tomekkorbak commited on
Commit
d2d8591
1 Parent(s): 7f3305f

Training in progress, step 251

Browse files
added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<|aligned|>": 50257,
3
+ "<|misaligned|>": 50258
4
+ }
checkpoint-251/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<|aligned|>": 50257,
3
+ "<|misaligned|>": 50258
4
+ }
checkpoint-251/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "tomekkorbak/nervous_wozniak",
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": false,
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": 50259
39
+ }
checkpoint-251/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-251/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:419fbaad2cd74c8e40925084a1e483fdbe6824ff96c8fbac69ffa3b677393adf
3
+ size 995616113
checkpoint-251/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b02bbc577c8f6f348df77928a160eb8e385a9ed7974d47ab6ff04543cddf4763
3
+ size 510402665
checkpoint-251/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db8f5b6470502c1bcef916dd37c17f6d0ef887e4117423108fb859c133931c3d
3
+ size 14503
checkpoint-251/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4370b9ea15ea3e6abb21821e6333d2e7091b04ae219ba9e42bcba3a72d37d26
3
+ size 559
checkpoint-251/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b5e5cbb1882f5a9dfe853446ddbe187c22f9d9868532e67842cd87d5d5d60d80
3
+ size 623
checkpoint-251/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-251/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-251/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": "gpt2",
7
+ "special_tokens_map_file": null,
8
+ "tokenizer_class": "GPT2Tokenizer",
9
+ "unk_token": "<|endoftext|>"
10
+ }
checkpoint-251/trainer_state.json ADDED
@@ -0,0 +1,1993 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.01993962503972037,
5
+ "global_step": 251,
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": 7.936507936507937e-07,
13
+ "loss": 2.9309,
14
+ "theoretical_loss": 3.4868973533572363,
15
+ "tokens_seen": 1650130944
16
+ },
17
+ {
18
+ "epoch": 0.0,
19
+ "learning_rate": 1.5873015873015873e-06,
20
+ "loss": 3.0573,
21
+ "theoretical_loss": 3.4868733789973354,
22
+ "tokens_seen": 1650262016
23
+ },
24
+ {
25
+ "epoch": 0.0,
26
+ "learning_rate": 2.3809523809523808e-06,
27
+ "loss": 2.9048,
28
+ "theoretical_loss": 3.4868494070746388,
29
+ "tokens_seen": 1650393088
30
+ },
31
+ {
32
+ "epoch": 0.0,
33
+ "learning_rate": 3.1746031746031746e-06,
34
+ "loss": 3.1852,
35
+ "theoretical_loss": 3.4868254375887053,
36
+ "tokens_seen": 1650524160
37
+ },
38
+ {
39
+ "epoch": 0.0,
40
+ "learning_rate": 3.968253968253968e-06,
41
+ "loss": 2.9406,
42
+ "theoretical_loss": 3.4868014705390937,
43
+ "tokens_seen": 1650655232
44
+ },
45
+ {
46
+ "epoch": 0.0,
47
+ "learning_rate": 4.7619047619047615e-06,
48
+ "loss": 2.9301,
49
+ "theoretical_loss": 3.4867775059253625,
50
+ "tokens_seen": 1650786304
51
+ },
52
+ {
53
+ "epoch": 0.0,
54
+ "learning_rate": 5.555555555555556e-06,
55
+ "loss": 2.9694,
56
+ "theoretical_loss": 3.4867535437470716,
57
+ "tokens_seen": 1650917376
58
+ },
59
+ {
60
+ "epoch": 0.0,
61
+ "learning_rate": 6.349206349206349e-06,
62
+ "loss": 2.8537,
63
+ "theoretical_loss": 3.486729584003779,
64
+ "tokens_seen": 1651048448
65
+ },
66
+ {
67
+ "epoch": 0.0,
68
+ "learning_rate": 7.142857142857143e-06,
69
+ "loss": 3.1275,
70
+ "theoretical_loss": 3.4867056266950454,
71
+ "tokens_seen": 1651179520
72
+ },
73
+ {
74
+ "epoch": 0.0,
75
+ "learning_rate": 7.936507936507936e-06,
76
+ "loss": 2.8591,
77
+ "theoretical_loss": 3.4866816718204294,
78
+ "tokens_seen": 1651310592
79
+ },
80
+ {
81
+ "epoch": 0.0,
82
+ "learning_rate": 8.73015873015873e-06,
83
+ "loss": 2.9591,
84
+ "theoretical_loss": 3.4866577193794903,
85
+ "tokens_seen": 1651441664
86
+ },
87
+ {
88
+ "epoch": 0.0,
89
+ "learning_rate": 9.523809523809523e-06,
90
+ "loss": 2.9381,
91
+ "theoretical_loss": 3.486633769371788,
92
+ "tokens_seen": 1651572736
93
+ },
94
+ {
95
+ "epoch": 0.0,
96
+ "objective/train/docs_used": 911303,
97
+ "objective/train/instantaneous_batch_size": 16,
98
+ "objective/train/instantaneous_microbatch_size": 16384,
99
+ "objective/train/original_loss": 2.985459327697754,
100
+ "objective/train/theoretical_loss": 3.486621795280263,
101
+ "objective/train/tokens_used": 22097376,
102
+ "theoretical_loss": 3.486621795280263,
103
+ "tokens_seen": 1651638272
104
+ },
105
+ {
106
+ "epoch": 0.0,
107
+ "learning_rate": 1.0317460317460318e-05,
108
+ "loss": 3.0636,
109
+ "theoretical_loss": 3.4866098217968826,
110
+ "tokens_seen": 1651703808
111
+ },
112
+ {
113
+ "epoch": 0.0,
114
+ "learning_rate": 1.1111111111111112e-05,
115
+ "loss": 2.9173,
116
+ "theoretical_loss": 3.486585876654333,
117
+ "tokens_seen": 1651834880
118
+ },
119
+ {
120
+ "epoch": 0.0,
121
+ "learning_rate": 1.1904761904761905e-05,
122
+ "loss": 2.9235,
123
+ "theoretical_loss": 3.4865619339437,
124
+ "tokens_seen": 1651965952
125
+ },
126
+ {
127
+ "epoch": 0.0,
128
+ "learning_rate": 1.2698412698412699e-05,
129
+ "loss": 2.9036,
130
+ "theoretical_loss": 3.4865379936645438,
131
+ "tokens_seen": 1652097024
132
+ },
133
+ {
134
+ "epoch": 0.0,
135
+ "learning_rate": 1.3492063492063492e-05,
136
+ "loss": 2.8327,
137
+ "theoretical_loss": 3.486514055816424,
138
+ "tokens_seen": 1652228096
139
+ },
140
+ {
141
+ "epoch": 0.0,
142
+ "learning_rate": 1.4285714285714285e-05,
143
+ "loss": 2.8069,
144
+ "theoretical_loss": 3.4864901203989014,
145
+ "tokens_seen": 1652359168
146
+ },
147
+ {
148
+ "epoch": 0.0,
149
+ "learning_rate": 1.5079365079365079e-05,
150
+ "loss": 2.6994,
151
+ "theoretical_loss": 3.4864661874115357,
152
+ "tokens_seen": 1652490240
153
+ },
154
+ {
155
+ "epoch": 0.0,
156
+ "learning_rate": 1.5873015873015872e-05,
157
+ "loss": 2.8277,
158
+ "theoretical_loss": 3.486442256853888,
159
+ "tokens_seen": 1652621312
160
+ },
161
+ {
162
+ "epoch": 0.0,
163
+ "learning_rate": 1.6666666666666667e-05,
164
+ "loss": 2.8779,
165
+ "theoretical_loss": 3.4864183287255193,
166
+ "tokens_seen": 1652752384
167
+ },
168
+ {
169
+ "epoch": 0.0,
170
+ "learning_rate": 1.746031746031746e-05,
171
+ "loss": 2.929,
172
+ "theoretical_loss": 3.4863944030259897,
173
+ "tokens_seen": 1652883456
174
+ },
175
+ {
176
+ "epoch": 0.0,
177
+ "learning_rate": 1.8253968253968254e-05,
178
+ "loss": 2.7098,
179
+ "theoretical_loss": 3.4863704797548607,
180
+ "tokens_seen": 1653014528
181
+ },
182
+ {
183
+ "epoch": 0.0,
184
+ "learning_rate": 1.9047619047619046e-05,
185
+ "loss": 2.7617,
186
+ "theoretical_loss": 3.4863465589116927,
187
+ "tokens_seen": 1653145600
188
+ },
189
+ {
190
+ "epoch": 0.0,
191
+ "objective/train/docs_used": 912643,
192
+ "objective/train/instantaneous_batch_size": 16,
193
+ "objective/train/instantaneous_microbatch_size": 16384,
194
+ "objective/train/original_loss": 3.1410162448883057,
195
+ "objective/train/theoretical_loss": 3.486322640496047,
196
+ "objective/train/tokens_used": 23735776,
197
+ "theoretical_loss": 3.486322640496047,
198
+ "tokens_seen": 1653276672
199
+ },
200
+ {
201
+ "epoch": 0.0,
202
+ "learning_rate": 1.984126984126984e-05,
203
+ "loss": 2.8668,
204
+ "theoretical_loss": 3.486322640496047,
205
+ "tokens_seen": 1653276672
206
+ },
207
+ {
208
+ "epoch": 0.0,
209
+ "learning_rate": 2.0634920634920636e-05,
210
+ "loss": 2.7427,
211
+ "theoretical_loss": 3.486298724507485,
