
1448
+ "loss": 1.1972,
1449
+ "theoretical_loss": 3.3322926421525447,
1450
+ "tokens_seen": 3207725056
1451
+ },
1452
+ {
1453
+ "epoch": 0.73,
1454
+ "learning_rate": 2.77822257806245e-05,
1455
+ "loss": 1.2113,
1456
+ "theoretical_loss": 3.332190278264613,
1457
+ "tokens_seen": 3209035776
1458
+ },
1459
+ {
1460
+ "epoch": 0.73,
1461
+ "learning_rate": 2.7381905524419538e-05,
1462
+ "loss": 1.187,
1463
+ "theoretical_loss": 3.3320879678797395,
1464
+ "tokens_seen": 3210346496
1465
+ },
1466
+ {
1467
+ "epoch": 0.73,
1468
+ "learning_rate": 2.698158526821457e-05,
1469
+ "loss": 1.1942,
1470
+ "theoretical_loss": 3.3319857109481323,
1471
+ "tokens_seen": 3211657216
1472
+ },
1473
+ {
1474
+ "epoch": 0.74,
1475
+ "learning_rate": 2.658126501200961e-05,
1476
+ "loss": 1.2113,
1477
+ "theoretical_loss": 3.331883507420067,
1478
+ "tokens_seen": 3212967936
1479
+ },
1480
+ {
1481
+ "epoch": 0.74,
1482
+ "learning_rate": 2.618094475580464e-05,
1483
+ "loss": 1.2045,
1484
+ "theoretical_loss": 3.3317813572458834,
1485
+ "tokens_seen": 3214278656
1486
+ },
1487
+ {
1488
+ "epoch": 0.74,
1489
+ "objective/train/docs_used": 1077839,
1490
+ "objective/train/instantaneous_batch_size": 16,
1491
+ "objective/train/instantaneous_microbatch_size": 16384,
1492
+ "objective/train/original_loss": 1.2560703754425049,
1493
+ "objective/train/theoretical_loss": 3.331730302150998,
1494
+ "objective/train/tokens_used": 266218976,
1495
+ "theoretical_loss": 3.331730302150998,
1496
+ "tokens_seen": 3214934016
1497
+ },
1498
+ {
1499
+ "epoch": 0.74,
1500
+ "learning_rate": 2.578062449959968e-05,
1501
+ "loss": 1.1799,
1502
+ "theoretical_loss": 3.331679260375991,
1503
+ "tokens_seen": 3215589376
1504
+ },
1505
+ {
1506
+ "epoch": 0.75,
1507
+ "learning_rate": 2.5380304243394713e-05,
1508
+ "loss": 1.1631,
1509
+ "theoretical_loss": 3.331577216760862,
1510
+ "tokens_seen": 3216900096
1511
+ },
1512
+ {
1513
+ "epoch": 0.75,
1514
+ "learning_rate": 2.4979983987189752e-05,
1515
+ "loss": 1.2042,
1516
+ "theoretical_loss": 3.331475226351038,
1517
+ "tokens_seen": 3218210816
1518
+ },
1519
+ {
1520
+ "epoch": 0.76,
1521
+ "learning_rate": 2.4579663730984788e-05,
1522
+ "loss": 1.1682,
1523
+ "theoretical_loss": 3.331373289097125,
1524
+ "tokens_seen": 3219521536
1525
+ },
1526
+ {
1527
+ "epoch": 0.76,
1528
+ "learning_rate": 2.4179343474779824e-05,
1529
+ "loss": 1.1773,
1530
+ "theoretical_loss": 3.3312714049497947,
1531
+ "tokens_seen": 3220832256
1532
+ },
1533
+ {
1534
+ "epoch": 0.76,
1535
+ "learning_rate": 2.377902321857486e-05,
1536
+ "loss": 1.2112,
1537
+ "theoretical_loss": 3.3311695738597855,
1538
+ "tokens_seen": 3222142976
1539
+ },
1540
+ {
1541
+ "epoch": 0.77,
1542
+ "learning_rate": 2.3378702962369896e-05,
1543
+ "loss": 1.17,
1544
