craa commited on
Commit
dbc5d25
·
verified ·
1 Parent(s): 88d52b4

Training in progress, step 10000, checkpoint

Browse files
checkpoint-10000/config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "GPT2LMHeadModel"
5
+ ],
6
+ "attn_pdrop": 0.1,
7
+ "bos_token_id": 50256,
8
+ "embd_pdrop": 0.1,
9
+ "eos_token_id": 50256,
10
+ "initializer_range": 0.02,
11
+ "layer_norm_epsilon": 1e-05,
12
+ "model_type": "gpt2",
13
+ "n_embd": 768,
14
+ "n_head": 12,
15
+ "n_inner": null,
16
+ "n_layer": 12,
17
+ "n_positions": 1024,
18
+ "reorder_and_upcast_attn": false,
19
+ "resid_pdrop": 0.1,
20
+ "scale_attn_by_inverse_layer_idx": false,
21
+ "scale_attn_weights": true,
22
+ "summary_activation": null,
23
+ "summary_first_dropout": 0.1,
24
+ "summary_proj_to_labels": true,
25
+ "summary_type": "cls_index",
26
+ "summary_use_proj": true,
27
+ "torch_dtype": "float32",
28
+ "transformers_version": "4.47.0.dev0",
29
+ "use_cache": true,
30
+ "vocab_size": 52000
31
+ }
checkpoint-10000/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.47.0.dev0"
6
+ }
checkpoint-10000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2649d43c7fc7792c518322fba4f4b6fc0d05b878feabff86a5bbac8189c7014
3
+ size 503128704
checkpoint-10000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6220bf647422ca97741ee52e9023675439a3233ecc1737dfcc428d081dbfc038
3
+ size 1006351290
checkpoint-10000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7ea6aba18815b6bff32277260718425e74ea8e5c5878dbd8a44da097ca7838de
3
+ size 14244
checkpoint-10000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:90c311f0069114271a5dba5d72dcffc82e7a33081dd5f150aa8c9b2278c1c497
3
+ size 1064
checkpoint-10000/special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {}
checkpoint-10000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-10000/tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-10000/trainer_state.json ADDED
@@ -0,0 +1,1523 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 3.7489354610443115,
3
+ "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_8397/checkpoint-10000",
4
+ "epoch": 1.0763104079216446,
5
+ "eval_steps": 1000,
6
+ "global_step": 10000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.005381552039608223,
13
+ "grad_norm": 1.3847299814224243,
14
+ "learning_rate": 0.0003,
15
+ "loss": 8.4655,
16
+ "step": 50
17
+ },
18
+ {
19
+ "epoch": 0.010763104079216447,
20
+ "grad_norm": 1.0632425546646118,
21
+ "learning_rate": 0.0006,
22
+ "loss": 6.8249,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 0.01614465611882467,
27
+ "grad_norm": 2.088347911834717,
28
+ "learning_rate": 0.0005996767589699385,
29
+ "loss": 6.4221,
30
+ "step": 150
31
+ },
32
+ {
33
+ "epoch": 0.021526208158432893,
34
+ "grad_norm": 1.0710954666137695,
35
+ "learning_rate": 0.0005993535179398771,
36
+ "loss": 6.1909,
37
+ "step": 200
38
+ },
39
+ {
40
+ "epoch": 0.026907760198041114,
41
+ "grad_norm": 1.5236896276474,
42
+ "learning_rate": 0.0005990302769098158,
43
+ "loss": 6.0442,
44
+ "step": 250
45
+ },
46
+ {
47
+ "epoch": 0.03228931223764934,
48
+ "grad_norm": 1.8247500658035278,
49
+ "learning_rate": 0.0005987070358797543,
50
+ "loss": 5.9513,
51
+ "step": 300
52
+ },
53
+ {
54
+ "epoch": 0.03767086427725756,
55
+ "grad_norm": 1.2985302209854126,
56
+ "learning_rate": 0.0005983837948496929,
57
+ "loss": 5.8643,
58
+ "step": 350
59
+ },
60
+ {
61
+ "epoch": 0.04305241631686579,
62
+ "grad_norm": 1.256401538848877,
63
+ "learning_rate": 0.0005980605538196314,
64
+ "loss": 5.8018,
65
+ "step": 400
66
+ },
67
+ {
68
+ "epoch": 0.048433968356474004,
69
+ "grad_norm": 0.8569295406341553,
70
+ "learning_rate": 0.0005977373127895701,
71
+ "loss": 5.7053,
72
+ "step": 450
73
+ },
74
+ {
75
+ "epoch": 0.05381552039608223,
76
+ "grad_norm": 0.860461413860321,
77
+ "learning_rate": 0.0005974140717595086,
78
+ "loss": 5.6413,
79
+ "step": 500
80
+ },
81
+ {
82
+ "epoch": 0.05919707243569045,
83
+ "grad_norm": 1.6800360679626465,
84
+ "learning_rate": 0.0005970908307294472,
85
+ "loss": 5.5789,
86
+ "step": 550
87
+ },
88
+ {
89
+ "epoch": 0.06457862447529868,
90
+ "grad_norm": 1.408766508102417,
91
+ "learning_rate": 0.0005967675896993858,
92
+ "loss": 5.4957,
93
+ "step": 600
94
+ },
95
+ {
96
+ "epoch": 0.0699601765149069,
97
+ "grad_norm": 1.263355016708374,
98
+ "learning_rate": 0.0005964443486693243,
99
+ "loss": 5.421,
100
+ "step": 650
101
+ },
102
+ {
103
+ "epoch": 0.07534172855451512,
104
+ "grad_norm": 1.617833137512207,
105
+ "learning_rate": 0.000596121107639263,
106
+ "loss": 5.334,
107
+ "step": 700
108
+ },
109
+ {
110
+ "epoch": 0.08072328059412334,
111
+ "grad_norm": 1.2163816690444946,
112
+ "learning_rate": 0.0005957978666092015,
113
+ "loss": 5.2925,
114
+ "step": 750
115
+ },
116
+ {
117
+ "epoch": 0.08610483263373157,
118
+ "grad_norm": 1.4054591655731201,
119
+ "learning_rate": 0.0005954746255791401,
120
+ "loss": 5.222,
121
+ "step": 800
122
+ },
123
+ {
124
+ "epoch": 0.09148638467333979,
125
+ "grad_norm": 1.0674018859863281,
126
+ "learning_rate": 0.0005951513845490787,
127
+ "loss": 5.2071,
128
+ "step": 850
129
+ },
130
+ {
131
+ "epoch": 0.09686793671294801,
132
+ "grad_norm": 1.4989951848983765,
133
