craa commited on
Commit
c70f58c
·
verified ·
1 Parent(s): b9dd874

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:ad96b9c160abbb286c9901f6668f5c89fb035af44adc8dec20a236ee5a73aac9
3
+ size 503128704
checkpoint-10000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bb52d7abd4dcfc37ce37351b954e2a755221104864201eeae51a11baf4ab3d6b
3
+ size 1006351290
checkpoint-10000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c8f1b0ec2ac4559b9a7317e5177712fc8549e54b1b53dfcf14de603fbc071d0
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": 3.7523183822631836,
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": 1.4122834205627441,
14
+ "learning_rate": 0.00028799999999999995,
15
+ "loss": 8.5357,
16
+ "step": 50
17
+ },
18
+ {
19
+ "epoch": 0.010763104079216447,
20
+ "grad_norm": 3.4872183799743652,
21
+ "learning_rate": 0.000588,
22
+ "loss": 6.9207,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 0.01614465611882467,
27
+ "grad_norm": 2.1926398277282715,
28
+ "learning_rate": 0.000599689688611141,
29
+ "loss": 6.4649,
30
+ "step": 150
31
+ },
32
+ {
33
+ "epoch": 0.021526208158432893,
34
+ "grad_norm": 1.9389663934707642,
35
+ "learning_rate": 0.0005993664475810796,
36
+ "loss": 6.2572,
37
+ "step": 200
38
+ },
39
+ {
40
+ "epoch": 0.026907760198041114,
41
+ "grad_norm": 1.5312936305999756,
42
+ "learning_rate": 0.0005990432065510182,
43
+ "loss": 6.1012,
44
+ "step": 250
45
+ },
46
+ {
47
+ "epoch": 0.03228931223764934,
48
+ "grad_norm": 1.838247299194336,
49
+ "learning_rate": 0.0005987199655209567,
50
+ "loss": 5.9633,
51
+ "step": 300
52
+ },
53
+ {
54
+ "epoch": 0.03767086427725756,
55
+ "grad_norm": 1.8276017904281616,
56
+ "learning_rate": 0.0005983967244908953,
57
+ "loss": 5.8677,
58
+ "step": 350
59
+ },
60
+ {
61
+ "epoch": 0.04305241631686579,
62
+ "grad_norm": 1.031121015548706,
63
+ "learning_rate": 0.0005980734834608338,
64
+ "loss": 5.7804,
65
+ "step": 400
66
+ },
67
+ {
68
+ "epoch": 0.048433968356474004,
69
+ "grad_norm": 2.2222535610198975,
70
+ "learning_rate": 0.0005977502424307725,
71
+ "loss": 5.6924,
72
+ "step": 450
73
+ },
74
+ {
75
+ "epoch": 0.05381552039608223,
76
+ "grad_norm": 1.698441982269287,
77
+ "learning_rate": 0.0005974270014007111,
78
+ "loss": 5.654,
79
+ "step": 500
80
+ },
81
+ {
82
+ "epoch": 0.05919707243569045,
83
+ "grad_norm": 1.586506962776184,
84
+ "learning_rate": 0.0005971037603706497,
85
+ "loss": 5.5769,
86
+ "step": 550
87
+ },
88
+ {
89
+ "epoch": 0.06457862447529868,
90
+ "grad_norm": 1.3522851467132568,
91
+ "learning_rate": 0.0005967805193405882,
92
+ "loss": 5.4942,
93
+ "step": 600
94
+ },
95
+ {
96
+ "epoch": 0.0699601765149069,
97
+ "grad_norm": 1.594616413116455,
98
+ "learning_rate": 0.0005964572783105269,
99
+ "loss": 5.4497,
100
+ "step": 650
101
+ },
102
+ {
103
+ "epoch": 0.07534172855451512,
104
+ "grad_norm": 1.3532969951629639,
105
+ "learning_rate": 0.0005961340372804654,
106
+ "loss": 5.3653,
107
+ "step": 700
108
+ },
109
+ {
110
+ "epoch": 0.08072328059412334,
111
+ "grad_norm": 1.399992823600769,
112
+ "learning_rate": 0.000595810796250404,
113
+ "loss": 5.3081,
114
+ "step": 750
115
+ },
116
+ {
117
+ "epoch": 0.08610483263373157,
118
+ "grad_norm": 1.2744522094726562,
119
+ "learning_rate": 0.0005954875552203426,
120
+ "loss": 5.2635,
121
+ "step": 800
122
+ },
123
+ {
124
+ "epoch": 0.09148638467333979,
125
+ "grad_norm": 1.311295747756958,
126
+ "learning_rate": 0.0005951643141902811,
127
+ "loss": 5.2182,
128
+ "step": 850
129
+ },
130
+ {
131
+ "epoch": 0.09686793671294801,
132
+ "grad_norm": 1.1952992677688599,
133
+ "learning_rate": 0.0005948410731602198,
134
+ "loss": 5.1703,
135
+ "step": 900
136
+ },
137
+ {
138
+ "epoch": 0.10224948875255624,
139
+ "grad_norm": 1.2511310577392578,
140
+ "learning_rate": 0.0005945178321301583,
141
+ "loss": 5.132,
142
+ "step": 950
143
+ },
144
+ {
145
+ "epoch": 0.10763104079216446,
146
+ "grad_norm": 1.1788593530654907,
147
+ "learning_rate": 0.000594194591100097,
148
+ "loss": 5.0646,
149
+ "step": 1000
150
+ },
151
+ {
152
+ "epoch": 0.10763104079216446,
153
+ "eval_accuracy": 0.22730227591944518,
154
+ "eval_loss": 5.024896144866943,
155
+ "eval_runtime": 187.644,
156
+ "eval_samples_per_second": 95.985,
157
+ "eval_steps_per_second": 6.001,
158
+ "step": 1000
159
+ },
160
+ {
161
+ "epoch": 0.11301259283177269,
162
+ "grad_norm": 1.5058064460754395,
163
+ "learning_rate": 0.0005938713500700355,
164
+ "loss": 5.0372,
165
+ "step": 1050
166
+ },
167
+ {
168
+ "epoch": 0.1183941448713809,
169
+ "grad_norm": 1.2659214735031128,
170
+ "learning_rate": 0.000593548109039974,
171
+ "loss": 5.0221,
172
+ "step": 1100
173
+ },
174
+ {
175
+ "epoch": 0.12377569691098914,
176
+ "grad_norm": 1.4247196912765503,