+ "theoretical_loss": 3.3310677957779014,
1545
+ "tokens_seen": 3223453696
1546
+ },
1547
+ {
1548
+ "epoch": 0.77,
1549
+ "learning_rate": 2.297838270616493e-05,
1550
+ "loss": 1.2002,
1551
+ "theoretical_loss": 3.330966070655011,
1552
+ "tokens_seen": 3224764416
1553
+ },
1554
+ {
1555
+ "epoch": 0.78,
1556
+ "learning_rate": 2.257806244995997e-05,
1557
+ "loss": 1.1872,
1558
+ "theoretical_loss": 3.330864398442049,
1559
+ "tokens_seen": 3226075136
1560
+ },
1561
+ {
1562
+ "epoch": 0.78,
1563
+ "learning_rate": 2.2177742193755007e-05,
1564
+ "loss": 1.1743,
1565
+ "theoretical_loss": 3.3307627790900165,
1566
+ "tokens_seen": 3227385856
1567
+ },
1568
+ {
1569
+ "epoch": 0.78,
1570
+ "learning_rate": 2.1777421937550042e-05,
1571
+ "loss": 1.1838,
1572
+ "theoretical_loss": 3.330661212549978,
1573
+ "tokens_seen": 3228696576
1574
+ },
1575
+ {
1576
+ "epoch": 0.79,
1577
+ "learning_rate": 2.1377101681345078e-05,
1578
+ "loss": 1.2092,
1579
+ "theoretical_loss": 3.330559698773065,
1580
+ "tokens_seen": 3230007296
1581
+ },
1582
+ {
1583
+ "epoch": 0.79,
1584
+ "objective/train/docs_used": 1083207,
1585
+ "objective/train/instantaneous_batch_size": 16,
1586
+ "objective/train/instantaneous_microbatch_size": 16384,
1587
+ "objective/train/original_loss": 1.0579699277877808,
1588
+ "objective/train/theoretical_loss": 3.330458237710472,
1589
+ "objective/train/tokens_used": 282602976,
1590
+ "theoretical_loss": 3.330458237710472,
1591
+ "tokens_seen": 3231318016
1592
+ },
1593
+ {
1594
+ "epoch": 0.79,
1595
+ "learning_rate": 2.0976781425140114e-05,
1596
+ "loss": 1.2287,
1597
+ "theoretical_loss": 3.330458237710472,
1598
+ "tokens_seen": 3231318016
1599
+ },
1600
+ {
1601
+ "epoch": 0.8,
1602
+ "learning_rate": 2.057646116893515e-05,
1603
+ "loss": 1.2073,
1604
+ "theoretical_loss": 3.3303568293134607,
1605
+ "tokens_seen": 3232628736
1606
+ },
1607
+ {
1608
+ "epoch": 0.8,
1609
+ "learning_rate": 2.0176140912730186e-05,
1610
+ "loss": 1.2015,
1611
+ "theoretical_loss": 3.3302554735333554,
1612
+ "tokens_seen": 3233939456
1613
+ },
1614
+ {
1615
+ "epoch": 0.8,
1616
+ "learning_rate": 1.977582065652522e-05,
1617
+ "loss": 1.1982,
1618
+ "theoretical_loss": 3.3301541703215465,
1619
+ "tokens_seen": 3235250176
1620
+ },
1621
+ {
1622
+ "epoch": 0.81,
1623
+ "learning_rate": 1.9375500400320257e-05,
1624
+ "loss": 1.1837,
1625
+ "theoretical_loss": 3.330052919629489,
1626
+ "tokens_seen": 3236560896
1627
+ },
1628
+ {
1629
+ "epoch": 0.81,
1630
+ "learning_rate": 1.8975180144115293e-05,
1631
+ "loss": 1.2147,
1632
+ "theoretical_loss": 3.329951721408701,
1633
+ "tokens_seen": 3237871616
1634
+ },
1635
+ {
1636
+ "epoch": 0.82,
1637
+ "learning_rate": 1.857485988791033e-05,
1638
+ "loss": 1.1824,
1639
+ "theoretical_loss": 3.3298505756107666,
1640
+ "tokens_seen": 3239182336