+ "learning_rate": 0.0005948281435190174,
134
+ "loss": 5.1526,
135
+ "step": 900
136
+ },
137
+ {
138
+ "epoch": 0.10224948875255624,
139
+ "grad_norm": 1.028799057006836,
140
+ "learning_rate": 0.0005945049024889559,
141
+ "loss": 5.1098,
142
+ "step": 950
143
+ },
144
+ {
145
+ "epoch": 0.10763104079216446,
146
+ "grad_norm": 0.955634355545044,
147
+ "learning_rate": 0.0005941816614588944,
148
+ "loss": 5.0633,
149
+ "step": 1000
150
+ },
151
+ {
152
+ "epoch": 0.10763104079216446,
153
+ "eval_accuracy": 0.22915491630956988,
154
+ "eval_loss": 4.996466636657715,
155
+ "eval_runtime": 202.5064,
156
+ "eval_samples_per_second": 88.94,
157
+ "eval_steps_per_second": 5.56,
158
+ "step": 1000
159
+ },
160
+ {
161
+ "epoch": 0.11301259283177269,
162
+ "grad_norm": 1.2216582298278809,
163
+ "learning_rate": 0.000593858420428833,
164
+ "loss": 5.0219,
165
+ "step": 1050
166
+ },
167
+ {
168
+ "epoch": 0.1183941448713809,
169
+ "grad_norm": 1.2919893264770508,
170
+ "learning_rate": 0.0005935351793987716,
171
+ "loss": 4.9962,
172
+ "step": 1100
173
+ },
174
+ {
175
+ "epoch": 0.12377569691098914,
176
+ "grad_norm": 1.842529058456421,
177
+ "learning_rate": 0.0005932119383687103,
178
+ "loss": 4.9736,
179
+ "step": 1150
180
+ },
181
+ {
182
+ "epoch": 0.12915724895059735,
183
+ "grad_norm": 0.9416630268096924,
184
+ "learning_rate": 0.0005928886973386488,
185
+ "loss": 4.9687,
186
+ "step": 1200
187
+ },
188
+ {
189
+ "epoch": 0.13453880099020557,
190
+ "grad_norm": 1.340468406677246,
191
+ "learning_rate": 0.0005925654563085874,
192
+ "loss": 4.9251,
193
+ "step": 1250
194
+ },
195
+ {
196
+ "epoch": 0.1399203530298138,
197
+ "grad_norm": 1.4057867527008057,
198
+ "learning_rate": 0.000592242215278526,
199
+ "loss": 4.8782,
200
+ "step": 1300
201
+ },
202
+ {
203
+ "epoch": 0.14530190506942203,
204
+ "grad_norm": 1.3462789058685303,
205
+ "learning_rate": 0.0005919189742484645,
206
+ "loss": 4.8431,
207
+ "step": 1350
208
+ },
209
+ {
210
+ "epoch": 0.15068345710903025,
211
+ "grad_norm": 0.9291279911994934,
212
+ "learning_rate": 0.0005915957332184032,
213
+ "loss": 4.8316,
214
+ "step": 1400
215
+ },
216
+ {
217
+ "epoch": 0.15606500914863847,
218
+ "grad_norm": 0.7789126634597778,
219
+ "learning_rate": 0.0005912724921883417,
220
+ "loss": 4.8028,
221
+ "step": 1450
222
+ },
223
+ {
224
+ "epoch": 0.16144656118824668,
225
+ "grad_norm": 1.2111430168151855,
226
+ "learning_rate": 0.0005909492511582803,
227
+ "loss": 4.7837,
228
+ "step": 1500
229
+ },
230
+ {
231
+ "epoch": 0.1668281132278549,
232
+ "grad_norm": 1.1703121662139893,
233
+ "learning_rate": 0.0005906260101282189,
234
+ "loss": 4.7456,
235
+ "step": 1550
236
+ },
237
+ {
238
+ "epoch": 0.17220966526746315,
239
+ "grad_norm": 0.9755436182022095,
240
+ "learning_rate": 0.0005903027690981575,
241
+ "loss": 4.7303,
242
+ "step": 1600
243
+ },
244
+ {
245
+ "epoch": 0.17759121730707136,
246
+ "grad_norm": 0.9121830463409424,
247
+ "learning_rate": 0.000589979528068096,
248
+ "loss": 4.7437,
249
+ "step": 1650
250
+ },
251
+ {
252
+ "epoch": 0.18297276934667958,
253
+ "grad_norm": 1.2337779998779297,
254
+ "learning_rate": 0.0005896562870380347,
255
+ "loss": 4.6991,
256
+ "step": 1700
257
+ },
258
+ {
259
+ "epoch": 0.1883543213862878,
260
+ "grad_norm": 0.7923302054405212,
261
+ "learning_rate": 0.0005893330460079732,
262
+ "loss": 4.6819,
263
+ "step": 1750
264
+ },
265
+ {
266
+ "epoch": 0.19373587342589602,
267
+ "grad_norm": 0.9414947032928467,
268
+ "learning_rate": 0.0005890098049779118,
269
+ "loss": 4.6401,
270
+ "step": 1800
271
+ },
272
+ {
273
+ "epoch": 0.19911742546550426,
274
+ "grad_norm": 0.8575401306152344,
275
+ "learning_rate": 0.0005886865639478504,
276
+ "loss": 4.6503,
277
+ "step": 1850
278
+ },
279
+ {
280
+ "epoch": 0.20449897750511248,
281
+ "grad_norm": 0.874113917350769,
282
+ "learning_rate": 0.0005883633229177889,
283
+ "loss": 4.6262,
284
+ "step": 1900
285
+ },
286
+ {
287
+ "epoch": 0.2098805295447207,
288
+ "grad_norm": 0.7579442262649536,
289
+ "learning_rate": 0.0005880400818877276,
290
+ "loss": 4.5946,
291
+ "step": 1950
292
+ },
293
+ {
294
+ "epoch": 0.2152620815843289,
295
+ "grad_norm": 0.8062272071838379,
296
+ "learning_rate": 0.0005877168408576662,
297
+ "loss": 4.579,
298
+ "step": 2000
299
+ },
300
+ {
301
+ "epoch": 0.2152620815843289,
302
+ "eval_accuracy": 0.2712283567521957,
303
+ "eval_loss": 4.50425910949707,
304
+ "eval_runtime": 211.1194,
305
+ "eval_samples_per_second": 85.312,
306
+ "eval_steps_per_second": 5.333,
307
+ "step": 2000
308
+ },
309
+ {
310
+ "epoch": 0.22064363362393713,
311
+ "grad_norm": 1.1572167873382568,
312
+ "learning_rate": 0.0005873935998276048,
313
+ "loss": 4.5564,
314
+ "step": 2050
315
+ },
316
+ {
317
+ "epoch": 0.22602518566354537,
318
+ "grad_norm": 1.1494579315185547,
319
+ "learning_rate": 0.0005870703587975433,
320
+ "loss": 4.5425,
321
+ "step": 2100
322
+ },
323
+ {
324
+ "epoch": 0.2314067377031536,
325
+ "grad_norm": 0.9105241298675537,
326
+ "learning_rate": 0.0005867471177674818,
327
+ "loss": 4.5083,
328
+ "step": 2150
329
+ },
330
+ {
331
+ "epoch": 0.2367882897427618,
332
+ "grad_norm": 0.9526411890983582,
333
+ "learning_rate": 0.0005864238767374205,