177
+ "learning_rate": 0.0005932248680099127,
178
+ "loss": 4.9852,
179
+ "step": 1150
180
+ },
181
+ {
182
+ "epoch": 0.12915724895059735,
183
+ "grad_norm": 1.7715990543365479,
184
+ "learning_rate": 0.0005929016269798512,
185
+ "loss": 4.943,
186
+ "step": 1200
187
+ },
188
+ {
189
+ "epoch": 0.13453880099020557,
190
+ "grad_norm": 1.0012747049331665,
191
+ "learning_rate": 0.0005925783859497898,
192
+ "loss": 4.9166,
193
+ "step": 1250
194
+ },
195
+ {
196
+ "epoch": 0.1399203530298138,
197
+ "grad_norm": 1.266779899597168,
198
+ "learning_rate": 0.0005922551449197284,
199
+ "loss": 4.8965,
200
+ "step": 1300
201
+ },
202
+ {
203
+ "epoch": 0.14530190506942203,
204
+ "grad_norm": 0.9170055389404297,
205
+ "learning_rate": 0.0005919319038896671,
206
+ "loss": 4.8706,
207
+ "step": 1350
208
+ },
209
+ {
210
+ "epoch": 0.15068345710903025,
211
+ "grad_norm": 0.9823539853096008,
212
+ "learning_rate": 0.0005916086628596056,
213
+ "loss": 4.8504,
214
+ "step": 1400
215
+ },
216
+ {
217
+ "epoch": 0.15606500914863847,
218
+ "grad_norm": 1.2595840692520142,
219
+ "learning_rate": 0.0005912854218295442,
220
+ "loss": 4.813,
221
+ "step": 1450
222
+ },
223
+ {
224
+ "epoch": 0.16144656118824668,
225
+ "grad_norm": 0.8510214686393738,
226
+ "learning_rate": 0.0005909621807994827,
227
+ "loss": 4.7813,
228
+ "step": 1500
229
+ },
230
+ {
231
+ "epoch": 0.1668281132278549,
232
+ "grad_norm": 1.1120244264602661,
233
+ "learning_rate": 0.0005906389397694213,
234
+ "loss": 4.7637,
235
+ "step": 1550
236
+ },
237
+ {
238
+ "epoch": 0.17220966526746315,
239
+ "grad_norm": 1.0976771116256714,
240
+ "learning_rate": 0.00059031569873936,
241
+ "loss": 4.7563,
242
+ "step": 1600
243
+ },
244
+ {
245
+ "epoch": 0.17759121730707136,
246
+ "grad_norm": 0.9833604693412781,
247
+ "learning_rate": 0.0005899924577092985,
248
+ "loss": 4.7421,
249
+ "step": 1650
250
+ },
251
+ {
252
+ "epoch": 0.18297276934667958,
253
+ "grad_norm": 0.8846125602722168,
254
+ "learning_rate": 0.0005896692166792371,
255
+ "loss": 4.7031,
256
+ "step": 1700
257
+ },
258
+ {
259
+ "epoch": 0.1883543213862878,
260
+ "grad_norm": 1.0292339324951172,
261
+ "learning_rate": 0.0005893459756491757,
262
+ "loss": 4.6573,
263
+ "step": 1750
264
+ },
265
+ {
266
+ "epoch": 0.19373587342589602,
267
+ "grad_norm": 1.0036226511001587,
268
+ "learning_rate": 0.0005890227346191143,
269
+ "loss": 4.6505,
270
+ "step": 1800
271
+ },
272
+ {
273
+ "epoch": 0.19911742546550426,
274
+ "grad_norm": 0.9965416789054871,
275
+ "learning_rate": 0.0005886994935890529,
276
+ "loss": 4.6383,
277
+ "step": 1850
278
+ },
279
+ {
280
+ "epoch": 0.20449897750511248,
281
+ "grad_norm": 1.1083658933639526,
282
+ "learning_rate": 0.0005883762525589915,
283
+ "loss": 4.6478,
284
+ "step": 1900
285
+ },
286
+ {
287
+ "epoch": 0.2098805295447207,
288
+ "grad_norm": 0.93230140209198,
289
+ "learning_rate": 0.00058805301152893,
290
+ "loss": 4.5879,
291
+ "step": 1950
292
+ },
293
+ {
294
+ "epoch": 0.2152620815843289,
295
+ "grad_norm": 1.1331114768981934,
296
+ "learning_rate": 0.0005877297704988686,
297
+ "loss": 4.5685,
298
+ "step": 2000
299
+ },
300
+ {
301
+ "epoch": 0.2152620815843289,
302
+ "eval_accuracy": 0.271470652692022,
303
+ "eval_loss": 4.505092620849609,
304
+ "eval_runtime": 185.4709,
305
+ "eval_samples_per_second": 97.11,
306
+ "eval_steps_per_second": 6.071,
307
+ "step": 2000
308
+ },
309
+ {
310
+ "epoch": 0.22064363362393713,
311
+ "grad_norm": 0.9773377776145935,
312
+ "learning_rate": 0.0005874065294688072,
313
+ "loss": 4.5528,
314
+ "step": 2050
315
+ },
316
+ {
317
+ "epoch": 0.22602518566354537,
318
+ "grad_norm": 0.7613441348075867,
319
+ "learning_rate": 0.0005870832884387457,
320
+ "loss": 4.5416,
321
+ "step": 2100
322
+ },
323
+ {
324
+ "epoch": 0.2314067377031536,
325
+ "grad_norm": 0.9072310924530029,
326
+ "learning_rate": 0.0005867600474086844,
327
+ "loss": 4.5192,
328
+ "step": 2150
329
+ },
330
+ {
331
+ "epoch": 0.2367882897427618,
332
+ "grad_norm": 0.9957188963890076,
333
+ "learning_rate": 0.0005864368063786229,
334
+ "loss": 4.5067,
335
+ "step": 2200
336
+ },
337
+ {
338
+ "epoch": 0.24216984178237003,
339
+ "grad_norm": 1.090625286102295,
340
+ "learning_rate": 0.0005861135653485616,
341
+ "loss": 4.4835,
342
+ "step": 2250
343
+ },
344
+ {
345
+ "epoch": 0.24755139382197827,
346
+ "grad_norm": 1.0402551889419556,
347
+ "learning_rate": 0.0005857903243185001,
348
+ "loss": 4.4773,
349
+ "step": 2300
350
+ },
351
+ {
352
+ "epoch": 0.2529329458615865,
353
+ "grad_norm": 0.9853291511535645,
354
+ "learning_rate": 0.0005854670832884386,
355
+ "loss": 4.4576,
356
+ "step": 2350
357
+ },
358
+ {
359
+ "epoch": 0.2583144979011947,
360
+ "grad_norm": 0.777766227722168,