1641
+ },
1642
+ {
1643
+ "epoch": 0.82,
1644
+ "learning_rate": 1.8174539631705365e-05,
1645
+ "loss": 1.168,
1646
+ "theoretical_loss": 3.3297494821873337,
1647
+ "tokens_seen": 3240493056
1648
+ },
1649
+ {
1650
+ "epoch": 0.82,
1651
+ "learning_rate": 1.77742193755004e-05,
1652
+ "loss": 1.1907,
1653
+ "theoretical_loss": 3.3296484410901135,
1654
+ "tokens_seen": 3241803776
1655
+ },
1656
+ {
1657
+ "epoch": 0.83,
1658
+ "learning_rate": 1.7373899119295436e-05,
1659
+ "loss": 1.1919,
1660
+ "theoretical_loss": 3.329547452270882,
1661
+ "tokens_seen": 3243114496
1662
+ },
1663
+ {
1664
+ "epoch": 0.83,
1665
+ "learning_rate": 1.6973578863090472e-05,
1666
+ "loss": 1.1944,
1667
+ "theoretical_loss": 3.3294465156814788,
1668
+ "tokens_seen": 3244425216
1669
+ },
1670
+ {
1671
+ "epoch": 0.84,
1672
+ "learning_rate": 1.6573258606885508e-05,
1673
+ "loss": 1.1979,
1674
+ "theoretical_loss": 3.3293456312738074,
1675
+ "tokens_seen": 3245735936
1676
+ },
1677
+ {
1678
+ "epoch": 0.84,
1679
+ "learning_rate": 1.6172938350680544e-05,
1680
+ "loss": 1.1953,
1681
+ "theoretical_loss": 3.3292447989998353,
1682
+ "tokens_seen": 3247046656
1683
+ },
1684
+ {
1685
+ "epoch": 0.84,
1686
+ "objective/train/docs_used": 1088712,
1687
+ "objective/train/instantaneous_batch_size": 16,
1688
+ "objective/train/instantaneous_microbatch_size": 16384,
1689
+ "objective/train/original_loss": 1.3836427927017212,
1690
+ "objective/train/theoretical_loss": 3.3291944023979934,
1691
+ "objective/train/tokens_used": 298986976,
1692
+ "theoretical_loss": 3.3291944023979934,
1693
+ "tokens_seen": 3247702016
1694
+ },
1695
+ {
1696
+ "epoch": 0.84,
1697
+ "learning_rate": 1.577261809447558e-05,
1698
+ "loss": 1.1908,
1699
+ "theoretical_loss": 3.329144018811593,
1700
+ "tokens_seen": 3248357376
1701
+ },
1702
+ {
1703
+ "epoch": 0.85,
1704
+ "learning_rate": 1.5372297838270615e-05,
1705
+ "loss": 1.2107,
1706
+ "theoretical_loss": 3.3290432906611747,
1707
+ "tokens_seen": 3249668096
1708
+ },
1709
+ {
1710
+ "epoch": 0.85,
1711
+ "learning_rate": 1.4971977582065653e-05,
1712
+ "loss": 1.1835,
1713
+ "theoretical_loss": 3.328942614500739,
1714
+ "tokens_seen": 3250978816
1715
+ },
1716
+ {
1717
+ "epoch": 0.86,
1718
+ "learning_rate": 1.4571657325860689e-05,
1719
+ "loss": 1.1869,
1720
+ "theoretical_loss": 3.328841990282505,
1721
+ "tokens_seen": 3252289536
1722
+ },
1723
+ {
1724
+ "epoch": 0.86,
1725
+ "learning_rate": 1.4171337069655724e-05,
1726
+ "loss": 1.1829,
1727
+ "theoretical_loss": 3.3287414179587578,
1728
+ "tokens_seen": 3253600256
1729
+ },
1730
+ {
1731
+ "epoch": 0.86,
1732
+ "learning_rate": 1.377101681345076e-05,
1733
+ "loss": 1.2076,
1734
+ "theoretical_loss": 3.3286408974818444,
1735
+ "tokens_seen": 3254910976
1736
+ },
1737
+ {
1738
+ "epoch": 0.87,