334
+ "loss": 4.5128,
335
+ "step": 2200
336
+ },
337
+ {
338
+ "epoch": 0.24216984178237003,
339
+ "grad_norm": 0.7726457715034485,
340
+ "learning_rate": 0.0005861006357073591,
341
+ "loss": 4.4899,
342
+ "step": 2250
343
+ },
344
+ {
345
+ "epoch": 0.24755139382197827,
346
+ "grad_norm": 1.0442702770233154,
347
+ "learning_rate": 0.0005857773946772977,
348
+ "loss": 4.4716,
349
+ "step": 2300
350
+ },
351
+ {
352
+ "epoch": 0.2529329458615865,
353
+ "grad_norm": 1.1448233127593994,
354
+ "learning_rate": 0.0005854541536472362,
355
+ "loss": 4.4607,
356
+ "step": 2350
357
+ },
358
+ {
359
+ "epoch": 0.2583144979011947,
360
+ "grad_norm": 0.9515467286109924,
361
+ "learning_rate": 0.0005851309126171749,
362
+ "loss": 4.4544,
363
+ "step": 2400
364
+ },
365
+ {
366
+ "epoch": 0.2636960499408029,
367
+ "grad_norm": 0.7929104566574097,
368
+ "learning_rate": 0.0005848076715871134,
369
+ "loss": 4.4392,
370
+ "step": 2450
371
+ },
372
+ {
373
+ "epoch": 0.26907760198041114,
374
+ "grad_norm": 1.1213116645812988,
375
+ "learning_rate": 0.000584484430557052,
376
+ "loss": 4.4407,
377
+ "step": 2500
378
+ },
379
+ {
380
+ "epoch": 0.27445915402001936,
381
+ "grad_norm": 1.006108045578003,
382
+ "learning_rate": 0.0005841611895269906,
383
+ "loss": 4.3997,
384
+ "step": 2550
385
+ },
386
+ {
387
+ "epoch": 0.2798407060596276,
388
+ "grad_norm": 0.7756773829460144,
389
+ "learning_rate": 0.0005838379484969291,
390
+ "loss": 4.3982,
391
+ "step": 2600
392
+ },
393
+ {
394
+ "epoch": 0.2852222580992358,
395
+ "grad_norm": 0.7442255020141602,
396
+ "learning_rate": 0.0005835147074668678,
397
+ "loss": 4.3751,
398
+ "step": 2650
399
+ },
400
+ {
401
+ "epoch": 0.29060381013884407,
402
+ "grad_norm": 0.7645350694656372,
403
+ "learning_rate": 0.0005831914664368063,
404
+ "loss": 4.3873,
405
+ "step": 2700
406
+ },
407
+ {
408
+ "epoch": 0.2959853621784523,
409
+ "grad_norm": 0.7028666734695435,
410
+ "learning_rate": 0.0005828682254067449,
411
+ "loss": 4.3659,
412
+ "step": 2750
413
+ },
414
+ {
415
+ "epoch": 0.3013669142180605,
416
+ "grad_norm": 0.8504071831703186,
417
+ "learning_rate": 0.0005825449843766835,
418
+ "loss": 4.3617,
419
+ "step": 2800
420
+ },
421
+ {
422
+ "epoch": 0.3067484662576687,
423
+ "grad_norm": 0.880102276802063,
424
+ "learning_rate": 0.0005822217433466221,
425
+ "loss": 4.3115,
426
+ "step": 2850
427
+ },
428
+ {
429
+ "epoch": 0.31213001829727693,
430
+ "grad_norm": 0.7290977835655212,
431
+ "learning_rate": 0.0005818985023165607,
432
+ "loss": 4.3215,
433
+ "step": 2900
434
+ },
435
+ {
436
+ "epoch": 0.31751157033688515,
437
+ "grad_norm": 0.8019910454750061,
438
+ "learning_rate": 0.0005815752612864992,
439
+ "loss": 4.3149,
440
+ "step": 2950
441
+ },
442
+ {
443
+ "epoch": 0.32289312237649337,
444
+ "grad_norm": 0.7739085555076599,
445
+ "learning_rate": 0.0005812520202564378,
446
+ "loss": 4.322,
447
+ "step": 3000
448
+ },
449
+ {
450
+ "epoch": 0.32289312237649337,
451
+ "eval_accuracy": 0.2992396144952079,
452
+ "eval_loss": 4.226711750030518,
453
+ "eval_runtime": 199.2176,
454
+ "eval_samples_per_second": 90.409,
455
+ "eval_steps_per_second": 5.652,
456
+ "step": 3000
457
+ },
458
+ {
459
+ "epoch": 0.3282746744161016,
460
+ "grad_norm": 1.0869840383529663,
461
+ "learning_rate": 0.0005809287792263764,
462
+ "loss": 4.3072,
463
+ "step": 3050
464
+ },
465
+ {
466
+ "epoch": 0.3336562264557098,
467
+ "grad_norm": 0.7337630391120911,
468
+ "learning_rate": 0.0005806055381963151,
469
+ "loss": 4.2759,
470
+ "step": 3100
471
+ },
472
+ {
473
+ "epoch": 0.3390377784953181,
474
+ "grad_norm": 0.7336916923522949,
475
+ "learning_rate": 0.0005802822971662536,
476
+ "loss": 4.2838,
477
+ "step": 3150
478
+ },
479
+ {
480
+ "epoch": 0.3444193305349263,
481
+ "grad_norm": 0.7901497483253479,
482
+ "learning_rate": 0.0005799590561361922,
483
+ "loss": 4.2757,
484
+ "step": 3200
485
+ },
486
+ {
487
+ "epoch": 0.3498008825745345,
488
+ "grad_norm": 0.9568474888801575,
489
+ "learning_rate": 0.0005796358151061307,
490
+ "loss": 4.2574,
491
+ "step": 3250
492
+ },
493
+ {
494
+ "epoch": 0.35518243461414273,
495
+ "grad_norm": 0.9219122529029846,
496
+ "learning_rate": 0.0005793125740760694,
497
+ "loss": 4.2423,
498
+ "step": 3300
499
+ },
500
+ {
501
+ "epoch": 0.36056398665375095,
502
+ "grad_norm": 0.6552030444145203,
503
+ "learning_rate": 0.0005789893330460079,
504
+ "loss": 4.2339,
505
+ "step": 3350
506
+ },
507
+ {
508
+ "epoch": 0.36594553869335916,
509
+ "grad_norm": 0.6318998336791992,
510
+ "learning_rate": 0.0005786660920159465,
511
+ "loss": 4.2253,
512
+ "step": 3400
513
+ },
514
+ {
515
+ "epoch": 0.3713270907329674,
516
+ "grad_norm": 0.7610264420509338,
517
+ "learning_rate": 0.0005783428509858851,
518
+ "loss": 4.2369,
519
+ "step": 3450
520
+ },
521
+ {
522
+ "epoch": 0.3767086427725756,
523
+ "grad_norm": 0.6418978571891785,
524
+ "learning_rate": 0.0005780196099558237,
525
+ "loss": 4.2239,
526
+ "step": 3500
527
+ },
528
+ {
529
+ "epoch": 0.3820901948121838,
530
+ "grad_norm": 0.7791884541511536,
531
+ "learning_rate": 0.0005776963689257623,
532
+ "loss": 4.2253,
533
+ "step": 3550
534
+ },
535
+ {
536
+ "epoch": 0.38747174685179203,