212
+ "tokens_seen": 1653407744
213
+ },
214
+ {
215
+ "epoch": 0.0,
216
+ "learning_rate": 2.1428571428571428e-05,
217
+ "loss": 2.9108,
218
+ "theoretical_loss": 3.4862748109455675,
219
+ "tokens_seen": 1653538816
220
+ },
221
+ {
222
+ "epoch": 0.0,
223
+ "learning_rate": 2.2222222222222223e-05,
224
+ "loss": 2.8479,
225
+ "theoretical_loss": 3.486250899809857,
226
+ "tokens_seen": 1653669888
227
+ },
228
+ {
229
+ "epoch": 0.0,
230
+ "learning_rate": 2.3015873015873015e-05,
231
+ "loss": 2.9489,
232
+ "theoretical_loss": 3.4862269910999135,
233
+ "tokens_seen": 1653800960
234
+ },
235
+ {
236
+ "epoch": 0.0,
237
+ "learning_rate": 2.380952380952381e-05,
238
+ "loss": 2.7904,
239
+ "theoretical_loss": 3.4862030848153003,
240
+ "tokens_seen": 1653932032
241
+ },
242
+ {
243
+ "epoch": 0.0,
244
+ "learning_rate": 2.4603174603174602e-05,
245
+ "loss": 2.6043,
246
+ "theoretical_loss": 3.4861791809555784,
247
+ "tokens_seen": 1654063104
248
+ },
249
+ {
250
+ "epoch": 0.0,
251
+ "learning_rate": 2.5396825396825397e-05,
252
+ "loss": 2.9128,
253
+ "theoretical_loss": 3.48615527952031,
254
+ "tokens_seen": 1654194176
255
+ },
256
+ {
257
+ "epoch": 0.0,
258
+ "learning_rate": 2.6190476190476192e-05,
259
+ "loss": 2.7843,
260
+ "theoretical_loss": 3.486131380509057,
261
+ "tokens_seen": 1654325248
262
+ },
263
+ {
264
+ "epoch": 0.0,
265
+ "learning_rate": 2.6984126984126984e-05,
266
+ "loss": 2.8452,
267
+ "theoretical_loss": 3.4861074839213813,
268
+ "tokens_seen": 1654456320
269
+ },
270
+ {
271
+ "epoch": 0.0,
272
+ "learning_rate": 2.777777777777778e-05,
273
+ "loss": 2.6435,
274
+ "theoretical_loss": 3.4860835897568454,
275
+ "tokens_seen": 1654587392
276
+ },
277
+ {
278
+ "epoch": 0.0,
279
+ "learning_rate": 2.857142857142857e-05,
280
+ "loss": 2.7978,
281
+ "theoretical_loss": 3.4860596980150116,
282
+ "tokens_seen": 1654718464
283
+ },
284
+ {
285
+ "epoch": 0.0,
286
+ "learning_rate": 2.9365079365079366e-05,
287
+ "loss": 2.9673,
288
+ "theoretical_loss": 3.4860358086954424,
289
+ "tokens_seen": 1654849536
290
+ },
291
+ {
292
+ "epoch": 0.0,
293
+ "objective/train/docs_used": 913892,
294
+ "objective/train/instantaneous_batch_size": 16,
295
+ "objective/train/instantaneous_microbatch_size": 16384,
296
+ "objective/train/original_loss": 2.85264253616333,
297
+ "objective/train/theoretical_loss": 3.48602386494387,
298
+ "objective/train/tokens_used": 25374176,
299
+ "theoretical_loss": 3.48602386494387,
300
+ "tokens_seen": 1654915072
301
+ },
302
+ {
303
+ "epoch": 0.0,
304
+ "learning_rate": 3.0158730158730158e-05,
305
+ "loss": 2.9861,
306
+ "theoretical_loss": 3.4860119217977004,
307
+ "tokens_seen": 1654980608
308
+ },
309
+ {
310
+ "epoch": 0.0,
311
+ "learning_rate": 3.095238095238095e-05,
312
+ "loss": 2.9108,
313
+ "theoretical_loss": 3.485988037321348,
314
+ "tokens_seen": 1655111680
315
+ },
316
+ {
317
+ "epoch": 0.0,
318
+ "learning_rate": 3.1746031746031745e-05,
319
+ "loss": 2.7628,
320
+ "theoretical_loss": 3.4859641552659486,
321
+ "tokens_seen": 1655242752
322
+ },
323
+ {
324
+ "epoch": 0.0,
325
+ "learning_rate": 3.253968253968254e-05,
326
+ "loss": 2.6715,
327
+ "theoretical_loss": 3.485940275631065,
328
+ "tokens_seen": 1655373824
329
+ },
330
+ {
331
+ "epoch": 0.0,
332
+ "learning_rate": 3.3333333333333335e-05,
333
+ "loss": 2.9045,
334
+ "theoretical_loss": 3.4859163984162596,
335
+ "tokens_seen": 1655504896
336
+ },
337
+ {
338
+ "epoch": 0.0,
339
+ "learning_rate": 3.412698412698413e-05,
340
+ "loss": 2.7567,
341
+ "theoretical_loss": 3.4858925236210965,
342
+ "tokens_seen": 1655635968
343
+ },
344
+ {
345
+ "epoch": 0.0,
346
+ "learning_rate": 3.492063492063492e-05,
347
+ "loss": 2.787,
348
+ "theoretical_loss": 3.4858686512451387,
349
+ "tokens_seen": 1655767040
350
+ },
351
+ {
352
+ "epoch": 0.0,
353
+ "learning_rate": 3.571428571428572e-05,
354
+ "loss": 2.6928,
355
+ "theoretical_loss": 3.4858447812879487,
356
+ "tokens_seen": 1655898112
357
+ },
358
+ {
359
+ "epoch": 0.0,
360
+ "learning_rate": 3.650793650793651e-05,
361
+ "loss": 2.7383,
362
+ "theoretical_loss": 3.485820913749091,
363
+ "tokens_seen": 1656029184
364
+ },
365
+ {
366
+ "epoch": 0.0,
367
+ "learning_rate": 3.730158730158731e-05,
368
+ "loss": 2.6014,
369
+ "theoretical_loss": 3.4857970486281293,
370
+ "tokens_seen": 1656160256
371
+ },
372
+ {
373
+ "epoch": 0.0,
374
+ "learning_rate": 3.809523809523809e-05,
375
+ "loss": 2.8116,
376
+ "theoretical_loss": 3.485773185924627,
377
+ "tokens_seen": 1656291328
378
+ },
379
+ {
380
+ "epoch": 0.0,
381
+ "learning_rate": 3.888888888888889e-05,
382
+ "loss": 2.6689,
383
+ "theoretical_loss": 3.4857493256381473,
384
+ "tokens_seen": 1656422400
385
+ },
386
+ {
387
+ "epoch": 0.0,
388
+ "objective/train/docs_used": 914561,
389
+ "objective/train/instantaneous_batch_size": 16,
390
+ "objective/train/instantaneous_microbatch_size": 16384,
391
+ "objective/train/original_loss": 2.774996280670166,
392
+ "objective/train/theoretical_loss": 3.4857254677682548,
393
+ "objective/train/tokens_used": 27012576,
394
+ "theoretical_loss": 3.4857254677682548,
395
+ "tokens_seen": 1656553472
396
+ },
397
+ {
398
+ "epoch": 0.0,
399
+ "learning_rate": 3.968253968253968e-05,
400
+ "loss": 2.6872,
401
+ "theoretical_loss": 3.4857254677682548,
402
+ "tokens_seen": 1656553472
403
+ },
404
+ {
405
+ "epoch": 0.0,
406
+ "learning_rate": 4.047619047619048e-05,
407
+ "loss": 2.5722,
408
+ "theoretical_loss": 3.485701612314514,
409
+ "tokens_seen": 1656684544
410
+ },
411
+ {
412
+ "epoch": 0.0,
413
+ "learning_rate": 4.126984126984127e-05,
414
+ "loss": 2.6782,
415
+ "theoretical_loss": 3.4856777592764883,
416
+ "tokens_seen": 1656815616
417
+ },
418
+ {
419
+ "epoch": 0.0,
420
+ "learning_rate": 4.2063492063492065e-05,
421
+ "loss": 2.6618,
422
+ "theoretical_loss": 3.4856539086537426,
423
+ "tokens_seen": 1656946688
424
+ },
425
+ {
426
+ "epoch": 0.0,
427
+ "learning_rate": 4.2857142857142856e-05,
428
+ "loss": 2.6818,
429
+ "theoretical_loss": 3.485630060445841,
430
+ "tokens_seen": 1657077760
431
+ },
432
+ {
433
+ "epoch": 0.0,
434
+ "learning_rate": 4.3650793650793655e-05,
435
+ "loss": 2.4444,
436
+ "theoretical_loss": 3.485606214652347,
437
+ "tokens_seen": 1657208832
438
+ },
439
+ {
440
+ "epoch": 0.0,
441
+ "learning_rate": 4.4444444444444447e-05,
442
+ "loss": 2.5045,
443
+ "theoretical_loss": 3.4855823712728276,
444
+ "tokens_seen": 1657339904
445
+ },
446
+ {
447
+ "epoch": 0.0,
448
+ "learning_rate": 4.523809523809524e-05,
449
+ "loss": 2.5876,
450
+ "theoretical_loss": 3.4855585303068453,
451
+ "tokens_seen": 1657470976
452
+ },
453
+ {
454
+ "epoch": 0.0,
455
+ "learning_rate": 4.603174603174603e-05,
456
+ "loss": 2.5061,
457
+ "theoretical_loss": 3.4855346917539665,
458
+ "tokens_seen": 1657602048
459
+ },
460
+ {
461
+ "epoch": 0.0,
462
+ "learning_rate": 4.682539682539683e-05,
463
+ "loss": 2.6817,
464
+ "theoretical_loss": 3.4855108556137546,
465
+ "tokens_seen": 1657733120
466
+ },
467
+ {
468
+ "epoch": 0.0,
469
+ "learning_rate": 4.761904761904762e-05,
470
+ "loss": 2.6159,
471
+ "theoretical_loss": 3.4854870218857763,
472
+ "tokens_seen": 1657864192
473
+ },
474
+ {
475
+ "epoch": 0.0,
476
+ "learning_rate": 4.841269841269841e-05,
477
+ "loss": 2.5539,
478
+ "theoretical_loss": 3.485463190569596,
479
+ "tokens_seen": 1657995264
480