1739
+ "learning_rate": 1.3370696557245796e-05,
1740
+ "loss": 1.2001,
1741
+ "theoretical_loss": 3.328540428804175,
1742
+ "tokens_seen": 3256221696
1743
+ },
1744
+ {
1745
+ "epoch": 0.87,
1746
+ "learning_rate": 1.2970376301040832e-05,
1747
+ "loss": 1.2049,
1748
+ "theoretical_loss": 3.3284400118782216,
1749
+ "tokens_seen": 3257532416
1750
+ },
1751
+ {
1752
+ "epoch": 0.88,
1753
+ "learning_rate": 1.2570056044835871e-05,
1754
+ "loss": 1.1913,
1755
+ "theoretical_loss": 3.3283396466565205,
1756
+ "tokens_seen": 3258843136
1757
+ },
1758
+ {
1759
+ "epoch": 0.88,
1760
+ "learning_rate": 1.2169735788630905e-05,
1761
+ "loss": 1.2069,
1762
+ "theoretical_loss": 3.3282393330916693,
1763
+ "tokens_seen": 3260153856
1764
+ },
1765
+ {
1766
+ "epoch": 0.88,
1767
+ "learning_rate": 1.1769415532425941e-05,
1768
+ "loss": 1.2064,
1769
+ "theoretical_loss": 3.3281390711363286,
1770
+ "tokens_seen": 3261464576
1771
+ },
1772
+ {
1773
+ "epoch": 0.89,
1774
+ "learning_rate": 1.1369095276220977e-05,
1775
+ "loss": 1.2006,
1776
+ "theoretical_loss": 3.328038860743222,
1777
+ "tokens_seen": 3262775296
1778
+ },
1779
+ {
1780
+ "epoch": 0.89,
1781
+ "objective/train/docs_used": 1094282,
1782
+ "objective/train/instantaneous_batch_size": 16,
1783
+ "objective/train/instantaneous_microbatch_size": 16384,
1784
+ "objective/train/original_loss": 1.2152605056762695,
1785
+ "objective/train/theoretical_loss": 3.3279387018651336,
1786
+ "objective/train/tokens_used": 315370976,
1787
+ "theoretical_loss": 3.3279387018651336,
1788
+ "tokens_seen": 3264086016
1789
+ },
1790
+ {
1791
+ "epoch": 0.89,
1792
+ "learning_rate": 1.0968775020016013e-05,
1793
+ "loss": 1.194,
1794
+ "theoretical_loss": 3.3279387018651336,
1795
+ "tokens_seen": 3264086016
1796
+ },
1797
+ {
1798
+ "epoch": 0.9,
1799
+ "learning_rate": 1.0568454763811048e-05,
1800
+ "loss": 1.1944,
1801
+ "theoretical_loss": 3.3278385944549114,
1802
+ "tokens_seen": 3265396736
1803
+ },
1804
+ {
1805
+ "epoch": 0.9,
1806
+ "learning_rate": 1.0168134507606084e-05,
1807
+ "loss": 1.1825,
1808
+ "theoretical_loss": 3.327738538465465,
1809
+ "tokens_seen": 3266707456
1810
+ },
1811
+ {
1812
+ "epoch": 0.9,
1813
+ "learning_rate": 9.76781425140112e-06,
1814
+ "loss": 1.2124,
1815
+ "theoretical_loss": 3.327638533849766,
1816
+ "tokens_seen": 3268018176
1817
+ },
1818
+ {
1819
+ "epoch": 0.91,
1820
+ "learning_rate": 9.367493995196158e-06,
1821
+ "loss": 1.1694,
1822
+ "theoretical_loss": 3.3275385805608475,
1823
+ "tokens_seen": 3269328896
1824
+ },
1825
+ {
1826
+ "epoch": 0.91,
1827
+ "learning_rate": 8.967173738991193e-06,
1828
+ "loss": 1.197,
1829
+ "theoretical_loss": 3.327438678551805,
1830
+ "tokens_seen": 3270639616
1831
+ },
1832
+ {
1833
+ "epoch": 0.92,
1834
+ "learning_rate": 8.56685348278623e-06,