537
+ "grad_norm": 0.7014104127883911,
538
+ "learning_rate": 0.0005773731278957008,
539
+ "loss": 4.1888,
540
+ "step": 3600
541
+ },
542
+ {
543
+ "epoch": 0.3928532988914003,
544
+ "grad_norm": 0.7691745758056641,
545
+ "learning_rate": 0.0005770498868656394,
546
+ "loss": 4.2055,
547
+ "step": 3650
548
+ },
549
+ {
550
+ "epoch": 0.3982348509310085,
551
+ "grad_norm": 0.7330165505409241,
552
+ "learning_rate": 0.000576726645835578,
553
+ "loss": 4.1917,
554
+ "step": 3700
555
+ },
556
+ {
557
+ "epoch": 0.40361640297061674,
558
+ "grad_norm": 0.6409156918525696,
559
+ "learning_rate": 0.0005764034048055167,
560
+ "loss": 4.1739,
561
+ "step": 3750
562
+ },
563
+ {
564
+ "epoch": 0.40899795501022496,
565
+ "grad_norm": 0.751600444316864,
566
+ "learning_rate": 0.0005760801637754552,
567
+ "loss": 4.1815,
568
+ "step": 3800
569
+ },
570
+ {
571
+ "epoch": 0.4143795070498332,
572
+ "grad_norm": 0.6341384649276733,
573
+ "learning_rate": 0.0005757569227453937,
574
+ "loss": 4.1752,
575
+ "step": 3850
576
+ },
577
+ {
578
+ "epoch": 0.4197610590894414,
579
+ "grad_norm": 0.8716601729393005,
580
+ "learning_rate": 0.0005754336817153324,
581
+ "loss": 4.1772,
582
+ "step": 3900
583
+ },
584
+ {
585
+ "epoch": 0.4251426111290496,
586
+ "grad_norm": 0.7268335223197937,
587
+ "learning_rate": 0.0005751104406852709,
588
+ "loss": 4.1579,
589
+ "step": 3950
590
+ },
591
+ {
592
+ "epoch": 0.4305241631686578,
593
+ "grad_norm": 0.7490660548210144,
594
+ "learning_rate": 0.0005747871996552096,
595
+ "loss": 4.1439,
596
+ "step": 4000
597
+ },
598
+ {
599
+ "epoch": 0.4305241631686578,
600
+ "eval_accuracy": 0.31218462820231296,
601
+ "eval_loss": 4.092499256134033,
602
+ "eval_runtime": 199.4467,
603
+ "eval_samples_per_second": 90.305,
604
+ "eval_steps_per_second": 5.646,
605
+ "step": 4000
606
+ },
607
+ {
608
+ "epoch": 0.43590571520826604,
609
+ "grad_norm": 0.9424974918365479,
610
+ "learning_rate": 0.0005744639586251481,
611
+ "loss": 4.1592,
612
+ "step": 4050
613
+ },
614
+ {
615
+ "epoch": 0.44128726724787426,
616
+ "grad_norm": 0.771391749382019,
617
+ "learning_rate": 0.0005741407175950867,
618
+ "loss": 4.1455,
619
+ "step": 4100
620
+ },
621
+ {
622
+ "epoch": 0.44666881928748253,
623
+ "grad_norm": 0.7916195392608643,
624
+ "learning_rate": 0.0005738174765650253,
625
+ "loss": 4.1329,
626
+ "step": 4150
627
+ },
628
+ {
629
+ "epoch": 0.45205037132709075,
630
+ "grad_norm": 0.7015873789787292,
631
+ "learning_rate": 0.0005734942355349638,
632
+ "loss": 4.1338,
633
+ "step": 4200
634
+ },
635
+ {
636
+ "epoch": 0.45743192336669897,
637
+ "grad_norm": 0.62185138463974,
638
+ "learning_rate": 0.0005731709945049025,
639
+ "loss": 4.1517,
640
+ "step": 4250
641
+ },
642
+ {
643
+ "epoch": 0.4628134754063072,
644
+ "grad_norm": 0.6499165296554565,
645
+ "learning_rate": 0.000572847753474841,
646
+ "loss": 4.1249,
647
+ "step": 4300
648
+ },
649
+ {
650
+ "epoch": 0.4681950274459154,
651
+ "grad_norm": 0.8142028450965881,
652
+ "learning_rate": 0.0005725245124447796,
653
+ "loss": 4.1094,
654
+ "step": 4350
655
+ },
656
+ {
657
+ "epoch": 0.4735765794855236,
658
+ "grad_norm": 0.686265230178833,
659
+ "learning_rate": 0.0005722012714147182,
660
+ "loss": 4.1156,
661
+ "step": 4400
662
+ },
663
+ {
664
+ "epoch": 0.47895813152513184,
665
+ "grad_norm": 0.6861261129379272,
666
+ "learning_rate": 0.0005718780303846568,
667
+ "loss": 4.0911,
668
+ "step": 4450
669
+ },
670
+ {
671
+ "epoch": 0.48433968356474005,
672
+ "grad_norm": 0.717899739742279,
673
+ "learning_rate": 0.0005715547893545953,
674
+ "loss": 4.0984,
675
+ "step": 4500
676
+ },
677
+ {
678
+ "epoch": 0.48972123560434827,
679
+ "grad_norm": 0.635183572769165,
680
+ "learning_rate": 0.000571231548324534,
681
+ "loss": 4.0915,
682
+ "step": 4550
683
+ },
684
+ {
685
+ "epoch": 0.49510278764395654,
686
+ "grad_norm": 0.6602552533149719,
687
+ "learning_rate": 0.0005709083072944725,
688
+ "loss": 4.0886,
689
+ "step": 4600
690
+ },
691
+ {
692
+ "epoch": 0.5004843396835648,
693
+ "grad_norm": 0.7243141531944275,
694
+ "learning_rate": 0.0005705850662644111,
695
+ "loss": 4.0928,
696
+ "step": 4650
697
+ },
698
+ {
699
+ "epoch": 0.505865891723173,
700
+ "grad_norm": 0.6310231685638428,
701
+ "learning_rate": 0.0005702618252343497,
702
+ "loss": 4.0986,
703
+ "step": 4700
704
+ },
705
+ {
706
+ "epoch": 0.5112474437627812,
707
+ "grad_norm": 0.5868064761161804,
708
+ "learning_rate": 0.0005699385842042882,
709
+ "loss": 4.0687,
710
+ "step": 4750
711
+ },
712
+ {
713
+ "epoch": 0.5166289958023894,
714
+ "grad_norm": 0.6389384269714355,
715
+ "learning_rate": 0.0005696153431742269,
716
+ "loss": 4.0789,
717
+ "step": 4800
718
+ },
719
+ {
720
+ "epoch": 0.5220105478419976,
721
+ "grad_norm": 0.7776073217391968,
722
+ "learning_rate": 0.0005692921021441655,
723
+ "loss": 4.0977,
724
+ "step": 4850
725
+ },
726
+ {
727
+ "epoch": 0.5273920998816058,
728
+ "grad_norm": 0.6234896779060364,
729
+ "learning_rate": 0.0005689688611141041,
730
+ "loss": 4.0547,
731
+ "step": 4900
732
+ },
733
+ {
734
+ "epoch": 0.5327736519212141,
735
+ "grad_norm": 0.8750380277633667,
736
+ "learning_rate": 0.0005686456200840426,
737
+ "loss": 4.0718,
738