+ },
481
+ {
482
+ "epoch": 0.0,
483
+ "learning_rate": 4.9206349206349204e-05,
484
+ "loss": 2.6999,
485
+ "theoretical_loss": 3.485439361664779,
486
+ "tokens_seen": 1658126336
487
+ },
488
+ {
489
+ "epoch": 0.0,
490
+ "objective/train/docs_used": 915678,
491
+ "objective/train/instantaneous_batch_size": 16,
492
+ "objective/train/instantaneous_microbatch_size": 16384,
493
+ "objective/train/original_loss": 2.6932287216186523,
494
+ "objective/train/theoretical_loss": 3.4854274481164964,
495
+ "objective/train/tokens_used": 28650976,
496
+ "theoretical_loss": 3.4854274481164964,
497
+ "tokens_seen": 1658191872
498
+ },
499
+ {
500
+ "epoch": 0.01,
501
+ "learning_rate": 5e-05,
502
+ "loss": 2.5566,
503
+ "theoretical_loss": 3.4854155351708913,
504
+ "tokens_seen": 1658257408
505
+ },
506
+ {
507
+ "epoch": 0.01,
508
+ "learning_rate": 5.0793650793650794e-05,
509
+ "loss": 2.7198,
510
+ "theoretical_loss": 3.4853917110874972,
511
+ "tokens_seen": 1658388480
512
+ },
513
+ {
514
+ "epoch": 0.01,
515
+ "learning_rate": 5.158730158730159e-05,
516
+ "loss": 2.4993,
517
+ "theoretical_loss": 3.4853678894141638,
518
+ "tokens_seen": 1658519552
519
+ },
520
+ {
521
+ "epoch": 0.01,
522
+ "learning_rate": 5.2380952380952384e-05,
523
+ "loss": 2.5541,
524
+ "theoretical_loss": 3.4853440701504557,
525
+ "tokens_seen": 1658650624
526
+ },
527
+ {
528
+ "epoch": 0.01,
529
+ "learning_rate": 5.3174603174603176e-05,
530
+ "loss": 2.5189,
531
+ "theoretical_loss": 3.4853202532959395,
532
+ "tokens_seen": 1658781696
533
+ },
534
+ {
535
+ "epoch": 0.01,
536
+ "learning_rate": 5.396825396825397e-05,
537
+ "loss": 2.322,
538
+ "theoretical_loss": 3.4852964388501806,
539
+ "tokens_seen": 1658912768
540
+ },
541
+ {
542
+ "epoch": 0.01,
543
+ "learning_rate": 5.4761904761904766e-05,
544
+ "loss": 2.6172,
545
+ "theoretical_loss": 3.485272626812746,
546
+ "tokens_seen": 1659043840
547
+ },
548
+ {
549
+ "epoch": 0.01,
550
+ "learning_rate": 5.555555555555556e-05,
551
+ "loss": 2.5944,
552
+ "theoretical_loss": 3.4852488171832006,
553
+ "tokens_seen": 1659174912
554
+ },
555
+ {
556
+ "epoch": 0.01,
557
+ "learning_rate": 5.634920634920635e-05,
558
+ "loss": 2.5423,
559
+ "theoretical_loss": 3.4852250099611117,
560
+ "tokens_seen": 1659305984
561
+ },
562
+ {
563
+ "epoch": 0.01,
564
+ "learning_rate": 5.714285714285714e-05,
565
+ "loss": 2.5272,
566
+ "theoretical_loss": 3.485201205146046,
567
+ "tokens_seen": 1659437056
568
+ },
569
+ {
570
+ "epoch": 0.01,
571
+ "learning_rate": 5.793650793650795e-05,
572
+ "loss": 2.6078,
573
+ "theoretical_loss": 3.4851774027375693,
574
+ "tokens_seen": 1659568128
575
+ },
576
+ {
577
+ "epoch": 0.01,
578
+ "learning_rate": 5.873015873015873e-05,
579
+ "loss": 2.5119,
580
+ "theoretical_loss": 3.4851536027352483,
581
+ "tokens_seen": 1659699200
582
+ },
583
+ {
584
+ "epoch": 0.01,
585
+ "objective/train/docs_used": 916417,
586
+ "objective/train/instantaneous_batch_size": 16,
587
+ "objective/train/instantaneous_microbatch_size": 16384,
588
+ "objective/train/original_loss": 2.456707239151001,
589
+ "objective/train/theoretical_loss": 3.48512980513865,
590
+ "objective/train/tokens_used": 30289376,
591
+ "theoretical_loss": 3.48512980513865,
592
+ "tokens_seen": 1659830272
593
+ },
594
+ {
595
+ "epoch": 0.01,
596
+ "learning_rate": 5.9523809523809524e-05,
597
+ "loss": 2.6637,
598
+ "theoretical_loss": 3.48512980513865,
599
+ "tokens_seen": 1659830272
600
+ },
601
+ {
602
+ "epoch": 0.01,
603
+ "learning_rate": 6.0317460317460316e-05,
604
+ "loss": 2.3894,
605
+ "theoretical_loss": 3.4851060099473417,
606
+ "tokens_seen": 1659961344
607
+ },
608
+ {
609
+ "epoch": 0.01,
610
+ "learning_rate": 6.111111111111112e-05,
611
+ "loss": 2.5928,
612
+ "theoretical_loss": 3.4850822171608895,
613
+ "tokens_seen": 1660092416
614
+ },
615
+ {
616
+ "epoch": 0.01,
617
+ "learning_rate": 6.19047619047619e-05,
618
+ "loss": 2.4428,
619
+ "theoretical_loss": 3.4850584267788607,
620
+ "tokens_seen": 1660223488
621
+ },
622
+ {
623
+ "epoch": 0.01,
624
+ "learning_rate": 6.26984126984127e-05,
625
+ "loss": 2.7376,
626
+ "theoretical_loss": 3.4850346388008235,
627
+ "tokens_seen": 1660354560
628
+ },
629
+ {
630
+ "epoch": 0.01,
631
+ "learning_rate": 6.349206349206349e-05,
632
+ "loss": 2.4582,
633
+ "theoretical_loss": 3.4850108532263437,
634
+ "tokens_seen": 1660485632
635
+ },
636
+ {
637
+ "epoch": 0.01,
638
+ "learning_rate": 6.428571428571429e-05,
639
+ "loss": 2.409,
640
+ "theoretical_loss": 3.48498707005499,
641
+ "tokens_seen": 1660616704
642
+ },
643
+ {
644
+ "epoch": 0.01,
645
+ "learning_rate": 6.507936507936509e-05,
646
+ "loss": 2.5177,
647
+ "theoretical_loss": 3.4849632892863296,
648
+ "tokens_seen": 1660747776
649
+ },
650
+ {
651
+ "epoch": 0.01,
652
+ "learning_rate": 6.587301587301587e-05,
653
+ "loss": 2.6281,
654
+ "theoretical_loss": 3.48493951091993,
655
+ "tokens_seen": 1660878848
656
+ },
657
+ {
658
+ "epoch": 0.01,
659
+ "learning_rate": 6.666666666666667e-05,
660
+ "loss": 2.5806,
661
+ "theoretical_loss": 3.4849157349553592,
662
+ "tokens_seen": 1661009920
663
+ },
664
+ {
665
+ "epoch": 0.01,
666
+ "learning_rate": 6.746031746031747e-05,
667
+ "loss": 2.5608,
668
+ "theoretical_loss": 3.4848919613921847,
669
+ "tokens_seen": 1661140992
670
+ },
671
+ {
672
+ "epoch": 0.01,
673
+ "learning_rate": 6.825396825396825e-05,
674
+ "loss": 2.5385,
675
+ "theoretical_loss": 3.484868190229975,
676
+ "tokens_seen": 1661272064
677
+ },
678
+ {
679
+ "epoch": 0.01,
680
+ "learning_rate": 6.904761904761905e-05,
681
+ "loss": 2.5632,
682
+ "theoretical_loss": 3.484844421468298,
683
+ "tokens_seen": 1661403136
684
+ },
685
+ {
686
+ "epoch": 0.01,
687
+ "objective/train/docs_used": 917689,
688
+ "objective/train/instantaneous_batch_size": 16,
689
+ "objective/train/instantaneous_microbatch_size": 16384,
690
+ "objective/train/original_loss": 2.0480103492736816,
691
+ "objective/train/theoretical_loss": 3.484832537987524,
692
+ "objective/train/tokens_used": 31927776,
693
+ "theoretical_loss": 3.484832537987524,
694
+ "tokens_seen": 1661468672
695
+ },
696
+ {
697
+ "epoch": 0.01,
698
+ "learning_rate": 6.984126984126984e-05,
699
+ "loss": 2.3842,
700
+ "theoretical_loss": 3.4848206551067213,
701
+ "tokens_seen": 1661534208
702
+ },
703
+ {
704
+ "epoch": 0.01,
705
+ "learning_rate": 7.063492063492065e-05,
706
+ "loss": 2.4595,
707
+ "theoretical_loss": 3.484796891144814,
708
+ "tokens_seen": 1661665280
709
+ },
710
+ {
711
+ "epoch": 0.01,
712
+ "learning_rate": 7.142857142857143e-05,
713
+ "loss": 2.5882,
714
+ "theoretical_loss": 3.4847731295821447,
715
+ "tokens_seen": 1661796352
716
+ },
717
+ {
718
+ "epoch": 0.01,
719
+ "learning_rate": 7.222222222222222e-05,
720
+ "loss": 2.3931,
721
+ "theoretical_loss": 3.4847493704182817,
722
+ "tokens_seen": 1661927424
723
+ },
724
+ {
725
+ "epoch": 0.01,
726
+ "learning_rate": 7.301587301587302e-05,
727
+ "loss": 2.622,
728
+ "theoretical_loss": 3.4847256136527935,
729
+ "tokens_seen": 1662058496
730
+ },
731
+ {
732
+ "epoch": 0.01,
733
+ "learning_rate": 7.380952380952382e-05,
734
+ "loss": 2.5425,
735
+ "theoretical_loss": 3.4847018592852494,
736
+ "tokens_seen": 1662189568
737
+ },
738
+ {
739
+ "epoch": 0.01,
740
+ "learning_rate": 7.460317460317461e-05,
741
+ "loss": 2.601,
742
+ "theoretical_loss": 3.4846781073152173,
743