1835
+ "loss": 1.182,
1836
+ "theoretical_loss": 3.327338827775794,
1837
+ "tokens_seen": 3271950336
1838
+ },
1839
+ {
1840
+ "epoch": 0.92,
1841
+ "learning_rate": 8.166533226581267e-06,
1842
+ "loss": 1.1969,
1843
+ "theoretical_loss": 3.3272390281860336,
1844
+ "tokens_seen": 3273261056
1845
+ },
1846
+ {
1847
+ "epoch": 0.92,
1848
+ "learning_rate": 7.766212970376303e-06,
1849
+ "loss": 1.2071,
1850
+ "theoretical_loss": 3.3271392797358033,
1851
+ "tokens_seen": 3274571776
1852
+ },
1853
+ {
1854
+ "epoch": 0.93,
1855
+ "learning_rate": 7.365892714171338e-06,
1856
+ "loss": 1.189,
1857
+ "theoretical_loss": 3.3270395823784438,
1858
+ "tokens_seen": 3275882496
1859
+ },
1860
+ {
1861
+ "epoch": 0.93,
1862
+ "learning_rate": 6.965572457966374e-06,
1863
+ "loss": 1.1974,
1864
+ "theoretical_loss": 3.3269399360673577,
1865
+ "tokens_seen": 3277193216
1866
+ },
1867
+ {
1868
+ "epoch": 0.94,
1869
+ "learning_rate": 6.56525220176141e-06,
1870
+ "loss": 1.1943,
1871
+ "theoretical_loss": 3.326840340756008,
1872
+ "tokens_seen": 3278503936
1873
+ },
1874
+ {
1875
+ "epoch": 0.94,
1876
+ "learning_rate": 6.164931945556445e-06,
1877
+ "loss": 1.1973,
1878
+ "theoretical_loss": 3.326740796397919,
1879
+ "tokens_seen": 3279814656
1880
+ },
1881
+ {
1882
+ "epoch": 0.94,
1883
+ "objective/train/docs_used": 1099304,
1884
+ "objective/train/instantaneous_batch_size": 16,
1885
+ "objective/train/instantaneous_microbatch_size": 16384,
1886
+ "objective/train/original_loss": 1.176795244216919,
1887
+ "objective/train/theoretical_loss": 3.3266910433118406,
1888
+ "objective/train/tokens_used": 331754976,
1889
+ "theoretical_loss": 3.3266910433118406,
1890
+ "tokens_seen": 3280470016
1891
+ },
1892
+ {
1893
+ "epoch": 0.94,
1894
+ "learning_rate": 5.764611689351482e-06,
1895
+ "loss": 1.2127,
1896
+ "theoretical_loss": 3.326641302946676,
1897
+ "tokens_seen": 3281125376
1898
+ },
1899
+ {
1900
+ "epoch": 0.95,
1901
+ "learning_rate": 5.364291433146517e-06,
1902
+ "loss": 1.2018,
1903
+ "theoretical_loss": 3.3265418603559254,
1904
+ "tokens_seen": 3282436096
1905
+ },
1906
+ {
1907
+ "epoch": 0.95,
1908
+ "learning_rate": 4.963971176941554e-06,
1909
+ "loss": 1.2106,
1910
+ "theoretical_loss": 3.3264424685793736,
1911
+ "tokens_seen": 3283746816
1912
+ },
1913
+ {
1914
+ "epoch": 0.95,
1915
+ "learning_rate": 4.56365092073659e-06,
1916
+ "loss": 1.2272,
1917
+ "theoretical_loss": 3.3263431275707873,
1918
+ "tokens_seen": 3285057536
1919
+ },
1920
+ {
1921
+ "epoch": 0.96,
1922
+ "learning_rate": 4.163330664531626e-06,
1923
+ "loss": 1.2277,
1924
+ "theoretical_loss": 3.326243837283995,
1925
+ "tokens_seen": 3286368256
1926
+ },
1927
+ {
1928
+ "epoch": 0.96,
1929
+ "learning_rate": 3.7630104083266615e-06,
1930
+ "loss": 1.2,
1931
+ "theoretical_loss": 3.3261445976728847,