+ "step": 4950
739
+ },
740
+ {
741
+ "epoch": 0.5381552039608223,
742
+ "grad_norm": 0.7700529098510742,
743
+ "learning_rate": 0.0005683223790539811,
744
+ "loss": 4.043,
745
+ "step": 5000
746
+ },
747
+ {
748
+ "epoch": 0.5381552039608223,
749
+ "eval_accuracy": 0.3214325100957547,
750
+ "eval_loss": 3.9926369190216064,
751
+ "eval_runtime": 225.5059,
752
+ "eval_samples_per_second": 79.869,
753
+ "eval_steps_per_second": 4.993,
754
+ "step": 5000
755
+ },
756
+ {
757
+ "epoch": 0.5435367560004305,
758
+ "grad_norm": 0.7643857002258301,
759
+ "learning_rate": 0.0005679991380239198,
760
+ "loss": 4.0605,
761
+ "step": 5050
762
+ },
763
+ {
764
+ "epoch": 0.5489183080400387,
765
+ "grad_norm": 0.554442286491394,
766
+ "learning_rate": 0.0005676758969938584,
767
+ "loss": 4.0446,
768
+ "step": 5100
769
+ },
770
+ {
771
+ "epoch": 0.5542998600796469,
772
+ "grad_norm": 0.7216308116912842,
773
+ "learning_rate": 0.000567352655963797,
774
+ "loss": 4.0313,
775
+ "step": 5150
776
+ },
777
+ {
778
+ "epoch": 0.5596814121192552,
779
+ "grad_norm": 0.6057455539703369,
780
+ "learning_rate": 0.0005670294149337355,
781
+ "loss": 4.0337,
782
+ "step": 5200
783
+ },
784
+ {
785
+ "epoch": 0.5650629641588634,
786
+ "grad_norm": 0.535860002040863,
787
+ "learning_rate": 0.0005667061739036742,
788
+ "loss": 4.0392,
789
+ "step": 5250
790
+ },
791
+ {
792
+ "epoch": 0.5704445161984716,
793
+ "grad_norm": 0.647204577922821,
794
+ "learning_rate": 0.0005663829328736127,
795
+ "loss": 4.0421,
796
+ "step": 5300
797
+ },
798
+ {
799
+ "epoch": 0.5758260682380799,
800
+ "grad_norm": 0.5126392841339111,
801
+ "learning_rate": 0.0005660596918435512,
802
+ "loss": 4.0318,
803
+ "step": 5350
804
+ },
805
+ {
806
+ "epoch": 0.5812076202776881,
807
+ "grad_norm": 0.6040687561035156,
808
+ "learning_rate": 0.0005657364508134899,
809
+ "loss": 4.0117,
810
+ "step": 5400
811
+ },
812
+ {
813
+ "epoch": 0.5865891723172963,
814
+ "grad_norm": 0.6971921324729919,
815
+ "learning_rate": 0.0005654132097834284,
816
+ "loss": 4.025,
817
+ "step": 5450
818
+ },
819
+ {
820
+ "epoch": 0.5919707243569046,
821
+ "grad_norm": 0.5343753099441528,
822
+ "learning_rate": 0.0005650899687533671,
823
+ "loss": 4.015,
824
+ "step": 5500
825
+ },
826
+ {
827
+ "epoch": 0.5973522763965128,
828
+ "grad_norm": 0.5600801706314087,
829
+ "learning_rate": 0.0005647667277233056,
830
+ "loss": 4.0087,
831
+ "step": 5550
832
+ },
833
+ {
834
+ "epoch": 0.602733828436121,
835
+ "grad_norm": 0.6542683243751526,
836
+ "learning_rate": 0.0005644434866932442,
837
+ "loss": 3.9957,
838
+ "step": 5600
839
+ },
840
+ {
841
+ "epoch": 0.6081153804757292,
842
+ "grad_norm": 0.6249606609344482,
843
+ "learning_rate": 0.0005641202456631828,
844
+ "loss": 4.0136,
845
+ "step": 5650
846
+ },
847
+ {
848
+ "epoch": 0.6134969325153374,
849
+ "grad_norm": 0.6572969555854797,
850
+ "learning_rate": 0.0005637970046331214,
851
+ "loss": 4.0046,
852
+ "step": 5700
853
+ },
854
+ {
855
+ "epoch": 0.6188784845549457,
856
+ "grad_norm": 0.6926669478416443,
857
+ "learning_rate": 0.00056347376360306,
858
+ "loss": 3.9869,
859
+ "step": 5750
860
+ },
861
+ {
862
+ "epoch": 0.6242600365945539,
863
+ "grad_norm": 0.6491366624832153,
864
+ "learning_rate": 0.0005631505225729985,
865
+ "loss": 4.0004,
866
+ "step": 5800
867
+ },
868
+ {
869
+ "epoch": 0.6296415886341621,
870
+ "grad_norm": 0.6138956546783447,
871
+ "learning_rate": 0.0005628272815429371,
872
+ "loss": 3.9814,
873
+ "step": 5850
874
+ },
875
+ {
876
+ "epoch": 0.6350231406737703,
877
+ "grad_norm": 0.6467788219451904,
878
+ "learning_rate": 0.0005625040405128757,
879
+ "loss": 3.9771,
880
+ "step": 5900
881
+ },
882
+ {
883
+ "epoch": 0.6404046927133785,
884
+ "grad_norm": 0.6469590067863464,
885
+ "learning_rate": 0.0005621807994828143,
886
+ "loss": 3.9882,
887
+ "step": 5950
888
+ },
889
+ {
890
+ "epoch": 0.6457862447529867,
891
+ "grad_norm": 0.7396846413612366,
892
+ "learning_rate": 0.0005618575584527529,
893
+ "loss": 3.9759,
894
+ "step": 6000
895
+ },
896
+ {
897
+ "epoch": 0.6457862447529867,
898
+ "eval_accuracy": 0.32833664054615025,
899
+ "eval_loss": 3.919649839401245,
900
+ "eval_runtime": 200.1239,
901
+ "eval_samples_per_second": 89.999,
902
+ "eval_steps_per_second": 5.627,
903
+ "step": 6000
904
+ },
905
+ {
906
+ "epoch": 0.651167796792595,
907
+ "grad_norm": 0.7368170619010925,
908
+ "learning_rate": 0.0005615343174226915,
909
+ "loss": 3.9789,
910
+ "step": 6050
911
+ },
912
+ {
913
+ "epoch": 0.6565493488322032,
914
+ "grad_norm": 0.593928873538971,
915
+ "learning_rate": 0.00056121107639263,
916
+ "loss": 3.962,
917
+ "step": 6100
918
+ },
919
+ {
920
+ "epoch": 0.6619309008718114,
921
+ "grad_norm": 0.5571395754814148,
922
+ "learning_rate": 0.0005608878353625687,
923
+ "loss": 3.9732,
924
+ "step": 6150
925
+ },
926
+ {
927
+ "epoch": 0.6673124529114196,
928
+ "grad_norm": 0.6184373497962952,
929
+ "learning_rate": 0.0005605645943325072,
930
+ "loss": 3.9618,
931
+ "step": 6200
932
+ },
933
+ {
934
+ "epoch": 0.6726940049510278,
935
+ "grad_norm": 0.6731172800064087,
936
+ "learning_rate": 0.0005602413533024458,
937
+ "loss": 3.9629,
938
+ "step": 6250
939
+ },
940
+ {
941