361
+ "learning_rate": 0.0005851438422583773,
362
+ "loss": 4.4518,
363
+ "step": 2400
364
+ },
365
+ {
366
+ "epoch": 0.2636960499408029,
367
+ "grad_norm": 1.0316593647003174,
368
+ "learning_rate": 0.0005848206012283159,
369
+ "loss": 4.43,
370
+ "step": 2450
371
+ },
372
+ {
373
+ "epoch": 0.26907760198041114,
374
+ "grad_norm": 0.7335191369056702,
375
+ "learning_rate": 0.0005844973601982545,
376
+ "loss": 4.4154,
377
+ "step": 2500
378
+ },
379
+ {
380
+ "epoch": 0.27445915402001936,
381
+ "grad_norm": 0.9138801097869873,
382
+ "learning_rate": 0.000584174119168193,
383
+ "loss": 4.396,
384
+ "step": 2550
385
+ },
386
+ {
387
+ "epoch": 0.2798407060596276,
388
+ "grad_norm": 0.6920284628868103,
389
+ "learning_rate": 0.0005838508781381316,
390
+ "loss": 4.4028,
391
+ "step": 2600
392
+ },
393
+ {
394
+ "epoch": 0.2852222580992358,
395
+ "grad_norm": 0.9748484492301941,
396
+ "learning_rate": 0.0005835276371080702,
397
+ "loss": 4.367,
398
+ "step": 2650
399
+ },
400
+ {
401
+ "epoch": 0.29060381013884407,
402
+ "grad_norm": 0.7634653449058533,
403
+ "learning_rate": 0.0005832043960780088,
404
+ "loss": 4.3595,
405
+ "step": 2700
406
+ },
407
+ {
408
+ "epoch": 0.2959853621784523,
409
+ "grad_norm": 0.6365559101104736,
410
+ "learning_rate": 0.0005828811550479474,
411
+ "loss": 4.3588,
412
+ "step": 2750
413
+ },
414
+ {
415
+ "epoch": 0.3013669142180605,
416
+ "grad_norm": 0.632849395275116,
417
+ "learning_rate": 0.0005825579140178859,
418
+ "loss": 4.3405,
419
+ "step": 2800
420
+ },
421
+ {
422
+ "epoch": 0.3067484662576687,
423
+ "grad_norm": 0.8527909517288208,
424
+ "learning_rate": 0.0005822346729878246,
425
+ "loss": 4.3207,
426
+ "step": 2850
427
+ },
428
+ {
429
+ "epoch": 0.31213001829727693,
430
+ "grad_norm": 0.948337197303772,
431
+ "learning_rate": 0.0005819114319577631,
432
+ "loss": 4.3312,
433
+ "step": 2900
434
+ },
435
+ {
436
+ "epoch": 0.31751157033688515,
437
+ "grad_norm": 0.8351525664329529,
438
+ "learning_rate": 0.0005815881909277017,
439
+ "loss": 4.3143,
440
+ "step": 2950
441
+ },
442
+ {
443
+ "epoch": 0.32289312237649337,
444
+ "grad_norm": 0.8625895977020264,
445
+ "learning_rate": 0.0005812649498976403,
446
+ "loss": 4.3002,
447
+ "step": 3000
448
+ },
449
+ {
450
+ "epoch": 0.32289312237649337,
451
+ "eval_accuracy": 0.29852044103076386,
452
+ "eval_loss": 4.236849308013916,
453
+ "eval_runtime": 184.7432,
454
+ "eval_samples_per_second": 97.492,
455
+ "eval_steps_per_second": 6.095,
456
+ "step": 3000
457
+ },
458
+ {
459
+ "epoch": 0.3282746744161016,
460
+ "grad_norm": 0.7663789391517639,
461
+ "learning_rate": 0.0005809417088675789,
462
+ "loss": 4.2959,
463
+ "step": 3050
464
+ },
465
+ {
466
+ "epoch": 0.3336562264557098,
467
+ "grad_norm": 0.7020168304443359,
468
+ "learning_rate": 0.0005806184678375175,
469
+ "loss": 4.2839,
470
+ "step": 3100
471
+ },
472
+ {
473
+ "epoch": 0.3390377784953181,
474
+ "grad_norm": 0.7872812151908875,
475
+ "learning_rate": 0.000580295226807456,
476
+ "loss": 4.2729,
477
+ "step": 3150
478
+ },
479
+ {
480
+ "epoch": 0.3444193305349263,
481
+ "grad_norm": 0.8375337719917297,
482
+ "learning_rate": 0.0005799719857773946,
483
+ "loss": 4.281,
484
+ "step": 3200
485
+ },
486
+ {
487
+ "epoch": 0.3498008825745345,
488
+ "grad_norm": 0.7012325525283813,
489
+ "learning_rate": 0.0005796487447473331,
490
+ "loss": 4.26,
491
+ "step": 3250
492
+ },
493
+ {
494
+ "epoch": 0.35518243461414273,
495
+ "grad_norm": 0.7597373723983765,
496
+ "learning_rate": 0.0005793255037172718,
497
+ "loss": 4.2512,
498
+ "step": 3300
499
+ },
500
+ {
501
+ "epoch": 0.36056398665375095,
502
+ "grad_norm": 0.7354339361190796,
503
+ "learning_rate": 0.0005790022626872104,
504
+ "loss": 4.2585,
505
+ "step": 3350
506
+ },
507
+ {
508
+ "epoch": 0.36594553869335916,
509
+ "grad_norm": 0.9357606768608093,
510
+ "learning_rate": 0.000578679021657149,
511
+ "loss": 4.2464,
512
+ "step": 3400
513
+ },
514
+ {
515
+ "epoch": 0.3713270907329674,
516
+ "grad_norm": 0.6800475716590881,
517
+ "learning_rate": 0.0005783557806270875,
518
+ "loss": 4.2159,
519
+ "step": 3450
520
+ },
521
+ {
522
+ "epoch": 0.3767086427725756,
523
+ "grad_norm": 0.8798818588256836,
524
+ "learning_rate": 0.0005780325395970262,
525
+ "loss": 4.2192,
526
+ "step": 3500
527
+ },
528
+ {
529
+ "epoch": 0.3820901948121838,
530
+ "grad_norm": 0.6990943551063538,
531
+ "learning_rate": 0.0005777092985669647,
532
+ "loss": 4.2043,
533
+ "step": 3550
534
+ },
535
+ {
536
+ "epoch": 0.38747174685179203,
537
+ "grad_norm": 0.732219398021698,
538
+ "learning_rate": 0.0005773860575369033,
539
+ "loss": 4.2124,
540
+ "step": 3600
541
+ },
542
+ {
543
+ "epoch": 0.3928532988914003,
544
+ "grad_norm": 0.768298327922821,
545