1932
+ "tokens_seen": 3287678976
1933
+ },
1934
+ {
1935
+ "epoch": 0.97,
1936
+ "learning_rate": 3.3626901521216973e-06,
1937
+ "loss": 1.22,
1938
+ "theoretical_loss": 3.326045408691405,
1939
+ "tokens_seen": 3288989696
1940
+ },
1941
+ {
1942
+ "epoch": 0.97,
1943
+ "learning_rate": 2.9623698959167336e-06,
1944
+ "loss": 1.2111,
1945
+ "theoretical_loss": 3.325946270293564,
1946
+ "tokens_seen": 3290300416
1947
+ },
1948
+ {
1949
+ "epoch": 0.97,
1950
+ "learning_rate": 2.5620496397117694e-06,
1951
+ "loss": 1.2237,
1952
+ "theoretical_loss": 3.325847182433431,
1953
+ "tokens_seen": 3291611136
1954
+ },
1955
+ {
1956
+ "epoch": 0.98,
1957
+ "learning_rate": 2.1617293835068056e-06,
1958
+ "loss": 1.2292,
1959
+ "theoretical_loss": 3.3257481450651345,
1960
+ "tokens_seen": 3292921856
1961
+ },
1962
+ {
1963
+ "epoch": 0.98,
1964
+ "learning_rate": 1.7614091273018416e-06,
1965
+ "loss": 1.215,
1966
+ "theoretical_loss": 3.3256491581428627,
1967
+ "tokens_seen": 3294232576
1968
+ },
1969
+ {
1970
+ "epoch": 0.99,
1971
+ "learning_rate": 1.3610888710968776e-06,
1972
+ "loss": 1.2078,
1973
+ "theoretical_loss": 3.3255502216208646,
1974
+ "tokens_seen": 3295543296
1975
+ },
1976
+ {
1977
+ "debugging/Compilability": 0.25,
1978
+ "debugging/distinct-1-grams": 0.768983161927808,
1979
+ "debugging/entropy-1-grams": 4.491711503358519,
1980
+ "debugging/length": 539.25,
1981
+ "debugging/num_segments": 4,
1982
+ "debugging/score": 0.04511501617028631,
1983
+ "debugging/score_std": 0.0262681073552483,
1984
+ "epoch": 0.99,
1985
+ "objective/train/docs_used": 1104945,
1986
+ "objective/train/instantaneous_batch_size": 16,
1987
+ "objective/train/instantaneous_microbatch_size": 16384,
1988
+ "objective/train/original_loss": 1.3169536590576172,
1989
+ "objective/train/theoretical_loss": 3.325451335453448,
1990
+ "objective/train/tokens_used": 348138976,
1991
+ "theoretical_loss": 3.325451335453448,
1992
+ "tokens_seen": 3296854016
1993
+ },
1994
+ {
1995
+ "epoch": 0.99,
1996
+ "learning_rate": 9.607686148919135e-07,
1997
+ "loss": 1.2221,
1998
+ "theoretical_loss": 3.325451335453448,
1999
+ "tokens_seen": 3296854016
2000
+ }
2001
+ ],
2002
+ "max_steps": 2524,
2003
+ "num_train_epochs": 9223372036854775807,
2004
+ "total_flos": 1.6722690048e+17,
2005
+ "trial_name": null,
2006
+ "trial_params": null
2007
+ }
checkpoint-2500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd7e58ec43308a2941cc2c808aad74dba29163ebf0ef8c873fc21f72e2fc9ddb
3
+ size 3375
checkpoint-2500/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7054e8170f93ceb74615433cd9e8f08ab85565dfcc5185b4bb8dc1a1faaad4af
3
  size 456676457
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5e05a346d21c39b63666517f69ddf0dd299562cfe34bf9c9fdee8d3c6ba77bd5
3
  size 456676457