+ "epoch": 0.6780755569906362,
942
+ "grad_norm": Infinity,
943
+ "learning_rate": 0.0005599245770929855,
944
+ "loss": 3.9661,
945
+ "step": 6300
946
+ },
947
+ {
948
+ "epoch": 0.6834571090302444,
949
+ "grad_norm": 0.6674894690513611,
950
+ "learning_rate": 0.0005596013360629242,
951
+ "loss": 3.9598,
952
+ "step": 6350
953
+ },
954
+ {
955
+ "epoch": 0.6888386610698526,
956
+ "grad_norm": 0.6270789504051208,
957
+ "learning_rate": 0.0005592780950328628,
958
+ "loss": 3.9443,
959
+ "step": 6400
960
+ },
961
+ {
962
+ "epoch": 0.6942202131094608,
963
+ "grad_norm": 0.6163814067840576,
964
+ "learning_rate": 0.0005589548540028014,
965
+ "loss": 3.9506,
966
+ "step": 6450
967
+ },
968
+ {
969
+ "epoch": 0.699601765149069,
970
+ "grad_norm": 0.6060782670974731,
971
+ "learning_rate": 0.0005586316129727399,
972
+ "loss": 3.9645,
973
+ "step": 6500
974
+ },
975
+ {
976
+ "epoch": 0.7049833171886772,
977
+ "grad_norm": 0.6246522068977356,
978
+ "learning_rate": 0.0005583083719426786,
979
+ "loss": 3.9614,
980
+ "step": 6550
981
+ },
982
+ {
983
+ "epoch": 0.7103648692282855,
984
+ "grad_norm": 0.603921115398407,
985
+ "learning_rate": 0.0005579851309126171,
986
+ "loss": 3.94,
987
+ "step": 6600
988
+ },
989
+ {
990
+ "epoch": 0.7157464212678937,
991
+ "grad_norm": 0.5655505061149597,
992
+ "learning_rate": 0.0005576618898825558,
993
+ "loss": 3.9318,
994
+ "step": 6650
995
+ },
996
+ {
997
+ "epoch": 0.7211279733075019,
998
+ "grad_norm": 0.604542076587677,
999
+ "learning_rate": 0.0005573386488524943,
1000
+ "loss": 3.9307,
1001
+ "step": 6700
1002
+ },
1003
+ {
1004
+ "epoch": 0.7265095253471101,
1005
+ "grad_norm": 0.600004255771637,
1006
+ "learning_rate": 0.0005570154078224328,
1007
+ "loss": 3.9442,
1008
+ "step": 6750
1009
+ },
1010
+ {
1011
+ "epoch": 0.7318910773867183,
1012
+ "grad_norm": 0.6960250735282898,
1013
+ "learning_rate": 0.0005566921667923715,
1014
+ "loss": 3.9425,
1015
+ "step": 6800
1016
+ },
1017
+ {
1018
+ "epoch": 0.7372726294263265,
1019
+ "grad_norm": 0.5619109869003296,
1020
+ "learning_rate": 0.00055636892576231,
1021
+ "loss": 3.9073,
1022
+ "step": 6850
1023
+ },
1024
+ {
1025
+ "epoch": 0.7426541814659348,
1026
+ "grad_norm": 0.6036113500595093,
1027
+ "learning_rate": 0.0005560456847322487,
1028
+ "loss": 3.9305,
1029
+ "step": 6900
1030
+ },
1031
+ {
1032
+ "epoch": 0.748035733505543,
1033
+ "grad_norm": 0.6234127283096313,
1034
+ "learning_rate": 0.0005557224437021872,
1035
+ "loss": 3.9285,
1036
+ "step": 6950
1037
+ },
1038
+ {
1039
+ "epoch": 0.7534172855451512,
1040
+ "grad_norm": 0.6131523251533508,
1041
+ "learning_rate": 0.0005553992026721258,
1042
+ "loss": 3.922,
1043
+ "step": 7000
1044
+ },
1045
+ {
1046
+ "epoch": 0.7534172855451512,
1047
+ "eval_accuracy": 0.3334209437785411,
1048
+ "eval_loss": 3.861107349395752,
1049
+ "eval_runtime": 205.1306,
1050
+ "eval_samples_per_second": 87.803,
1051
+ "eval_steps_per_second": 5.489,
1052
+ "step": 7000
1053
+ },
1054
+ {
1055
+ "epoch": 0.7587988375847594,
1056
+ "grad_norm": 0.6058487296104431,
1057
+ "learning_rate": 0.0005550759616420644,
1058
+ "loss": 3.9083,
1059
+ "step": 7050
1060
+ },
1061
+ {
1062
+ "epoch": 0.7641803896243676,
1063
+ "grad_norm": 0.624620258808136,
1064
+ "learning_rate": 0.000554752720612003,
1065
+ "loss": 3.9349,
1066
+ "step": 7100
1067
+ },
1068
+ {
1069
+ "epoch": 0.7695619416639758,
1070
+ "grad_norm": 0.5525732040405273,
1071
+ "learning_rate": 0.0005544294795819415,
1072
+ "loss": 3.9261,
1073
+ "step": 7150
1074
+ },
1075
+ {
1076
+ "epoch": 0.7749434937035841,
1077
+ "grad_norm": 0.5950748324394226,
1078
+ "learning_rate": 0.0005541062385518801,
1079
+ "loss": 3.9067,
1080
+ "step": 7200
1081
+ },
1082
+ {
1083
+ "epoch": 0.7803250457431924,
1084
+ "grad_norm": 0.5052813291549683,
1085
+ "learning_rate": 0.0005537829975218188,
1086
+ "loss": 3.9004,
1087
+ "step": 7250
1088
+ },
1089
+ {
1090
+ "epoch": 0.7857065977828006,
1091
+ "grad_norm": 0.5610913038253784,
1092
+ "learning_rate": 0.0005534662213123586,
1093
+ "loss": 3.8911,
1094
+ "step": 7300
1095
+ },
1096
+ {
1097
+ "epoch": 0.7910881498224088,
1098
+ "grad_norm": 0.5639758110046387,
1099
+ "learning_rate": 0.0005531429802822971,
1100
+ "loss": 3.931,
1101
+ "step": 7350
1102
+ },
1103
+ {
1104
+ "epoch": 0.796469701862017,
1105
+ "grad_norm": 0.6221727728843689,
1106
+ "learning_rate": 0.0005528197392522357,
1107
+ "loss": 3.9288,
1108
+ "step": 7400
1109
+ },
1110
+ {
1111
+ "epoch": 0.8018512539016253,
1112
+ "grad_norm": 0.5807225108146667,
1113
+ "learning_rate": 0.0005524964982221743,
1114
+ "loss": 3.9206,
1115
+ "step": 7450
1116
+ },
1117
+ {
1118
+ "epoch": 0.8072328059412335,
1119
+ "grad_norm": 0.6214258074760437,
1120
+ "learning_rate": 0.0005521732571921129,
1121
+ "loss": 3.9216,
1122
+ "step": 7500
1123
+ },
1124
+ {
1125
+ "epoch": 0.8126143579808417,
1126
+ "grad_norm": 0.6088549494743347,
1127
+ "learning_rate": 0.0005518500161620514,
1128
+ "loss": 3.9102,
1129
+ "step": 7550
1130
+ },
1131
+ {
1132
+ "epoch": 0.8179959100204499,
1133
+ "grad_norm": 0.615315318107605,
1134
+ "learning_rate": 0.00055152677513199,
1135
+ "loss": 3.8932,
1136
+ "step": 7600
1137
+ },
1138
+ {
1139
+ "epoch": 0.8233774620600581,