+ "tokens_seen": 1662320640
744
+ },
745
+ {
746
+ "epoch": 0.01,
747
+ "learning_rate": 7.53968253968254e-05,
748
+ "loss": 2.5712,
749
+ "theoretical_loss": 3.484654357742267,
750
+ "tokens_seen": 1662451712
751
+ },
752
+ {
753
+ "epoch": 0.01,
754
+ "learning_rate": 7.619047619047618e-05,
755
+ "loss": 2.2686,
756
+ "theoretical_loss": 3.4846306105659677,
757
+ "tokens_seen": 1662582784
758
+ },
759
+ {
760
+ "epoch": 0.01,
761
+ "learning_rate": 7.6984126984127e-05,
762
+ "loss": 2.3971,
763
+ "theoretical_loss": 3.4846068657858877,
764
+ "tokens_seen": 1662713856
765
+ },
766
+ {
767
+ "epoch": 0.01,
768
+ "learning_rate": 7.777777777777778e-05,
769
+ "loss": 2.5189,
770
+ "theoretical_loss": 3.4845831234015976,
771
+ "tokens_seen": 1662844928
772
+ },
773
+ {
774
+ "epoch": 0.01,
775
+ "learning_rate": 7.857142857142858e-05,
776
+ "loss": 2.3924,
777
+ "theoretical_loss": 3.484559383412666,
778
+ "tokens_seen": 1662976000
779
+ },
780
+ {
781
+ "epoch": 0.01,
782
+ "objective/train/docs_used": 918239,
783
+ "objective/train/instantaneous_batch_size": 16,
784
+ "objective/train/instantaneous_microbatch_size": 16384,
785
+ "objective/train/original_loss": 2.3232572078704834,
786
+ "objective/train/theoretical_loss": 3.4845356458186627,
787
+ "objective/train/tokens_used": 33566176,
788
+ "theoretical_loss": 3.4845356458186627,
789
+ "tokens_seen": 1663107072
790
+ },
791
+ {
792
+ "epoch": 0.01,
793
+ "learning_rate": 7.936507936507937e-05,
794
+ "loss": 2.5743,
795
+ "theoretical_loss": 3.4845356458186627,
796
+ "tokens_seen": 1663107072
797
+ },
798
+ {
799
+ "epoch": 0.01,
800
+ "learning_rate": 8.015873015873016e-05,
801
+ "loss": 2.4116,
802
+ "theoretical_loss": 3.4845119106191578,
803
+ "tokens_seen": 1663238144
804
+ },
805
+ {
806
+ "epoch": 0.01,
807
+ "learning_rate": 8.095238095238096e-05,
808
+ "loss": 2.5243,
809
+ "theoretical_loss": 3.48448817781372,
810
+ "tokens_seen": 1663369216
811
+ },
812
+ {
813
+ "epoch": 0.01,
814
+ "learning_rate": 8.174603174603175e-05,
815
+ "loss": 2.4798,
816
+ "theoretical_loss": 3.484464447401921,
817
+ "tokens_seen": 1663500288
818
+ },
819
+ {
820
+ "epoch": 0.01,
821
+ "learning_rate": 8.253968253968255e-05,
822
+ "loss": 2.3456,
823
+ "theoretical_loss": 3.4844407193833282,
824
+ "tokens_seen": 1663631360
825
+ },
826
+ {
827
+ "epoch": 0.01,
828
+ "learning_rate": 8.333333333333334e-05,
829
+ "loss": 2.4678,
830
+ "theoretical_loss": 3.484416993757514,
831
+ "tokens_seen": 1663762432
832
+ },
833
+ {
834
+ "epoch": 0.01,
835
+ "learning_rate": 8.412698412698413e-05,
836
+ "loss": 2.4106,
837
+ "theoretical_loss": 3.4843932705240483,
838
+ "tokens_seen": 1663893504
839
+ },
840
+ {
841
+ "epoch": 0.01,
842
+ "learning_rate": 8.492063492063493e-05,
843
+ "loss": 2.3669,
844
+ "theoretical_loss": 3.4843695496825005,
845
+ "tokens_seen": 1664024576
846
+ },
847
+ {
848
+ "epoch": 0.01,
849
+ "learning_rate": 8.571428571428571e-05,
850
+ "loss": 2.5092,
851
+ "theoretical_loss": 3.4843458312324413,
852
+ "tokens_seen": 1664155648
853
+ },
854
+ {
855
+ "epoch": 0.01,
856
+ "learning_rate": 8.650793650793651e-05,
857
+ "loss": 2.473,
858
+ "theoretical_loss": 3.484322115173442,
859
+ "tokens_seen": 1664286720
860
+ },
861
+ {
862
+ "epoch": 0.01,
863
+ "learning_rate": 8.730158730158731e-05,
864
+ "loss": 2.5359,
865
+ "theoretical_loss": 3.4842984015050726,
866
+ "tokens_seen": 1664417792
867
+ },
868
+ {
869
+ "epoch": 0.01,
870
+ "learning_rate": 8.80952380952381e-05,
871
+ "loss": 2.6096,
872
+ "theoretical_loss": 3.484274690226904,
873
+ "tokens_seen": 1664548864
874
+ },
875
+ {
876
+ "epoch": 0.01,
877
+ "learning_rate": 8.888888888888889e-05,
878
+ "loss": 2.4428,
879
+ "theoretical_loss": 3.484250981338507,
880
+ "tokens_seen": 1664679936
881
+ },
882
+ {
883
+ "epoch": 0.01,
884
+ "objective/train/docs_used": 919306,
885
+ "objective/train/instantaneous_batch_size": 16,
886
+ "objective/train/instantaneous_microbatch_size": 16384,
887
+ "objective/train/original_loss": 2.603062629699707,
888
+ "objective/train/theoretical_loss": 3.484239127790339,
889
+ "objective/train/tokens_used": 35204576,
890
+ "theoretical_loss": 3.484239127790339,
891
+ "tokens_seen": 1664745472
892
+ },
893
+ {
894
+ "epoch": 0.01,
895
+ "learning_rate": 8.968253968253969e-05,
896
+ "loss": 2.5448,
897
+ "theoretical_loss": 3.484227274839453,
898
+ "tokens_seen": 1664811008
899
+ },
900
+ {
901
+ "epoch": 0.01,
902
+ "learning_rate": 9.047619047619048e-05,
903
+ "loss": 2.5698,
904
+ "theoretical_loss": 3.484203570729313,
905
+ "tokens_seen": 1664942080
906
+ },
907
+ {
908
+ "epoch": 0.01,
909
+ "learning_rate": 9.126984126984128e-05,
910
+ "loss": 2.4996,
911
+ "theoretical_loss": 3.484179869007658,
912
+ "tokens_seen": 1665073152
913
+ },
914
+ {
915
+ "epoch": 0.01,
916
+ "learning_rate": 9.206349206349206e-05,
917
+ "loss": 2.5152,
918
+ "theoretical_loss": 3.4841561696740597,
919
+ "tokens_seen": 1665204224
920
+ },
921
+ {
922
+ "epoch": 0.01,
923
+ "learning_rate": 9.285714285714286e-05,
924
+ "loss": 2.4996,
925
+ "theoretical_loss": 3.4841324727280893,
926
+ "tokens_seen": 1665335296
927
+ },
928
+ {
929
+ "epoch": 0.01,
930
+ "learning_rate": 9.365079365079366e-05,
931
+ "loss": 2.47,
932
+ "theoretical_loss": 3.484108778169318,
933
+ "tokens_seen": 1665466368
934
+ },
935
+ {
936
+ "epoch": 0.01,
937
+ "learning_rate": 9.444444444444444e-05,
938
+ "loss": 2.3922,
939
+ "theoretical_loss": 3.484085085997318,
940
+ "tokens_seen": 1665597440
941
+ },
942
+ {
943
+ "epoch": 0.01,
944
+ "learning_rate": 9.523809523809524e-05,
945
+ "loss": 2.4767,
946
+ "theoretical_loss": 3.484061396211661,
947
+ "tokens_seen": 1665728512
948
+ },
949
+ {
950
+ "epoch": 0.01,
951
+ "learning_rate": 9.603174603174604e-05,
952
+ "loss": 2.5377,
953
+ "theoretical_loss": 3.4840377088119188,
954
+ "tokens_seen": 1665859584
955
+ },
956
+ {
957
+ "epoch": 0.01,
958
+ "learning_rate": 9.682539682539682e-05,
959
+ "loss": 2.61,
960
+ "theoretical_loss": 3.4840140237976636,
961
+ "tokens_seen": 1665990656
962
+ },
963
+ {
964
+ "epoch": 0.01,
965
+ "learning_rate": 9.761904761904762e-05,
966
+ "loss": 2.4506,
967
+ "theoretical_loss": 3.483990341168467,
968
+ "tokens_seen": 1666121728
969
+ },
970
+ {
971
+ "epoch": 0.01,
972
+ "learning_rate": 9.841269841269841e-05,
973
+ "loss": 2.4311,
974
+ "theoretical_loss": 3.483966660923902,
975
+ "tokens_seen": 1666252800
976
+ },
977
+ {
978
+ "epoch": 0.01,
979
+ "objective/train/docs_used": 919806,
980
+ "objective/train/instantaneous_batch_size": 16,
981
+ "objective/train/instantaneous_microbatch_size": 16384,
982
+ "objective/train/original_loss": 2.3898022174835205,
983
+ "objective/train/theoretical_loss": 3.4839429830635407,
984
+ "objective/train/tokens_used": 36842976,
985
+ "theoretical_loss": 3.4839429830635407,
986
+ "tokens_seen": 1666383872
987
+ },
988
+ {
989
+ "epoch": 0.01,
990
+ "learning_rate": 9.920634920634922e-05,
991
+ "loss": 2.543,
992
+ "theoretical_loss": 3.4839429830635407,
993
+ "tokens_seen": 1666383872
994
+ },
995
+ {
996
+ "epoch": 0.01,
997
+ "learning_rate": 0.0001,
998
+ "loss": 2.4395,
999
+ "theoretical_loss": 3.4839193075869543,
1000
+ "tokens_seen": 1666514944
1001
+ },
1002
+ {
1003
+ "epoch": 0.01,
1004
+ "learning_rate": 9.999197560584176e-05,
1005
+ "loss": 2.4916,
1006
+ "theoretical_loss": 3.4838956344937175,