+ "learning_rate": 0.0005770628165068419,
546
+ "loss": 4.194,
547
+ "step": 3650
548
+ },
549
+ {
550
+ "epoch": 0.3982348509310085,
551
+ "grad_norm": 0.7568483948707581,
552
+ "learning_rate": 0.0005767395754767804,
553
+ "loss": 4.1734,
554
+ "step": 3700
555
+ },
556
+ {
557
+ "epoch": 0.40361640297061674,
558
+ "grad_norm": 0.7396944165229797,
559
+ "learning_rate": 0.0005764163344467191,
560
+ "loss": 4.1768,
561
+ "step": 3750
562
+ },
563
+ {
564
+ "epoch": 0.40899795501022496,
565
+ "grad_norm": 0.7737883925437927,
566
+ "learning_rate": 0.0005760930934166576,
567
+ "loss": 4.1866,
568
+ "step": 3800
569
+ },
570
+ {
571
+ "epoch": 0.4143795070498332,
572
+ "grad_norm": 0.6525033116340637,
573
+ "learning_rate": 0.0005757698523865963,
574
+ "loss": 4.1749,
575
+ "step": 3850
576
+ },
577
+ {
578
+ "epoch": 0.4197610590894414,
579
+ "grad_norm": 0.7739155888557434,
580
+ "learning_rate": 0.0005754466113565348,
581
+ "loss": 4.18,
582
+ "step": 3900
583
+ },
584
+ {
585
+ "epoch": 0.4251426111290496,
586
+ "grad_norm": 0.7185007333755493,
587
+ "learning_rate": 0.0005751233703264733,
588
+ "loss": 4.1561,
589
+ "step": 3950
590
+ },
591
+ {
592
+ "epoch": 0.4305241631686578,
593
+ "grad_norm": 0.6594641804695129,
594
+ "learning_rate": 0.000574800129296412,
595
+ "loss": 4.1579,
596
+ "step": 4000
597
+ },
598
+ {
599
+ "epoch": 0.4305241631686578,
600
+ "eval_accuracy": 0.3129942008694187,
601
+ "eval_loss": 4.0840744972229,
602
+ "eval_runtime": 184.6939,
603
+ "eval_samples_per_second": 97.518,
604
+ "eval_steps_per_second": 6.097,
605
+ "step": 4000
606
+ },
607
+ {
608
+ "epoch": 0.43590571520826604,
609
+ "grad_norm": 0.6553388833999634,
610
+ "learning_rate": 0.0005744768882663505,
611
+ "loss": 4.1519,
612
+ "step": 4050
613
+ },
614
+ {
615
+ "epoch": 0.44128726724787426,
616
+ "grad_norm": 0.7919243574142456,
617
+ "learning_rate": 0.0005741536472362891,
618
+ "loss": 4.165,
619
+ "step": 4100
620
+ },
621
+ {
622
+ "epoch": 0.44666881928748253,
623
+ "grad_norm": 0.5930292010307312,
624
+ "learning_rate": 0.0005738304062062277,
625
+ "loss": 4.1667,
626
+ "step": 4150
627
+ },
628
+ {
629
+ "epoch": 0.45205037132709075,
630
+ "grad_norm": 0.7211865782737732,
631
+ "learning_rate": 0.0005735071651761664,
632
+ "loss": 4.1264,
633
+ "step": 4200
634
+ },
635
+ {
636
+ "epoch": 0.45743192336669897,
637
+ "grad_norm": 0.7048370242118835,
638
+ "learning_rate": 0.0005731839241461049,
639
+ "loss": 4.1397,
640
+ "step": 4250
641
+ },
642
+ {
643
+ "epoch": 0.4628134754063072,
644
+ "grad_norm": 0.7572526931762695,
645
+ "learning_rate": 0.0005728606831160435,
646
+ "loss": 4.1205,
647
+ "step": 4300
648
+ },
649
+ {
650
+ "epoch": 0.4681950274459154,
651
+ "grad_norm": 0.7019510865211487,
652
+ "learning_rate": 0.000572537442085982,
653
+ "loss": 4.1274,
654
+ "step": 4350
655
+ },
656
+ {
657
+ "epoch": 0.4735765794855236,
658
+ "grad_norm": 0.693013072013855,
659
+ "learning_rate": 0.0005722142010559206,
660
+ "loss": 4.1165,
661
+ "step": 4400
662
+ },
663
+ {
664
+ "epoch": 0.47895813152513184,
665
+ "grad_norm": 0.6860193014144897,
666
+ "learning_rate": 0.0005718909600258593,
667
+ "loss": 4.1139,
668
+ "step": 4450
669
+ },
670
+ {
671
+ "epoch": 0.48433968356474005,
672
+ "grad_norm": 0.6195184588432312,
673
+ "learning_rate": 0.0005715677189957978,
674
+ "loss": 4.1266,
675
+ "step": 4500
676
+ },
677
+ {
678
+ "epoch": 0.48972123560434827,
679
+ "grad_norm": 0.7034777402877808,
680
+ "learning_rate": 0.0005712444779657364,
681
+ "loss": 4.0937,
682
+ "step": 4550
683
+ },
684
+ {
685
+ "epoch": 0.49510278764395654,
686
+ "grad_norm": 0.6804658770561218,
687
+ "learning_rate": 0.000570921236935675,
688
+ "loss": 4.1195,
689
+ "step": 4600
690
+ },
691
+ {
692
+ "epoch": 0.5004843396835648,
693
+ "grad_norm": 0.7629151940345764,
694
+ "learning_rate": 0.0005705979959056136,
695
+ "loss": 4.0871,
696
+ "step": 4650
697
+ },
698
+ {
699
+ "epoch": 0.505865891723173,
700
+ "grad_norm": 0.5974324345588684,
701
+ "learning_rate": 0.0005702747548755522,
702
+ "loss": 4.1022,
703
+ "step": 4700
704
+ },
705
+ {
706
+ "epoch": 0.5112474437627812,
707
+ "grad_norm": 0.5412039160728455,
708
+ "learning_rate": 0.0005699515138454908,
709
+ "loss": 4.0876,
710
+ "step": 4750
711
+ },
712
+ {
713
+ "epoch": 0.5166289958023894,
714
+ "grad_norm": 0.7911316752433777,
715
+ "learning_rate": 0.0005696282728154293,
716
+ "loss": 4.0688,
717
+ "step": 4800
718
+ },
719
+ {
720
+ "epoch": 0.5220105478419976,
721
+ "grad_norm": 0.68361496925354,
722
+ "learning_rate": 0.0005693050317853679,
723
+ "loss": 4.0913,
724
+ "step": 4850
725
+ },
726
+ {
727
+ "epoch": 0.5273920998816058,
728
+ "grad_norm": 0.577229380607605,
729