1140
+ "grad_norm": 0.5932325124740601,
1141
+ "learning_rate": 0.0005512035341019286,
1142
+ "loss": 3.9037,
1143
+ "step": 7650
1144
+ },
1145
+ {
1146
+ "epoch": 0.8287590140996663,
1147
+ "grad_norm": 0.632127583026886,
1148
+ "learning_rate": 0.0005508802930718672,
1149
+ "loss": 3.9074,
1150
+ "step": 7700
1151
+ },
1152
+ {
1153
+ "epoch": 0.8341405661392746,
1154
+ "grad_norm": 0.5906286239624023,
1155
+ "learning_rate": 0.0005505570520418058,
1156
+ "loss": 3.8969,
1157
+ "step": 7750
1158
+ },
1159
+ {
1160
+ "epoch": 0.8395221181788828,
1161
+ "grad_norm": 0.7419958114624023,
1162
+ "learning_rate": 0.0005502338110117443,
1163
+ "loss": 3.911,
1164
+ "step": 7800
1165
+ },
1166
+ {
1167
+ "epoch": 0.844903670218491,
1168
+ "grad_norm": 0.5959650874137878,
1169
+ "learning_rate": 0.000549910569981683,
1170
+ "loss": 3.885,
1171
+ "step": 7850
1172
+ },
1173
+ {
1174
+ "epoch": 0.8502852222580992,
1175
+ "grad_norm": 0.5370450615882874,
1176
+ "learning_rate": 0.0005495873289516215,
1177
+ "loss": 3.8883,
1178
+ "step": 7900
1179
+ },
1180
+ {
1181
+ "epoch": 0.8556667742977074,
1182
+ "grad_norm": 0.6142526865005493,
1183
+ "learning_rate": 0.0005492640879215602,
1184
+ "loss": 3.8891,
1185
+ "step": 7950
1186
+ },
1187
+ {
1188
+ "epoch": 0.8610483263373157,
1189
+ "grad_norm": 0.5250054001808167,
1190
+ "learning_rate": 0.0005489408468914987,
1191
+ "loss": 3.8731,
1192
+ "step": 8000
1193
+ },
1194
+ {
1195
+ "epoch": 0.8610483263373157,
1196
+ "eval_accuracy": 0.3380041398923315,
1197
+ "eval_loss": 3.814573049545288,
1198
+ "eval_runtime": 218.7029,
1199
+ "eval_samples_per_second": 82.354,
1200
+ "eval_steps_per_second": 5.149,
1201
+ "step": 8000
1202
+ },
1203
+ {
1204
+ "epoch": 0.8664298783769239,
1205
+ "grad_norm": 0.5592741370201111,
1206
+ "learning_rate": 0.0005486176058614372,
1207
+ "loss": 3.8661,
1208
+ "step": 8050
1209
+ },
1210
+ {
1211
+ "epoch": 0.8718114304165321,
1212
+ "grad_norm": 0.558358371257782,
1213
+ "learning_rate": 0.0005482943648313759,
1214
+ "loss": 3.8665,
1215
+ "step": 8100
1216
+ },
1217
+ {
1218
+ "epoch": 0.8771929824561403,
1219
+ "grad_norm": 0.6065065860748291,
1220
+ "learning_rate": 0.0005479711238013145,
1221
+ "loss": 3.8815,
1222
+ "step": 8150
1223
+ },
1224
+ {
1225
+ "epoch": 0.8825745344957485,
1226
+ "grad_norm": 0.566906750202179,
1227
+ "learning_rate": 0.0005476478827712531,
1228
+ "loss": 3.8788,
1229
+ "step": 8200
1230
+ },
1231
+ {
1232
+ "epoch": 0.8879560865353568,
1233
+ "grad_norm": 0.5542231202125549,
1234
+ "learning_rate": 0.0005473246417411916,
1235
+ "loss": 3.8757,
1236
+ "step": 8250
1237
+ },
1238
+ {
1239
+ "epoch": 0.8933376385749651,
1240
+ "grad_norm": 0.5650736093521118,
1241
+ "learning_rate": 0.0005470014007111302,
1242
+ "loss": 3.867,
1243
+ "step": 8300
1244
+ },
1245
+ {
1246
+ "epoch": 0.8987191906145733,
1247
+ "grad_norm": 0.6359168887138367,
1248
+ "learning_rate": 0.0005466781596810688,
1249
+ "loss": 3.8704,
1250
+ "step": 8350
1251
+ },
1252
+ {
1253
+ "epoch": 0.9041007426541815,
1254
+ "grad_norm": 0.5609973669052124,
1255
+ "learning_rate": 0.0005463549186510073,
1256
+ "loss": 3.8704,
1257
+ "step": 8400
1258
+ },
1259
+ {
1260
+ "epoch": 0.9094822946937897,
1261
+ "grad_norm": 0.5966047048568726,
1262
+ "learning_rate": 0.000546031677620946,
1263
+ "loss": 3.8649,
1264
+ "step": 8450
1265
+ },
1266
+ {
1267
+ "epoch": 0.9148638467333979,
1268
+ "grad_norm": 0.555415689945221,
1269
+ "learning_rate": 0.0005457084365908845,
1270
+ "loss": 3.8775,
1271
+ "step": 8500
1272
+ },
1273
+ {
1274
+ "epoch": 0.9202453987730062,
1275
+ "grad_norm": 0.5719990134239197,
1276
+ "learning_rate": 0.0005453851955608232,
1277
+ "loss": 3.8601,
1278
+ "step": 8550
1279
+ },
1280
+ {
1281
+ "epoch": 0.9256269508126144,
1282
+ "grad_norm": 0.5464230179786682,
1283
+ "learning_rate": 0.0005450619545307617,
1284
+ "loss": 3.8599,
1285
+ "step": 8600
1286
+ },
1287
+ {
1288
+ "epoch": 0.9310085028522226,
1289
+ "grad_norm": 0.6069023013114929,
1290
+ "learning_rate": 0.0005447387135007003,
1291
+ "loss": 3.8584,
1292
+ "step": 8650
1293
+ },
1294
+ {
1295
+ "epoch": 0.9363900548918308,
1296
+ "grad_norm": 0.556610643863678,
1297
+ "learning_rate": 0.0005444154724706389,
1298
+ "loss": 3.8636,
1299
+ "step": 8700
1300
+ },
1301
+ {
1302
+ "epoch": 0.941771606931439,
1303
+ "grad_norm": 0.5988168120384216,
1304
+ "learning_rate": 0.0005440922314405775,
1305
+ "loss": 3.8366,
1306
+ "step": 8750
1307
+ },
1308
+ {
1309
+ "epoch": 0.9471531589710472,
1310
+ "grad_norm": 0.5634347200393677,
1311
+ "learning_rate": 0.0005437689904105161,
1312
+ "loss": 3.8604,
1313
+ "step": 8800
1314
+ },
1315
+ {
1316
+ "epoch": 0.9525347110106555,
1317
+ "grad_norm": 0.5349180102348328,
1318
+ "learning_rate": 0.0005434457493804546,
1319
+ "loss": 3.8395,
1320
+ "step": 8850
1321
+ },
1322
+ {
1323
+ "epoch": 0.9579162630502637,
1324
+ "grad_norm": 0.5537461042404175,
1325
+ "learning_rate": 0.0005431225083503932,
1326
+ "loss": 3.8644,
1327
+ "step": 8900
1328
+ },
1329
+ {
1330
+ "epoch": 0.9632978150898719,
1331
+ "grad_norm": 0.5711461901664734,
1332
+ "learning_rate": 0.0005427992673203318,
1333