1007
+ "tokens_seen": 1666646016
1008
+ },
1009
+ {
1010
+ "epoch": 0.01,
1011
+ "learning_rate": 9.998395121168352e-05,
1012
+ "loss": 2.3888,
1013
+ "theoretical_loss": 3.4838719637834012,
1014
+ "tokens_seen": 1666777088
1015
+ },
1016
+ {
1017
+ "epoch": 0.01,
1018
+ "learning_rate": 9.997592681752528e-05,
1019
+ "loss": 2.5063,
1020
+ "theoretical_loss": 3.483848295455579,
1021
+ "tokens_seen": 1666908160
1022
+ },
1023
+ {
1024
+ "epoch": 0.01,
1025
+ "learning_rate": 9.996790242336704e-05,
1026
+ "loss": 2.4951,
1027
+ "theoretical_loss": 3.483824629509824,
1028
+ "tokens_seen": 1667039232
1029
+ },
1030
+ {
1031
+ "epoch": 0.01,
1032
+ "learning_rate": 9.99598780292088e-05,
1033
+ "loss": 2.4226,
1034
+ "theoretical_loss": 3.4838009659457088,
1035
+ "tokens_seen": 1667170304
1036
+ },
1037
+ {
1038
+ "epoch": 0.01,
1039
+ "learning_rate": 9.995185363505056e-05,
1040
+ "loss": 2.4431,
1041
+ "theoretical_loss": 3.4837773047628064,
1042
+ "tokens_seen": 1667301376
1043
+ },
1044
+ {
1045
+ "epoch": 0.01,
1046
+ "learning_rate": 9.994382924089232e-05,
1047
+ "loss": 2.4434,
1048
+ "theoretical_loss": 3.4837536459606904,
1049
+ "tokens_seen": 1667432448
1050
+ },
1051
+ {
1052
+ "epoch": 0.01,
1053
+ "learning_rate": 9.993580484673407e-05,
1054
+ "loss": 2.5021,
1055
+ "theoretical_loss": 3.4837299895389333,
1056
+ "tokens_seen": 1667563520
1057
+ },
1058
+ {
1059
+ "epoch": 0.01,
1060
+ "learning_rate": 9.992778045257584e-05,
1061
+ "loss": 2.5622,
1062
+ "theoretical_loss": 3.48370633549711,
1063
+ "tokens_seen": 1667694592
1064
+ },
1065
+ {
1066
+ "epoch": 0.01,
1067
+ "learning_rate": 9.991975605841759e-05,
1068
+ "loss": 2.524,
1069
+ "theoretical_loss": 3.4836826838347923,
1070
+ "tokens_seen": 1667825664
1071
+ },
1072
+ {
1073
+ "epoch": 0.01,
1074
+ "learning_rate": 9.991173166425936e-05,
1075
+ "loss": 2.5055,
1076
+ "theoretical_loss": 3.483659034551555,
1077
+ "tokens_seen": 1667956736
1078
+ },
1079
+ {
1080
+ "epoch": 0.01,
1081
+ "objective/train/docs_used": 921005,
1082
+ "objective/train/instantaneous_batch_size": 16,
1083
+ "objective/train/instantaneous_microbatch_size": 16384,
1084
+ "objective/train/original_loss": 2.6571173667907715,
1085
+ "objective/train/theoretical_loss": 3.4836472108019585,
1086
+ "objective/train/tokens_used": 38481376,
1087
+ "theoretical_loss": 3.4836472108019585,
1088
+ "tokens_seen": 1668022272
1089
+ },
1090
+ {
1091
+ "epoch": 0.01,
1092
+ "learning_rate": 9.990370727010111e-05,
1093
+ "loss": 2.5956,
1094
+ "theoretical_loss": 3.483635387646972,
1095
+ "tokens_seen": 1668087808
1096
+ },
1097
+ {
1098
+ "epoch": 0.01,
1099
+ "learning_rate": 9.989568287594288e-05,
1100
+ "loss": 2.5279,
1101
+ "theoretical_loss": 3.4836117431206164,
1102
+ "tokens_seen": 1668218880
1103
+ },
1104
+ {
1105
+ "epoch": 0.01,
1106
+ "learning_rate": 9.988765848178463e-05,
1107
+ "loss": 2.45,
1108
+ "theoretical_loss": 3.483588100972063,
1109
+ "tokens_seen": 1668349952
1110
+ },
1111
+ {
1112
+ "epoch": 0.01,
1113
+ "learning_rate": 9.987963408762638e-05,
1114
+ "loss": 2.5056,
1115
+ "theoretical_loss": 3.483564461200885,
1116
+ "tokens_seen": 1668481024
1117
+ },
1118
+ {
1119
+ "epoch": 0.01,
1120
+ "learning_rate": 9.987160969346815e-05,
1121
+ "loss": 2.56,
1122
+ "theoretical_loss": 3.4835408238066567,
1123
+ "tokens_seen": 1668612096
1124
+ },
1125
+ {
1126
+ "epoch": 0.01,
1127
+ "learning_rate": 9.98635852993099e-05,
1128
+ "loss": 2.4673,
1129
+ "theoretical_loss": 3.4835171887889533,
1130
+ "tokens_seen": 1668743168
1131
+ },
1132
+ {
1133
+ "epoch": 0.01,
1134
+ "learning_rate": 9.985556090515167e-05,
1135
+ "loss": 2.3887,
1136
+ "theoretical_loss": 3.483493556147349,
1137
+ "tokens_seen": 1668874240
1138
+ },
1139
+ {
1140
+ "epoch": 0.01,
1141
+ "learning_rate": 9.984753651099342e-05,
1142
+ "loss": 2.478,
1143
+ "theoretical_loss": 3.4834699258814172,
1144
+ "tokens_seen": 1669005312
1145
+ },
1146
+ {
1147
+ "epoch": 0.01,
1148
+ "learning_rate": 9.983951211683519e-05,
1149
+ "loss": 2.6136,
1150
+ "theoretical_loss": 3.483446297990734,
1151
+ "tokens_seen": 1669136384
1152
+ },
1153
+ {
1154
+ "epoch": 0.01,
1155
+ "learning_rate": 9.983148772267694e-05,
1156
+ "loss": 2.5876,
1157
+ "theoretical_loss": 3.4834226724748736,
1158
+ "tokens_seen": 1669267456
1159
+ },
1160
+ {
1161
+ "epoch": 0.01,
1162
+ "learning_rate": 9.982346332851871e-05,
1163
+ "loss": 2.5149,
1164
+ "theoretical_loss": 3.4833990493334106,
1165
+ "tokens_seen": 1669398528
1166
+ },
1167
+ {
1168
+ "epoch": 0.01,
1169
+ "learning_rate": 9.981543893436046e-05,
1170
+ "loss": 2.5755,
1171
+ "theoretical_loss": 3.48337542856592,
1172
+ "tokens_seen": 1669529600
1173
+ },
1174
+ {
1175
+ "epoch": 0.01,
1176
+ "objective/train/docs_used": 921680,
1177
+ "objective/train/instantaneous_batch_size": 16,
1178
+ "objective/train/instantaneous_microbatch_size": 16384,
1179
+ "objective/train/original_loss": 2.329754590988159,
1180
+ "objective/train/theoretical_loss": 3.4833518101719774,
1181
+ "objective/train/tokens_used": 40119776,
1182
+ "theoretical_loss": 3.4833518101719774,
1183
+ "tokens_seen": 1669660672
1184
+ },
1185
+ {
1186
+ "epoch": 0.01,
1187
+ "learning_rate": 9.980741454020222e-05,
1188
+ "loss": 2.4671,
1189
+ "theoretical_loss": 3.4833518101719774,
1190
+ "tokens_seen": 1669660672
1191
+ },
1192
+ {
1193
+ "epoch": 0.01,
1194
+ "learning_rate": 9.979939014604398e-05,
1195
+ "loss": 2.3942,
1196
+ "theoretical_loss": 3.4833281941511576,
1197
+ "tokens_seen": 1669791744
1198
+ },
1199
+ {
1200
+ "epoch": 0.01,
1201
+ "learning_rate": 9.979136575188574e-05,
1202
+ "loss": 2.4857,
1203
+ "theoretical_loss": 3.4833045805030363,
1204
+ "tokens_seen": 1669922816
1205
+ },
1206
+ {
1207
+ "epoch": 0.01,
1208
+ "learning_rate": 9.97833413577275e-05,
1209
+ "loss": 2.4502,
1210
+ "theoretical_loss": 3.483280969227188,
1211
+ "tokens_seen": 1670053888
1212
+ },
1213
+ {
1214
+ "epoch": 0.01,
1215
+ "learning_rate": 9.977531696356926e-05,
1216
+ "loss": 2.5831,
1217
+ "theoretical_loss": 3.4832573603231887,
1218
+ "tokens_seen": 1670184960
1219
+ },
1220
+ {
1221
+ "epoch": 0.01,
1222
+ "learning_rate": 9.976729256941102e-05,
1223
+ "loss": 2.561,
1224
+ "theoretical_loss": 3.4832337537906146,
1225
+ "tokens_seen": 1670316032
1226
+ },
1227
+ {
1228
+ "epoch": 0.01,
1229
+ "learning_rate": 9.975926817525277e-05,
1230
+ "loss": 2.5869,
1231
+ "theoretical_loss": 3.4832101496290404,
1232
+ "tokens_seen": 1670447104
1233
+ },
1234
+ {
1235
+ "epoch": 0.01,
1236
+ "learning_rate": 9.975124378109453e-05,
1237
+ "loss": 2.4206,
1238
+ "theoretical_loss": 3.4831865478380433,
1239
+ "tokens_seen": 1670578176
1240
+ },
1241
+ {
1242
+ "epoch": 0.01,
1243
+ "learning_rate": 9.97432193869363e-05,
1244
+ "loss": 2.3991,
1245
+ "theoretical_loss": 3.4831629484171978,
1246
+ "tokens_seen": 1670709248
1247
+ },
1248
+ {
1249
+ "epoch": 0.01,
1250
+ "learning_rate": 9.973519499277805e-05,
1251
+ "loss": 2.3821,
1252
+ "theoretical_loss": 3.4831393513660807,
1253
+ "tokens_seen": 1670840320
1254
+ },
1255
+ {
1256
+ "epoch": 0.01,
1257
+ "learning_rate": 9.972717059861981e-05,
1258
+ "loss": 2.5258,
1259
+ "theoretical_loss": 3.483115756684268,
1260
+ "tokens_seen": 1670971392
1261
+ },
1262
+ {
1263
+ "epoch": 0.01,
1264
+ "learning_rate": 9.971914620446157e-05,
1265
+ "loss": 2.4604,