+ "learning_rate": 0.0005689817907553065,
730
+ "loss": 4.0676,
731
+ "step": 4900
732
+ },
733
+ {
734
+ "epoch": 0.5327736519212141,
735
+ "grad_norm": 0.7196305394172668,
736
+ "learning_rate": 0.000568658549725245,
737
+ "loss": 4.0467,
738
+ "step": 4950
739
+ },
740
+ {
741
+ "epoch": 0.5381552039608223,
742
+ "grad_norm": 0.5695436596870422,
743
+ "learning_rate": 0.0005683353086951837,
744
+ "loss": 4.062,
745
+ "step": 5000
746
+ },
747
+ {
748
+ "epoch": 0.5381552039608223,
749
+ "eval_accuracy": 0.3209510691498487,
750
+ "eval_loss": 3.9919753074645996,
751
+ "eval_runtime": 184.7142,
752
+ "eval_samples_per_second": 97.507,
753
+ "eval_steps_per_second": 6.096,
754
+ "step": 5000
755
+ },
756
+ {
757
+ "epoch": 0.5435367560004305,
758
+ "grad_norm": 0.5603345632553101,
759
+ "learning_rate": 0.0005680120676651222,
760
+ "loss": 4.0619,
761
+ "step": 5050
762
+ },
763
+ {
764
+ "epoch": 0.5489183080400387,
765
+ "grad_norm": 0.6007739305496216,
766
+ "learning_rate": 0.0005676888266350609,
767
+ "loss": 4.039,
768
+ "step": 5100
769
+ },
770
+ {
771
+ "epoch": 0.5542998600796469,
772
+ "grad_norm": 0.7048094272613525,
773
+ "learning_rate": 0.0005673655856049994,
774
+ "loss": 4.0414,
775
+ "step": 5150
776
+ },
777
+ {
778
+ "epoch": 0.5596814121192552,
779
+ "grad_norm": 0.6947103142738342,
780
+ "learning_rate": 0.0005670423445749379,
781
+ "loss": 4.0415,
782
+ "step": 5200
783
+ },
784
+ {
785
+ "epoch": 0.5650629641588634,
786
+ "grad_norm": 0.6355268359184265,
787
+ "learning_rate": 0.0005667191035448766,
788
+ "loss": 4.0565,
789
+ "step": 5250
790
+ },
791
+ {
792
+ "epoch": 0.5704445161984716,
793
+ "grad_norm": 0.6978614926338196,
794
+ "learning_rate": 0.0005663958625148152,
795
+ "loss": 4.0299,
796
+ "step": 5300
797
+ },
798
+ {
799
+ "epoch": 0.5758260682380799,
800
+ "grad_norm": 0.5592193007469177,
801
+ "learning_rate": 0.0005660726214847538,
802
+ "loss": 4.0185,
803
+ "step": 5350
804
+ },
805
+ {
806
+ "epoch": 0.5812076202776881,
807
+ "grad_norm": 0.6000945568084717,
808
+ "learning_rate": 0.0005657493804546923,
809
+ "loss": 4.0321,
810
+ "step": 5400
811
+ },
812
+ {
813
+ "epoch": 0.5865891723172963,
814
+ "grad_norm": 0.6557536721229553,
815
+ "learning_rate": 0.0005654261394246309,
816
+ "loss": 4.0062,
817
+ "step": 5450
818
+ },
819
+ {
820
+ "epoch": 0.5919707243569046,
821
+ "grad_norm": 0.6313118934631348,
822
+ "learning_rate": 0.0005651028983945695,
823
+ "loss": 4.0171,
824
+ "step": 5500
825
+ },
826
+ {
827
+ "epoch": 0.5973522763965128,
828
+ "grad_norm": 0.6380894184112549,
829
+ "learning_rate": 0.000564779657364508,
830
+ "loss": 4.0139,
831
+ "step": 5550
832
+ },
833
+ {
834
+ "epoch": 0.602733828436121,
835
+ "grad_norm": 0.5402942299842834,
836
+ "learning_rate": 0.0005644564163344467,
837
+ "loss": 4.0111,
838
+ "step": 5600
839
+ },
840
+ {
841
+ "epoch": 0.6081153804757292,
842
+ "grad_norm": 0.5897926688194275,
843
+ "learning_rate": 0.0005641331753043852,
844
+ "loss": 4.0321,
845
+ "step": 5650
846
+ },
847
+ {
848
+ "epoch": 0.6134969325153374,
849
+ "grad_norm": 0.6751918792724609,
850
+ "learning_rate": 0.0005638099342743239,
851
+ "loss": 4.014,
852
+ "step": 5700
853
+ },
854
+ {
855
+ "epoch": 0.6188784845549457,
856
+ "grad_norm": 0.5889696478843689,
857
+ "learning_rate": 0.0005634866932442624,
858
+ "loss": 3.9981,
859
+ "step": 5750
860
+ },
861
+ {
862
+ "epoch": 0.6242600365945539,
863
+ "grad_norm": 0.6100760698318481,
864
+ "learning_rate": 0.000563163452214201,
865
+ "loss": 4.0152,
866
+ "step": 5800
867
+ },
868
+ {
869
+ "epoch": 0.6296415886341621,
870
+ "grad_norm": 0.6364511847496033,
871
+ "learning_rate": 0.0005628402111841396,
872
+ "loss": 4.0011,
873
+ "step": 5850
874
+ },
875
+ {
876
+ "epoch": 0.6350231406737703,
877
+ "grad_norm": 0.6010422706604004,
878
+ "learning_rate": 0.0005625169701540782,
879
+ "loss": 3.9959,
880
+ "step": 5900
881
+ },
882
+ {
883
+ "epoch": 0.6404046927133785,
884
+ "grad_norm": 0.5707629919052124,
885
+ "learning_rate": 0.0005621937291240168,
886
+ "loss": 3.9918,
887
+ "step": 5950
888
+ },
889
+ {
890
+ "epoch": 0.6457862447529867,
891
+ "grad_norm": 0.6960643529891968,
892
+ "learning_rate": 0.0005618704880939553,
893
+ "loss": 3.9778,
894
+ "step": 6000
895
+ },
896
+ {
897
+ "epoch": 0.6457862447529867,
898
+ "eval_accuracy": 0.32842921280656817,
899
+ "eval_loss": 3.9220640659332275,
900
+ "eval_runtime": 184.5812,
901
+ "eval_samples_per_second": 97.578,
902
+ "eval_steps_per_second": 6.1,
903
+ "step": 6000
904
+ },
905
+ {
906
+ "epoch": 0.651167796792595,
907
+ "grad_norm": 0.5975825190544128,
908
+ "learning_rate": 0.0005615472470638939,
909
+ "loss": 3.9885,
910
+ "step": 6050
911
+ },
912
+ {
913
+ "epoch": 0.6565493488322032,