+ "loss": 3.8352,
1334
+ "step": 8950
1335
+ },
1336
+ {
1337
+ "epoch": 0.9686793671294801,
1338
+ "grad_norm": 0.6128414273262024,
1339
+ "learning_rate": 0.0005424760262902704,
1340
+ "loss": 3.8405,
1341
+ "step": 9000
1342
+ },
1343
+ {
1344
+ "epoch": 0.9686793671294801,
1345
+ "eval_accuracy": 0.3417073562677125,
1346
+ "eval_loss": 3.779106378555298,
1347
+ "eval_runtime": 210.7209,
1348
+ "eval_samples_per_second": 85.473,
1349
+ "eval_steps_per_second": 5.344,
1350
+ "step": 9000
1351
+ },
1352
+ {
1353
+ "epoch": 0.9740609191690883,
1354
+ "grad_norm": 0.6156434416770935,
1355
+ "learning_rate": 0.000542152785260209,
1356
+ "loss": 3.8333,
1357
+ "step": 9050
1358
+ },
1359
+ {
1360
+ "epoch": 0.9794424712086965,
1361
+ "grad_norm": 0.5652556419372559,
1362
+ "learning_rate": 0.0005418295442301476,
1363
+ "loss": 3.8327,
1364
+ "step": 9100
1365
+ },
1366
+ {
1367
+ "epoch": 0.9848240232483048,
1368
+ "grad_norm": 0.6443074941635132,
1369
+ "learning_rate": 0.0005415063032000861,
1370
+ "loss": 3.842,
1371
+ "step": 9150
1372
+ },
1373
+ {
1374
+ "epoch": 0.9902055752879131,
1375
+ "grad_norm": 0.632297158241272,
1376
+ "learning_rate": 0.0005411830621700248,
1377
+ "loss": 3.8312,
1378
+ "step": 9200
1379
+ },
1380
+ {
1381
+ "epoch": 0.9955871273275213,
1382
+ "grad_norm": 0.5863285064697266,
1383
+ "learning_rate": 0.0005408598211399633,
1384
+ "loss": 3.8298,
1385
+ "step": 9250
1386
+ },
1387
+ {
1388
+ "epoch": 1.0009686793671295,
1389
+ "grad_norm": 0.5393086671829224,
1390
+ "learning_rate": 0.0005405365801099019,
1391
+ "loss": 3.8069,
1392
+ "step": 9300
1393
+ },
1394
+ {
1395
+ "epoch": 1.0063502314067376,
1396
+ "grad_norm": 0.5285312533378601,
1397
+ "learning_rate": 0.0005402133390798405,
1398
+ "loss": 3.7619,
1399
+ "step": 9350
1400
+ },
1401
+ {
1402
+ "epoch": 1.011731783446346,
1403
+ "grad_norm": 0.5921708941459656,
1404
+ "learning_rate": 0.000539890098049779,
1405
+ "loss": 3.7701,
1406
+ "step": 9400
1407
+ },
1408
+ {
1409
+ "epoch": 1.017113335485954,
1410
+ "grad_norm": 0.5384525060653687,
1411
+ "learning_rate": 0.0005395668570197177,
1412
+ "loss": 3.7657,
1413
+ "step": 9450
1414
+ },
1415
+ {
1416
+ "epoch": 1.0224948875255624,
1417
+ "grad_norm": 0.5624929666519165,
1418
+ "learning_rate": 0.0005392436159896562,
1419
+ "loss": 3.7754,
1420
+ "step": 9500
1421
+ },
1422
+ {
1423
+ "epoch": 1.0278764395651705,
1424
+ "grad_norm": 0.5882208943367004,
1425
+ "learning_rate": 0.0005389203749595948,
1426
+ "loss": 3.7848,
1427
+ "step": 9550
1428
+ },
1429
+ {
1430
+ "epoch": 1.0332579916047788,
1431
+ "grad_norm": 0.5588924884796143,
1432
+ "learning_rate": 0.0005385971339295334,
1433
+ "loss": 3.7603,
1434
+ "step": 9600
1435
+ },
1436
+ {
1437
+ "epoch": 1.0386395436443872,
1438
+ "grad_norm": 0.5837873220443726,
1439
+ "learning_rate": 0.000538273892899472,
1440
+ "loss": 3.7785,
1441
+ "step": 9650
1442
+ },
1443
+ {
1444
+ "epoch": 1.0440210956839953,
1445
+ "grad_norm": 0.5880899429321289,
1446
+ "learning_rate": 0.0005379506518694106,
1447
+ "loss": 3.7506,
1448
+ "step": 9700
1449
+ },
1450
+ {
1451
+ "epoch": 1.0494026477236036,
1452
+ "grad_norm": 0.5579642653465271,
1453
+ "learning_rate": 0.0005376274108393491,
1454
+ "loss": 3.7714,
1455
+ "step": 9750
1456
+ },
1457
+ {
1458
+ "epoch": 1.0547841997632117,
1459
+ "grad_norm": 0.576766312122345,
1460
+ "learning_rate": 0.0005373041698092877,
1461
+ "loss": 3.7653,
1462
+ "step": 9800
1463
+ },
1464
+ {
1465
+ "epoch": 1.06016575180282,
1466
+ "grad_norm": 0.5471925735473633,
1467
+ "learning_rate": 0.0005369809287792263,
1468
+ "loss": 3.7686,
1469
+ "step": 9850
1470
+ },
1471
+ {
1472
+ "epoch": 1.0655473038424281,
1473
+ "grad_norm": 0.59092116355896,
1474
+ "learning_rate": 0.000536657687749165,
1475
+ "loss": 3.7667,
1476
+ "step": 9900
1477
+ },
1478
+ {
1479
+ "epoch": 1.0709288558820365,
1480
+ "grad_norm": 0.638380765914917,
1481
+ "learning_rate": 0.0005363344467191035,
1482
+ "loss": 3.7722,
1483
+ "step": 9950
1484
+ },
1485
+ {
1486
+ "epoch": 1.0763104079216446,
1487
+ "grad_norm": 0.5594536066055298,
1488
+ "learning_rate": 0.000536011205689042,
1489
+ "loss": 3.7655,
1490
+ "step": 10000
1491
+ },
1492
+ {
1493
+ "epoch": 1.0763104079216446,
1494
+ "eval_accuracy": 0.3451049320696713,
1495
+ "eval_loss": 3.7489354610443115,
1496
+ "eval_runtime": 202.8767,
1497
+ "eval_samples_per_second": 88.778,
1498
+ "eval_steps_per_second": 5.55,
1499
+ "step": 10000
1500
+ }
1501
+ ],
1502
+ "logging_steps": 50,
1503
+ "max_steps": 92910,
1504
+ "num_input_tokens_seen": 0,
1505
+ "num_train_epochs": 10,
1506
+ "save_steps": 10000,
1507
+ "stateful_callbacks": {
1508
+ "TrainerControl": {
1509
+ "args": {
1510
+ "should_epoch_stop": false,
1511
+ "should_evaluate": false,
1512
+ "should_log": false,
1513
+ "should_save": true,
1514
+ "should_training_stop": false
1515
+ },
1516
+ "attributes": {}
1517
+ }
1518
+ },
1519
+ "total_flos": 8.3610314735616e+16,
1520
+ "train_batch_size": 32,
1521
+ "trial_name": null,
1522
+ "trial_params": null
1523
+ }
checkpoint-10000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8af7b84dbc09e2405215b378e28c9508606efb6c1cb6a8674665fa2271606b7
3
+ size 5304