1266
+ "theoretical_loss": 3.4830921643713366,
1267
+ "tokens_seen": 1671102464
1268
+ },
1269
+ {
1270
+ "epoch": 0.01,
1271
+ "learning_rate": 9.971112181030333e-05,
1272
+ "loss": 2.5027,
1273
+ "theoretical_loss": 3.483068574426862,
1274
+ "tokens_seen": 1671233536
1275
+ },
1276
+ {
1277
+ "epoch": 0.01,
1278
+ "objective/train/docs_used": 922598,
1279
+ "objective/train/instantaneous_batch_size": 16,
1280
+ "objective/train/instantaneous_microbatch_size": 16384,
1281
+ "objective/train/original_loss": 2.674171209335327,
1282
+ "objective/train/theoretical_loss": 3.4830567803426637,
1283
+ "objective/train/tokens_used": 41758176,
1284
+ "theoretical_loss": 3.4830567803426637,
1285
+ "tokens_seen": 1671299072
1286
+ },
1287
+ {
1288
+ "epoch": 0.01,
1289
+ "learning_rate": 9.970309741614509e-05,
1290
+ "loss": 2.5194,
1291
+ "theoretical_loss": 3.483044986850421,
1292
+ "tokens_seen": 1671364608
1293
+ },
1294
+ {
1295
+ "epoch": 0.01,
1296
+ "learning_rate": 9.969507302198684e-05,
1297
+ "loss": 2.4984,
1298
+ "theoretical_loss": 3.48302140164159,
1299
+ "tokens_seen": 1671495680
1300
+ },
1301
+ {
1302
+ "epoch": 0.01,
1303
+ "learning_rate": 9.968704862782861e-05,
1304
+ "loss": 2.6113,
1305
+ "theoretical_loss": 3.482997818799947,
1306
+ "tokens_seen": 1671626752
1307
+ },
1308
+ {
1309
+ "epoch": 0.01,
1310
+ "learning_rate": 9.967902423367036e-05,
1311
+ "loss": 2.6013,
1312
+ "theoretical_loss": 3.4829742383250673,
1313
+ "tokens_seen": 1671757824
1314
+ },
1315
+ {
1316
+ "epoch": 0.01,
1317
+ "learning_rate": 9.967099983951213e-05,
1318
+ "loss": 2.4007,
1319
+ "theoretical_loss": 3.4829506602165283,
1320
+ "tokens_seen": 1671888896
1321
+ },
1322
+ {
1323
+ "epoch": 0.01,
1324
+ "learning_rate": 9.966297544535388e-05,
1325
+ "loss": 2.4473,
1326
+ "theoretical_loss": 3.482927084473907,
1327
+ "tokens_seen": 1672019968
1328
+ },
1329
+ {
1330
+ "epoch": 0.01,
1331
+ "learning_rate": 9.965495105119565e-05,
1332
+ "loss": 2.3941,
1333
+ "theoretical_loss": 3.482903511096781,
1334
+ "tokens_seen": 1672151040
1335
+ },
1336
+ {
1337
+ "epoch": 0.01,
1338
+ "learning_rate": 9.96469266570374e-05,
1339
+ "loss": 2.5564,
1340
+ "theoretical_loss": 3.482879940084727,
1341
+ "tokens_seen": 1672282112
1342
+ },
1343
+ {
1344
+ "epoch": 0.01,
1345
+ "learning_rate": 9.963890226287915e-05,
1346
+ "loss": 2.4979,
1347
+ "theoretical_loss": 3.4828563714373226,
1348
+ "tokens_seen": 1672413184
1349
+ },
1350
+ {
1351
+ "epoch": 0.01,
1352
+ "learning_rate": 9.963087786872092e-05,
1353
+ "loss": 2.4506,
1354
+ "theoretical_loss": 3.4828328051541453,
1355
+ "tokens_seen": 1672544256
1356
+ },
1357
+ {
1358
+ "epoch": 0.01,
1359
+ "learning_rate": 9.962285347456267e-05,
1360
+ "loss": 2.5131,
1361
+ "theoretical_loss": 3.482809241234773,
1362
+ "tokens_seen": 1672675328
1363
+ },
1364
+ {
1365
+ "epoch": 0.01,
1366
+ "learning_rate": 9.961482908040444e-05,
1367
+ "loss": 2.4079,
1368
+ "theoretical_loss": 3.482785679678783,
1369
+ "tokens_seen": 1672806400
1370
+ },
1371
+ {
1372
+ "epoch": 0.01,
1373
+ "objective/train/docs_used": 923236,
1374
+ "objective/train/instantaneous_batch_size": 16,
1375
+ "objective/train/instantaneous_microbatch_size": 16384,
1376
+ "objective/train/original_loss": 2.478933095932007,
1377
+ "objective/train/theoretical_loss": 3.482762120485753,
1378
+ "objective/train/tokens_used": 43396576,
1379
+ "theoretical_loss": 3.482762120485753,
1380
+ "tokens_seen": 1672937472
1381
+ },
1382
+ {
1383
+ "epoch": 0.01,
1384
+ "learning_rate": 9.960680468624619e-05,
1385
+ "loss": 2.5005,
1386
+ "theoretical_loss": 3.482762120485753,
1387
+ "tokens_seen": 1672937472
1388
+ },
1389
+ {
1390
+ "epoch": 0.01,
1391
+ "learning_rate": 9.959878029208796e-05,
1392
+ "loss": 2.669,
1393
+ "theoretical_loss": 3.482738563655261,
1394
+ "tokens_seen": 1673068544
1395
+ },
1396
+ {
1397
+ "epoch": 0.01,
1398
+ "learning_rate": 9.959075589792971e-05,
1399
+ "loss": 2.5706,
1400
+ "theoretical_loss": 3.4827150091868853,
1401
+ "tokens_seen": 1673199616
1402
+ },
1403
+ {
1404
+ "epoch": 0.01,
1405
+ "learning_rate": 9.958273150377147e-05,
1406
+ "loss": 2.5295,
1407
+ "theoretical_loss": 3.482691457080204,
1408
+ "tokens_seen": 1673330688
1409
+ },
1410
+ {
1411
+ "epoch": 0.01,
1412
+ "learning_rate": 9.957470710961323e-05,
1413
+ "loss": 2.4358,
1414
+ "theoretical_loss": 3.482667907334795,
1415
+ "tokens_seen": 1673461760
1416
+ },
1417
+ {
1418
+ "epoch": 0.01,
1419
+ "learning_rate": 9.956668271545499e-05,
1420
+ "loss": 2.4467,
1421
+ "theoretical_loss": 3.482644359950237,
1422
+ "tokens_seen": 1673592832
1423
+ },
1424
+ {
1425
+ "epoch": 0.01,
1426
+ "learning_rate": 9.955865832129675e-05,
1427
+ "loss": 2.3988,
1428
+ "theoretical_loss": 3.4826208149261078,
1429
+ "tokens_seen": 1673723904
1430
+ },
1431
+ {
1432
+ "epoch": 0.01,
1433
+ "learning_rate": 9.95506339271385e-05,
1434
+ "loss": 2.521,
1435
+ "theoretical_loss": 3.482597272261987,
1436
+ "tokens_seen": 1673854976
1437
+ },
1438
+ {
1439
+ "epoch": 0.01,
1440
+ "learning_rate": 9.954260953298027e-05,
1441
+ "loss": 2.409,
1442
+ "theoretical_loss": 3.4825737319574523,
1443
+ "tokens_seen": 1673986048
1444
+ },
1445
+ {
1446
+ "epoch": 0.01,
1447
+ "learning_rate": 9.953458513882203e-05,
1448
+ "loss": 2.5548,
1449
+ "theoretical_loss": 3.4825501940120835,
1450
+ "tokens_seen": 1674117120
1451
+ },
1452
+ {
1453
+ "epoch": 0.01,
1454
+ "learning_rate": 9.952656074466378e-05,
1455
+ "loss": 2.4541,
1456
+ "theoretical_loss": 3.4825266584254586,
1457
+ "tokens_seen": 1674248192
1458
+ },
1459
+ {
1460
+ "epoch": 0.01,
1461
+ "learning_rate": 9.951853635050554e-05,
1462
+ "loss": 2.5549,
1463
+ "theoretical_loss": 3.482503125197157,
1464
+ "tokens_seen": 1674379264
1465
+ },
1466
+ {
1467
+ "epoch": 0.01,
1468
+ "learning_rate": 9.95105119563473e-05,
1469
+ "loss": 2.4687,
1470
+ "theoretical_loss": 3.4824795943267577,
1471
+ "tokens_seen": 1674510336
1472
+ },
1473
+ {
1474
+ "epoch": 0.01,
1475
+ "objective/train/docs_used": 924425,
1476
+ "objective/train/instantaneous_batch_size": 16,
1477
+ "objective/train/instantaneous_microbatch_size": 16384,
1478
+ "objective/train/original_loss": 2.032245397567749,
1479
+ "objective/train/theoretical_loss": 3.48246782977564,
1480
+ "objective/train/tokens_used": 45034976,
1481
+ "theoretical_loss": 3.48246782977564,
1482
+ "tokens_seen": 1674575872
1483
+ },
1484
+ {
1485
+ "epoch": 0.01,
1486
+ "learning_rate": 9.950248756218906e-05,
1487
+ "loss": 2.4781,
1488
+ "theoretical_loss": 3.4824560658138397,
1489
+ "tokens_seen": 1674641408
1490
+ },
1491
+ {
1492
+ "epoch": 0.02,
1493
+ "learning_rate": 9.949446316803082e-05,
1494
+ "loss": 2.6007,
1495
+ "theoretical_loss": 3.482432539657983,
1496
+ "tokens_seen": 1674772480
1497
+ },
1498
+ {
1499
+ "epoch": 0.02,
1500
+ "learning_rate": 9.948643877387258e-05,
1501
+ "loss": 2.4198,
1502
+ "theoretical_loss": 3.4824090158587664,
1503
+ "tokens_seen": 1674903552
1504
+ },
1505
+ {
1506
+ "epoch": 0.02,
1507
+ "learning_rate": 9.947841437971434e-05,
1508
+ "loss": 2.49,
1509
+ "theoretical_loss": 3.4823854944157695,
1510
+ "tokens_seen": 1675034624
1511
+ },
1512
+ {
1513
+ "epoch": 0.02,
1514
+ "learning_rate": 9.94703899855561e-05,
1515
+ "loss": 2.4733,
1516
+ "theoretical_loss": 3.482361975328572,
1517
+ "tokens_seen": 1675165696
1518
+ },