914
+ "grad_norm": 0.5811153650283813,
915
+ "learning_rate": 0.0005612240060338325,
916
+ "loss": 3.9683,
917
+ "step": 6100
918
+ },
919
+ {
920
+ "epoch": 0.6619309008718114,
921
+ "grad_norm": 0.6349126696586609,
922
+ "learning_rate": 0.0005609007650037711,
923
+ "loss": 3.978,
924
+ "step": 6150
925
+ },
926
+ {
927
+ "epoch": 0.6673124529114196,
928
+ "grad_norm": 0.5223124623298645,
929
+ "learning_rate": 0.0005605775239737097,
930
+ "loss": 3.9877,
931
+ "step": 6200
932
+ },
933
+ {
934
+ "epoch": 0.6726940049510278,
935
+ "grad_norm": 0.5649115443229675,
936
+ "learning_rate": 0.0005602542829436483,
937
+ "loss": 3.9654,
938
+ "step": 6250
939
+ },
940
+ {
941
+ "epoch": 0.6780755569906362,
942
+ "grad_norm": 0.5698593258857727,
943
+ "learning_rate": 0.0005599310419135868,
944
+ "loss": 3.9748,
945
+ "step": 6300
946
+ },
947
+ {
948
+ "epoch": 0.6834571090302444,
949
+ "grad_norm": 0.5707403421401978,
950
+ "learning_rate": 0.0005596078008835255,
951
+ "loss": 3.9573,
952
+ "step": 6350
953
+ },
954
+ {
955
+ "epoch": 0.6888386610698526,
956
+ "grad_norm": 0.6776087284088135,
957
+ "learning_rate": 0.000559284559853464,
958
+ "loss": 3.9625,
959
+ "step": 6400
960
+ },
961
+ {
962
+ "epoch": 0.6942202131094608,
963
+ "grad_norm": 0.5434262752532959,
964
+ "learning_rate": 0.0005589613188234026,
965
+ "loss": 3.9497,
966
+ "step": 6450
967
+ },
968
+ {
969
+ "epoch": 0.699601765149069,
970
+ "grad_norm": 0.621336817741394,
971
+ "learning_rate": 0.0005586380777933412,
972
+ "loss": 3.9608,
973
+ "step": 6500
974
+ },
975
+ {
976
+ "epoch": 0.7049833171886772,
977
+ "grad_norm": 0.6960099339485168,
978
+ "learning_rate": 0.0005583148367632797,
979
+ "loss": 3.9539,
980
+ "step": 6550
981
+ },
982
+ {
983
+ "epoch": 0.7103648692282855,
984
+ "grad_norm": 0.6768355965614319,
985
+ "learning_rate": 0.0005579915957332184,
986
+ "loss": 3.9431,
987
+ "step": 6600
988
+ },
989
+ {
990
+ "epoch": 0.7157464212678937,
991
+ "grad_norm": 0.7968887686729431,
992
+ "learning_rate": 0.0005576683547031569,
993
+ "loss": 3.963,
994
+ "step": 6650
995
+ },
996
+ {
997
+ "epoch": 0.7211279733075019,
998
+ "grad_norm": 0.6040535569190979,
999
+ "learning_rate": 0.0005573451136730956,
1000
+ "loss": 3.9357,
1001
+ "step": 6700
1002
+ },
1003
+ {
1004
+ "epoch": 0.7265095253471101,
1005
+ "grad_norm": 0.6244126558303833,
1006
+ "learning_rate": 0.0005570218726430341,
1007
+ "loss": 3.9461,
1008
+ "step": 6750
1009
+ },
1010
+ {
1011
+ "epoch": 0.7318910773867183,
1012
+ "grad_norm": 0.6094521284103394,
1013
+ "learning_rate": 0.0005566986316129728,
1014
+ "loss": 3.9556,
1015
+ "step": 6800
1016
+ },
1017
+ {
1018
+ "epoch": 0.7372726294263265,
1019
+ "grad_norm": 0.6165975332260132,
1020
+ "learning_rate": 0.0005563753905829113,
1021
+ "loss": 3.9451,
1022
+ "step": 6850
1023
+ },
1024
+ {
1025
+ "epoch": 0.7426541814659348,
1026
+ "grad_norm": 0.6186493635177612,
1027
+ "learning_rate": 0.0005560521495528498,
1028
+ "loss": 3.9314,
1029
+ "step": 6900
1030
+ },
1031
+ {
1032
+ "epoch": 0.748035733505543,
1033
+ "grad_norm": 0.60491943359375,
1034
+ "learning_rate": 0.0005557289085227884,
1035
+ "loss": 3.9439,
1036
+ "step": 6950
1037
+ },
1038
+ {
1039
+ "epoch": 0.7534172855451512,
1040
+ "grad_norm": 0.5608195662498474,
1041
+ "learning_rate": 0.000555405667492727,
1042
+ "loss": 3.9329,
1043
+ "step": 7000
1044
+ },
1045
+ {
1046
+ "epoch": 0.7534172855451512,
1047
+ "eval_accuracy": 0.3331331222787205,
1048
+ "eval_loss": 3.8616700172424316,
1049
+ "eval_runtime": 184.9857,
1050
+ "eval_samples_per_second": 97.364,
1051
+ "eval_steps_per_second": 6.087,
1052
+ "step": 7000
1053
+ },
1054
+ {
1055
+ "epoch": 0.7587988375847594,
1056
+ "grad_norm": 0.6202792525291443,
1057
+ "learning_rate": 0.0005550824264626657,
1058
+ "loss": 3.9319,
1059
+ "step": 7050
1060
+ },
1061
+ {
1062
+ "epoch": 0.7641803896243676,
1063
+ "grad_norm": 0.5904057025909424,
1064
+ "learning_rate": 0.0005547591854326042,
1065
+ "loss": 3.9246,
1066
+ "step": 7100
1067
+ },
1068
+ {
1069
+ "epoch": 0.7695619416639758,
1070
+ "grad_norm": 0.5333248972892761,
1071
+ "learning_rate": 0.0005544359444025428,
1072
+ "loss": 3.9287,
1073
+ "step": 7150
1074
+ },
1075
+ {
1076
+ "epoch": 0.7749434937035841,
1077
+ "grad_norm": 0.6606253981590271,
1078
+ "learning_rate": 0.0005541127033724813,
1079
+ "loss": 3.9263,
1080
+ "step": 7200
1081
+ },
1082
+ {
1083
+ "epoch": 0.7803250457431924,
1084
+ "grad_norm": 0.6061251759529114,
1085
+ "learning_rate": 0.0005537894623424199,
1086
+ "loss": 3.9207,
1087
+ "step": 7250
1088
+ },
1089
+ {
1090
+ "epoch": 0.7857065977828006,
1091
+ "grad_norm": 0.5618706345558167,
1092
+ "learning_rate": 0.0005534662213123586,
1093
+ "loss": 3.9275,
1094
+ "step": 7300
1095
+ },
1096
+ {
1097
+ "epoch": 0.7910881498224088,
1098
+ "grad_norm": 0.5616447329521179,
1099
+ "learning_rate": 0.0005531429802822971,
1100
+ "loss": 3.9249,
1101
+ "step": 7350
1102
+ },
1103
+ {
1104
+ "epoch": 0.796469701862017,
1105
+ "grad_norm": 0.6080626845359802,
1106
+ "learning_rate": 0.0005528197392522357,
1107
+ "loss": 3.9069,
1108
+ "step": 7400
1109
+ },
1110
+ {
1111
+ "epoch": 0.8018512539016253,
1112
+ "grad_norm": 0.6487952470779419,
1113
+ "learning_rate": 0.0005524964982221743,
1114
+ "loss": 3.9027,
1115
+ "step": 7450
1116
+ },
1117
+ {
1118
+ "epoch": 0.8072328059412335,
1119
+ "grad_norm": 0.5572729706764221,
1120
+ "learning_rate": 0.0005521732571921129,
1121
+ "loss": 3.8851,
1122
+ "step": 7500
1123
+ },
1124
+ {
1125
+ "epoch": 0.8126143579808417,
1126
+ "grad_norm": 0.6633656620979309,
1127
+ "learning_rate": 0.0005518500161620514,
1128
+ "loss": 3.9135,
1129
+ "step": 7550
1130
+ },
1131
+ {
1132
+ "epoch": 0.8179959100204499,
1133
+ "grad_norm": 0.5475841164588928,
1134
+ "learning_rate": 0.00055152677513199,
1135
+ "loss": 3.9074,
1136
+ "step": 7600
1137
+ },
1138
+ {
1139
+ "epoch": 0.8233774620600581,
1140
+ "grad_norm": 0.5891292691230774,
1141
+ "learning_rate": 0.0005512035341019286,
1142
+ "loss": 3.9187,
1143
+ "step": 7650
1144
+ },
1145
+ {
1146
+ "epoch": 0.8287590140996663,
1147
+ "grad_norm": 0.6392335891723633,
1148
+ "learning_rate": 0.0005508802930718672,
1149
+ "loss": 3.8935,
1150
+ "step": 7700
1151
+ },
1152
+ {
1153
+ "epoch": 0.8341405661392746,
1154
+ "grad_norm": 0.6276848316192627,
1155
+ "learning_rate": 0.0005505570520418058,
1156
+ "loss": 3.8951,
1157
+ "step": 7750
1158
+ },
1159
+ {
1160
+ "epoch": 0.8395221181788828,
1161
+ "grad_norm": 0.5669248700141907,
1162
+ "learning_rate": 0.0005502338110117443,
1163
+ "loss": 3.9057,
1164
+ "step": 7800
1165
+ },
1166
+ {
1167
+ "epoch": 0.844903670218491,
1168
+ "grad_norm": 0.666827380657196,
1169
+ "learning_rate": 0.000549910569981683,
1170
+ "loss": 3.8974,
1171
+ "step": 7850
1172
+ },
1173
+ {
1174
+ "epoch": 0.8502852222580992,
1175
+ "grad_norm": 0.6152955293655396,
1176
+ "learning_rate": 0.0005495873289516215,
1177
+ "loss": 3.9008,
1178
+ "step": 7900
1179
+ },
1180
+ {
1181
+ "epoch": 0.8556667742977074,
1182
+ "grad_norm": 0.5386107563972473,
1183
+ "learning_rate": 0.0005492640879215602,
1184
+ "loss": 3.8832,
1185
+ "step": 7950
1186
+ },
1187
+ {
1188
+ "epoch": 0.8610483263373157,
1189
+ "grad_norm": 0.553949236869812,
1190
+ "learning_rate": 0.0005489408468914987,
1191
+ "loss": 3.8799,
1192
+ "step": 8000
1193
+ },
1194
+ {
1195
+ "epoch": 0.8610483263373157,
1196
+ "eval_accuracy": 0.3376146192895166,
1197
+ "eval_loss": 3.817253589630127,
1198
+ "eval_runtime": 184.9597,
1199
+ "eval_samples_per_second": 97.378,
1200
+ "eval_steps_per_second": 6.088,
1201
+ "step": 8000
1202
+ },
1203
+ {
1204
+ "epoch": 0.8664298783769239,
1205
+ "grad_norm": 0.5128387212753296,
1206
+ "learning_rate": 0.0005486176058614372,
1207
+ "loss": 3.8821,
1208
+ "step": 8050
1209
+ },
1210
+ {
1211
+ "epoch": 0.8718114304165321,
1212
+ "grad_norm": 0.6132450699806213,
1213
+ "learning_rate": 0.0005482943648313759,
1214
+ "loss": 3.8667,
1215
+ "step": 8100
1216
+ },
1217
+ {
1218
+ "epoch": 0.8771929824561403,
1219
+ "grad_norm": 0.6007569432258606,
1220
+ "learning_rate": 0.0005479711238013145,
1221
+ "loss": 3.8853,
1222
+ "step": 8150
1223
+ },
1224
+ {
1225
+ "epoch": 0.8825745344957485,
1226
+ "grad_norm": 0.599909782409668,
1227
+ "learning_rate": 0.0005476478827712531,
1228
+ "loss": 3.871,
1229
+ "step": 8200
1230
+ },
1231
+ {
1232
+ "epoch": 0.8879560865353568,
1233
+ "grad_norm": 0.6086229681968689,
1234
+ "learning_rate": 0.0005473246417411916,
1235
+ "loss": 3.8779,
1236
+ "step": 8250
1237
+ },
1238
+ {
1239
+ "epoch": 0.8933376385749651,
1240
+ "grad_norm": 0.6416434645652771,
1241
+ "learning_rate": 0.0005470014007111302,
1242
+ "loss": 3.8791,
1243
+ "step": 8300
1244
+ },
1245
+ {
1246
+ "epoch": 0.8987191906145733,
1247
+ "grad_norm": 0.549599289894104,
1248
+ "learning_rate": 0.0005466781596810688,
1249
+ "loss": 3.8874,
1250
+ "step": 8350
1251
+ },
1252
+ {
1253
+ "epoch": 0.9041007426541815,
1254
+ "grad_norm": 0.5874208211898804,
1255
+ "learning_rate": 0.0005463549186510073,
1256
+ "loss": 3.8738,
1257
+ "step": 8400
1258
+ },
1259
+ {
1260
+ "epoch": 0.9094822946937897,
1261
+ "grad_norm": 0.5840559005737305,
1262
+ "learning_rate": 0.000546031677620946,
1263
+ "loss": 3.864,
1264
+ "step": 8450
1265
+ },
1266
+ {
1267
+ "epoch": 0.9148638467333979,
1268
+ "grad_norm": 0.5824464559555054,
1269
+ "learning_rate": 0.0005457084365908845,
1270
+ "loss": 3.8568,