1519
+ {
1520
+ "epoch": 0.02,
1521
+ "learning_rate": 9.946236559139786e-05,
1522
+ "loss": 2.6164,
1523
+ "theoretical_loss": 3.482338458596754,
1524
+ "tokens_seen": 1675296768
1525
+ },
1526
+ {
1527
+ "epoch": 0.02,
1528
+ "learning_rate": 9.945434119723961e-05,
1529
+ "loss": 2.4729,
1530
+ "theoretical_loss": 3.482314944219895,
1531
+ "tokens_seen": 1675427840
1532
+ },
1533
+ {
1534
+ "epoch": 0.02,
1535
+ "learning_rate": 9.944631680308138e-05,
1536
+ "loss": 2.5106,
1537
+ "theoretical_loss": 3.482291432197575,
1538
+ "tokens_seen": 1675558912
1539
+ },
1540
+ {
1541
+ "epoch": 0.02,
1542
+ "learning_rate": 9.943829240892313e-05,
1543
+ "loss": 2.4255,
1544
+ "theoretical_loss": 3.482267922529374,
1545
+ "tokens_seen": 1675689984
1546
+ },
1547
+ {
1548
+ "epoch": 0.02,
1549
+ "learning_rate": 9.94302680147649e-05,
1550
+ "loss": 2.4172,
1551
+ "theoretical_loss": 3.482244415214873,
1552
+ "tokens_seen": 1675821056
1553
+ },
1554
+ {
1555
+ "epoch": 0.02,
1556
+ "learning_rate": 9.942224362060665e-05,
1557
+ "loss": 2.4485,
1558
+ "theoretical_loss": 3.482220910253651,
1559
+ "tokens_seen": 1675952128
1560
+ },
1561
+ {
1562
+ "epoch": 0.02,
1563
+ "learning_rate": 9.941421922644842e-05,
1564
+ "loss": 2.5783,
1565
+ "theoretical_loss": 3.482197407645289,
1566
+ "tokens_seen": 1676083200
1567
+ },
1568
+ {
1569
+ "epoch": 0.02,
1570
+ "objective/train/docs_used": 924948,
1571
+ "objective/train/instantaneous_batch_size": 16,
1572
+ "objective/train/instantaneous_microbatch_size": 16384,
1573
+ "objective/train/original_loss": 2.1431517601013184,
1574
+ "objective/train/theoretical_loss": 3.4821739073893676,
1575
+ "objective/train/tokens_used": 46673376,
1576
+ "theoretical_loss": 3.4821739073893676,
1577
+ "tokens_seen": 1676214272
1578
+ },
1579
+ {
1580
+ "epoch": 0.02,
1581
+ "learning_rate": 9.940619483229017e-05,
1582
+ "loss": 2.4842,
1583
+ "theoretical_loss": 3.4821739073893676,
1584
+ "tokens_seen": 1676214272
1585
+ },
1586
+ {
1587
+ "epoch": 0.02,
1588
+ "learning_rate": 9.939817043813192e-05,
1589
+ "loss": 2.5048,
1590
+ "theoretical_loss": 3.482150409485467,
1591
+ "tokens_seen": 1676345344
1592
+ },
1593
+ {
1594
+ "epoch": 0.02,
1595
+ "learning_rate": 9.939014604397369e-05,
1596
+ "loss": 2.4242,
1597
+ "theoretical_loss": 3.482126913933169,
1598
+ "tokens_seen": 1676476416
1599
+ },
1600
+ {
1601
+ "epoch": 0.02,
1602
+ "learning_rate": 9.938212164981544e-05,
1603
+ "loss": 2.4716,
1604
+ "theoretical_loss": 3.482103420732053,
1605
+ "tokens_seen": 1676607488
1606
+ },
1607
+ {
1608
+ "epoch": 0.02,
1609
+ "learning_rate": 9.937409725565721e-05,
1610
+ "loss": 2.4006,
1611
+ "theoretical_loss": 3.482079929881701,
1612
+ "tokens_seen": 1676738560
1613
+ },
1614
+ {
1615
+ "epoch": 0.02,
1616
+ "learning_rate": 9.936607286149896e-05,
1617
+ "loss": 2.5438,
1618
+ "theoretical_loss": 3.482056441381694,
1619
+ "tokens_seen": 1676869632
1620
+ },
1621
+ {
1622
+ "epoch": 0.02,
1623
+ "learning_rate": 9.935804846734073e-05,
1624
+ "loss": 2.4948,
1625
+ "theoretical_loss": 3.4820329552316123,
1626
+ "tokens_seen": 1677000704
1627
+ },
1628
+ {
1629
+ "epoch": 0.02,
1630
+ "learning_rate": 9.935002407318248e-05,
1631
+ "loss": 2.4135,
1632
+ "theoretical_loss": 3.482009471431038,
1633
+ "tokens_seen": 1677131776
1634
+ },
1635
+ {
1636
+ "epoch": 0.02,
1637
+ "learning_rate": 9.934199967902424e-05,
1638
+ "loss": 2.4302,
1639
+ "theoretical_loss": 3.4819859899795516,
1640
+ "tokens_seen": 1677262848
1641
+ },
1642
+ {
1643
+ "epoch": 0.02,
1644
+ "learning_rate": 9.9333975284866e-05,
1645
+ "loss": 2.4168,
1646
+ "theoretical_loss": 3.481962510876736,
1647
+ "tokens_seen": 1677393920
1648
+ },
1649
+ {
1650
+ "epoch": 0.02,
1651
+ "learning_rate": 9.932595089070776e-05,
1652
+ "loss": 2.648,
1653
+ "theoretical_loss": 3.481939034122171,
1654
+ "tokens_seen": 1677524992
1655
+ },
1656
+ {
1657
+ "epoch": 0.02,
1658
+ "learning_rate": 9.931792649654952e-05,
1659
+ "loss": 2.4332,
1660
+ "theoretical_loss": 3.4819155597154396,
1661
+ "tokens_seen": 1677656064
1662
+ },
1663
+ {
1664
+ "epoch": 0.02,
1665
+ "learning_rate": 9.930990210239128e-05,
1666
+ "loss": 2.4625,
1667
+ "theoretical_loss": 3.4818920876561235,
1668
+ "tokens_seen": 1677787136
1669
+ },
1670
+ {
1671
+ "epoch": 0.02,
1672
+ "objective/train/docs_used": 926165,
1673
+ "objective/train/instantaneous_batch_size": 16,
1674
+ "objective/train/instantaneous_microbatch_size": 16384,
1675
+ "objective/train/original_loss": 2.494450569152832,
1676
+ "objective/train/theoretical_loss": 3.4818803525066153,
1677
+ "objective/train/tokens_used": 48311776,
1678
+ "theoretical_loss": 3.4818803525066153,
1679
+ "tokens_seen": 1677852672
1680
+ },
1681
+ {
1682
+ "epoch": 0.02,
1683
+ "learning_rate": 9.930187770823304e-05,
1684
+ "loss": 2.4562,
1685
+ "theoretical_loss": 3.481868617943804,
1686
+ "tokens_seen": 1677918208
1687
+ },
1688
+ {
1689
+ "epoch": 0.02,
1690
+ "learning_rate": 9.92938533140748e-05,
1691
+ "loss": 2.4934,
1692
+ "theoretical_loss": 3.481845150578063,
1693
+ "tokens_seen": 1678049280
1694
+ },
1695
+ {
1696
+ "epoch": 0.02,
1697
+ "learning_rate": 9.928582891991655e-05,
1698
+ "loss": 2.4625,
1699
+ "theoretical_loss": 3.481821685558484,
1700
+ "tokens_seen": 1678180352
1701
+ },
1702
+ {
1703
+ "epoch": 0.02,
1704
+ "learning_rate": 9.927780452575832e-05,
1705
+ "loss": 2.3664,
1706
+ "theoretical_loss": 3.4817982228846476,
1707
+ "tokens_seen": 1678311424
1708
+ },
1709
+ {
1710
+ "epoch": 0.02,
1711
+ "learning_rate": 9.926978013160007e-05,
1712
+ "loss": 2.4554,
1713
+ "theoretical_loss": 3.481774762556137,
1714
+ "tokens_seen": 1678442496
1715
+ },
1716
+ {
1717
+ "epoch": 0.02,
1718
+ "learning_rate": 9.926175573744183e-05,
1719
+ "loss": 2.6398,
1720
+ "theoretical_loss": 3.4817513045725343,
1721
+ "tokens_seen": 1678573568
1722
+ },
1723
+ {
1724
+ "epoch": 0.02,
1725
+ "learning_rate": 9.925373134328359e-05,
1726
+ "loss": 2.3632,
1727
+ "theoretical_loss": 3.4817278489334225,
1728
+ "tokens_seen": 1678704640
1729
+ },
1730
+ {
1731
+ "epoch": 0.02,
1732
+ "learning_rate": 9.924570694912535e-05,
1733
+ "loss": 2.6002,
1734
+ "theoretical_loss": 3.481704395638383,
1735
+ "tokens_seen": 1678835712
1736
+ },
1737
+ {
1738
+ "epoch": 0.02,
1739
+ "learning_rate": 9.923768255496711e-05,
1740
+ "loss": 2.5935,
1741
+ "theoretical_loss": 3.4816809446870005,
1742
+ "tokens_seen": 1678966784
1743
+ },
1744
+ {
1745
+ "epoch": 0.02,
1746
+ "learning_rate": 9.922965816080886e-05,
1747
+ "loss": 2.6889,
1748
+ "theoretical_loss": 3.4816574960788564,
1749
+ "tokens_seen": 1679097856
1750
+ },
1751
+ {
1752
+ "epoch": 0.02,
1753
+ "learning_rate": 9.922163376665063e-05,
1754
+ "loss": 2.2718,
1755
+ "theoretical_loss": 3.4816340498135343,
1756
+ "tokens_seen": 1679228928
1757
+ },
1758
+ {
1759
+ "epoch": 0.02,
1760
+ "learning_rate": 9.921360937249238e-05,
1761
+ "loss": 2.4179,
1762
+ "theoretical_loss": 3.4816106058906175,
1763
+ "tokens_seen": 1679360000
1764
+ },
1765
+ {
1766
+ "epoch": 0.02,
1767
+ "objective/train/docs_used": 926923,
1768
+ "objective/train/instantaneous_batch_size": 16,
1769
+ "objective/train/instantaneous_microbatch_size": 16384,
1770
+ "objective/train/original_loss": 2.7133426666259766,
1771
+ "objective/train/theoretical_loss": 3.481587164309688,
1772
+ "objective/train/tokens_used": 49950176,
1773
+ "theoretical_loss": 3.481587164309688,
1774
+ "tokens_seen": 1679491072
1775
+ },
1776
+ {
1777
+ "epoch": 0.02,
1778
+ "learning_rate": 9.920558497833415e-05,
1779
+ "loss": 2.4835,
1780
+ "theoretical_loss": 3.481587164309688,
1781
+ "tokens_seen": 1679491072
1782
+ },
1783
+ {
1784
+ "epoch": 0.02,
1785
+ "learning_rate": 9.91975605841759e-05,
1786
+ "loss": 2.5025,
1787
+ "theoretical_loss": 3.4815637250703304,
1788
+ "tokens_seen": 1679622144
1789
+ },
1790
+ {
1791
+ "epoch": 0.02,
1792
+ "learning_rate": 9.918953619001767e-05,
1793
+ "loss": 2.4519,
1794
+ "theoretical_loss": 3.4815402881721274,
1795
+ "tokens_seen": 1679753216
1796
+ },
1797
+ {
1798
+ "epoch": 0.02,
1799
+ "learning_rate": 9.918151179585942e-05,
1800
+ "loss": 2.5021,
1801
+ "theoretical_loss": 3.4815168536146626,
1802
+ "tokens_seen": 1679884288
1803
+ },
1804
+ {
1805
+ "epoch": 0.02,
1806
+ "learning_rate": 9.917348740170119e-05,
1807
+ "loss": 2.4785,
1808
+ "theoretical_loss": 3.48149342139752,
1809
+ "tokens_seen": 1680015360
1810
+ },
1811
+ {
1812
+ "epoch": 0.02,
1813
+ "learning_rate": 9.916546300754294e-05,
1814
+ "loss": 2.5138,
1815
+ "theoretical_loss": 3.481469991520283,
1816
+ "tokens_seen": 1680146432
1817
+ },
1818
+ {
1819
+ "epoch": 0.02,
1820
+ "learning_rate": 9.91574386133847e-05,
1821
+ "loss": 2.4019,
1822
+ "theoretical_loss": 3.4814465639825354,
1823
+ "tokens_seen": 1680277504
1824
+ },
1825
+ {
1826
+ "epoch": 0.02,
1827
+ "learning_rate": 9.914941421922646e-05,
1828
+ "loss": 2.5178,
1829
+ "theoretical_loss": 3.481423138783861,
1830
+ "tokens_seen": 1680408576
1831
+ },
1832
+ {
1833
+ "epoch": 0.02,
1834
+ "learning_rate": 9.914138982506821e-05,
1835
+ "loss": 2.4615,
1836
+ "theoretical_loss": 3.4813997159238443,
1837
+ "tokens_seen": 1680539648
1838
+ },
1839
+ {
1840
+ "epoch": 0.02,
1841
+ "learning_rate": 9.913336543090998e-05,
1842
+ "loss": 2.533,
1843
+ "theoretical_loss": 3.481376295402069,
1844
+ "tokens_seen": 1680670720
1845
+ },
1846
+ {
1847
+ "epoch": 0.02,
1848
+ "learning_rate": 9.912534103675173e-05,
1849
+ "loss": 2.5408,
1850
+ "theoretical_loss": 3.4813528772181193,
1851
+ "tokens_seen": 1680801792
1852
+ },
1853
+ {
1854
+ "epoch": 0.02,
1855
+ "learning_rate": 9.91173166425935e-05,
1856
+ "loss": 2.4864,
1857
+ "theoretical_loss": 3.4813294613715797,
1858
+ "tokens_seen": 1680932864
1859
+ },
1860
+ {
1861
+ "epoch": 0.02,
1862
+ "learning_rate": 9.910929224843525e-05,
1863
+ "loss": 2.567,
1864
+ "theoretical_loss": 3.4813060478620352,
1865
+ "tokens_seen": 1681063936
1866
+ },
1867
+ {
1868
+ "epoch": 0.02,
1869
+ "objective/train/docs_used": 927976,
1870
+ "objective/train/instantaneous_batch_size": 16,
1871
+ "objective/train/instantaneous_microbatch_size": 16384,
1872
+ "objective/train/original_loss": 2.612856864929199,
1873
+ "objective/train/theoretical_loss": 3.4812943419835065,
1874
+ "objective/train/tokens_used": 51588576,
1875
+ "theoretical_loss": 3.4812943419835065,
1876
+ "tokens_seen": 1681129472
1877
+ },
1878
+ {
1879
+ "epoch": 0.02,
1880
+ "learning_rate": 9.9101267854277e-05,
1881
+ "loss": 2.515,
1882
+ "theoretical_loss": 3.48128263668907,
1883
+ "tokens_seen": 1681195008
1884
+ },
1885
+ {
1886
+ "epoch": 0.02,
1887
+ "learning_rate": 9.909324346011877e-05,
1888
+ "loss": 2.5095,
1889
+ "theoretical_loss": 3.4812592278522687,
1890
+ "tokens_seen": 1681326080
1891
+ },
1892
+ {
1893
+ "epoch": 0.02,
1894
+ "learning_rate": 9.908521906596053e-05,
1895
+ "loss": 2.6043,
1896
+ "theoretical_loss": 3.4812358213512162,
1897
+ "tokens_seen": 1681457152
1898
+ },
1899
+ {
1900
+ "epoch": 0.02,
1901
+ "learning_rate": 9.907719467180229e-05,
1902
+ "loss": 2.6276,
1903
+ "theoretical_loss": 3.4812124171854966,
1904
+ "tokens_seen": 1681588224
1905
+ },
1906
+ {
1907
+ "epoch": 0.02,
1908
+ "learning_rate": 9.906917027764405e-05,
1909
+ "loss": 2.3657,
1910
+ "theoretical_loss": 3.481189015354696,
1911
+ "tokens_seen": 1681719296
1912
+ },
1913
+ {
1914
+ "epoch": 0.02,
1915
+ "learning_rate": 9.906114588348581e-05,
1916
+ "loss": 2.4345,
1917
+ "theoretical_loss": 3.481165615858399,
1918
+ "tokens_seen": 1681850368
1919
+ },
1920
+ {
1921
+ "epoch": 0.02,
1922
+ "learning_rate": 9.905312148932757e-05,
1923
+ "loss": 2.6024,
1924
+ "theoretical_loss": 3.481142218696191,
1925
+ "tokens_seen": 1681981440
1926
+ },
1927
+ {
1928
+ "epoch": 0.02,
1929
+ "learning_rate": 9.904509709516932e-05,
1930
+ "loss": 2.5844,
1931
+ "theoretical_loss": 3.481118823867657,
1932
+ "tokens_seen": 1682112512
1933
+ },
1934
+ {
1935
+ "epoch": 0.02,
1936
+ "learning_rate": 9.903707270101109e-05,
1937
+ "loss": 2.4985,
1938
+ "theoretical_loss": 3.481095431372383,
1939
+ "tokens_seen": 1682243584
1940
+ },
1941
+ {
1942
+ "epoch": 0.02,
1943
+ "learning_rate": 9.902904830685284e-05,
1944
+ "loss": 2.4409,
1945
+ "theoretical_loss": 3.481072041209954,
1946
+ "tokens_seen": 1682374656
1947
+ },
1948
+ {
1949
+ "epoch": 0.02,
1950
+ "learning_rate": 9.90210239126946e-05,
1951
+ "loss": 2.4786,
1952
+ "theoretical_loss": 3.481048653379955,
1953
+ "tokens_seen": 1682505728
1954
+ },
1955
+ {
1956
+ "epoch": 0.02,
1957
+ "learning_rate": 9.901299951853636e-05,
1958
+ "loss": 2.4033,
1959
+ "theoretical_loss": 3.4810252678819733,
1960
+ "tokens_seen": 1682636800
1961
+ },
1962
+ {
1963
+ "epoch": 0.02,
1964
+ "objective/train/docs_used": 929273,
1965
+ "objective/train/instantaneous_batch_size": 16,
1966
+ "objective/train/instantaneous_microbatch_size": 16384,
1967
+ "objective/train/original_loss": 2.6835052967071533,
1968
+ "objective/train/theoretical_loss": 3.481001884715594,
1969
+ "objective/train/tokens_used": 53226976,
1970
+ "theoretical_loss": 3.481001884715594,
1971
+ "tokens_seen": 1682767872
1972
+ },
1973
+ {
1974
+ "epoch": 0.02,
1975
+ "learning_rate": 9.900497512437812e-05,
1976
+ "loss": 2.5045,
1977
+ "theoretical_loss": 3.481001884715594,
1978
+ "tokens_seen": 1682767872
1979
+ },
1980
+ {
1981
+ "epoch": 0.02,
1982
+ "learning_rate": 9.899695073021988e-05,
1983
+ "loss": 2.4493,
1984
+ "theoretical_loss": 3.4809785038804026,
1985
+ "tokens_seen": 1682898944
1986
+ }
1987
+ ],
1988
+ "max_steps": 12588,
1989
+ "num_train_epochs": 9223372036854775807,
1990
+ "total_flos": 1.6789580808192e+16,
1991
+ "trial_name": null,
1992
+ "trial_params": null
1993
+ }
checkpoint-251/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51bec2d3e2d88152dfe0a6734219bf08800b6233ea9aabb72d1cd25d9d6a0463
3
+ size 3375
checkpoint-251/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "tomekkorbak/nervous_wozniak",
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": false,
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": 50259
39
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b02bbc577c8f6f348df77928a160eb8e385a9ed7974d47ab6ff04543cddf4763
3
+ size 510402665
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
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
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": "gpt2",
7
+ "special_tokens_map_file": null,
8
+ "tokenizer_class": "GPT2Tokenizer",
9
+ "unk_token": "<|endoftext|>"
10
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51bec2d3e2d88152dfe0a6734219bf08800b6233ea9aabb72d1cd25d9d6a0463
3
+ size 3375
vocab.json ADDED
The diff for this file is too large to render. See raw diff