1271
+ "step": 8500
1272
+ },
1273
+ {
1274
+ "epoch": 0.9202453987730062,
1275
+ "grad_norm": 0.6089414358139038,
1276
+ "learning_rate": 0.0005453851955608232,
1277
+ "loss": 3.8317,
1278
+ "step": 8550
1279
+ },
1280
+ {
1281
+ "epoch": 0.9256269508126144,
1282
+ "grad_norm": 0.5441460609436035,
1283
+ "learning_rate": 0.0005450619545307617,
1284
+ "loss": 3.8802,
1285
+ "step": 8600
1286
+ },
1287
+ {
1288
+ "epoch": 0.9310085028522226,
1289
+ "grad_norm": 0.5227726697921753,
1290
+ "learning_rate": 0.0005447387135007003,
1291
+ "loss": 3.8594,
1292
+ "step": 8650
1293
+ },
1294
+ {
1295
+ "epoch": 0.9363900548918308,
1296
+ "grad_norm": 0.5715673565864563,
1297
+ "learning_rate": 0.0005444154724706389,
1298
+ "loss": 3.8604,
1299
+ "step": 8700
1300
+ },
1301
+ {
1302
+ "epoch": 0.941771606931439,
1303
+ "grad_norm": 0.5205848813056946,
1304
+ "learning_rate": 0.0005440922314405775,
1305
+ "loss": 3.8593,
1306
+ "step": 8750
1307
+ },
1308
+ {
1309
+ "epoch": 0.9471531589710472,
1310
+ "grad_norm": 0.6547507047653198,
1311
+ "learning_rate": 0.0005437689904105161,
1312
+ "loss": 3.8436,
1313
+ "step": 8800
1314
+ },
1315
+ {
1316
+ "epoch": 0.9525347110106555,
1317
+ "grad_norm": 0.61313396692276,
1318
+ "learning_rate": 0.0005434457493804546,
1319
+ "loss": 3.848,
1320
+ "step": 8850
1321
+ },
1322
+ {
1323
+ "epoch": 0.9579162630502637,
1324
+ "grad_norm": 0.644686758518219,
1325
+ "learning_rate": 0.0005431225083503932,
1326
+ "loss": 3.844,
1327
+ "step": 8900
1328
+ },
1329
+ {
1330
+ "epoch": 0.9632978150898719,
1331
+ "grad_norm": 0.5130710005760193,
1332
+ "learning_rate": 0.0005427992673203318,
1333
+ "loss": 3.8736,
1334
+ "step": 8950
1335
+ },
1336
+ {
1337
+ "epoch": 0.9686793671294801,
1338
+ "grad_norm": 0.5622977018356323,
1339
+ "learning_rate": 0.0005424760262902704,
1340
+ "loss": 3.8502,
1341
+ "step": 9000
1342
+ },
1343
+ {
1344
+ "epoch": 0.9686793671294801,
1345
+ "eval_accuracy": 0.3408010825304519,
1346
+ "eval_loss": 3.7829928398132324,
1347
+ "eval_runtime": 184.3628,
1348
+ "eval_samples_per_second": 97.693,
1349
+ "eval_steps_per_second": 6.108,
1350
+ "step": 9000
1351
+ },
1352
+ {
1353
+ "epoch": 0.9740609191690883,
1354
+ "grad_norm": 0.5300456285476685,
1355
+ "learning_rate": 0.000542152785260209,
1356
+ "loss": 3.8365,
1357
+ "step": 9050
1358
+ },
1359
+ {
1360
+ "epoch": 0.9794424712086965,
1361
+ "grad_norm": 0.5774227976799011,
1362
+ "learning_rate": 0.0005418295442301476,
1363
+ "loss": 3.8359,
1364
+ "step": 9100
1365
+ },
1366
+ {
1367
+ "epoch": 0.9848240232483048,
1368
+ "grad_norm": 0.5969952344894409,
1369
+ "learning_rate": 0.0005415063032000861,
1370
+ "loss": 3.8564,
1371
+ "step": 9150
1372
+ },
1373
+ {
1374
+ "epoch": 0.9902055752879131,
1375
+ "grad_norm": 0.6162004470825195,
1376
+ "learning_rate": 0.0005411830621700248,
1377
+ "loss": 3.8361,
1378
+ "step": 9200
1379
+ },
1380
+ {
1381
+ "epoch": 0.9955871273275213,
1382
+ "grad_norm": 0.5422845482826233,
1383
+ "learning_rate": 0.0005408598211399633,
1384
+ "loss": 3.8151,
1385
+ "step": 9250
1386
+ },
1387
+ {
1388
+ "epoch": 1.0009686793671295,
1389
+ "grad_norm": 0.5742968916893005,
1390
+ "learning_rate": 0.0005405365801099019,
1391
+ "loss": 3.8236,
1392
+ "step": 9300
1393
+ },
1394
+ {
1395
+ "epoch": 1.0063502314067376,
1396
+ "grad_norm": 0.5724886655807495,
1397
+ "learning_rate": 0.0005402133390798405,
1398
+ "loss": 3.7718,
1399
+ "step": 9350
1400
+ },
1401
+ {
1402
+ "epoch": 1.011731783446346,
1403
+ "grad_norm": 0.6036879420280457,
1404
+ "learning_rate": 0.000539890098049779,
1405
+ "loss": 3.7636,
1406
+ "step": 9400
1407
+ },
1408
+ {
1409
+ "epoch": 1.017113335485954,
1410
+ "grad_norm": 0.5474029779434204,
1411
+ "learning_rate": 0.0005395668570197177,
1412
+ "loss": 3.7549,
1413
+ "step": 9450
1414
+ },
1415
+ {
1416
+ "epoch": 1.0224948875255624,
1417
+ "grad_norm": 0.5884007811546326,
1418
+ "learning_rate": 0.0005392436159896562,
1419
+ "loss": 3.7614,
1420
+ "step": 9500
1421
+ },
1422
+ {
1423
+ "epoch": 1.0278764395651705,
1424
+ "grad_norm": 0.5328190326690674,
1425
+ "learning_rate": 0.0005389203749595948,
1426
+ "loss": 3.7569,
1427
+ "step": 9550
1428
+ },
1429
+ {
1430
+ "epoch": 1.0332579916047788,
1431
+ "grad_norm": 0.5494911670684814,
1432
+ "learning_rate": 0.0005385971339295334,
1433
+ "loss": 3.7681,
1434
+ "step": 9600
1435
+ },
1436
+ {
1437
+ "epoch": 1.0386395436443872,
1438
+ "grad_norm": 0.6018995046615601,
1439
+ "learning_rate": 0.000538273892899472,
1440
+ "loss": 3.7921,
1441
+ "step": 9650
1442
+ },
1443
+ {
1444
+ "epoch": 1.0440210956839953,
1445
+ "grad_norm": 0.6208644509315491,
1446
+ "learning_rate": 0.0005379506518694106,
1447
+ "loss": 3.7763,
1448
+ "step": 9700
1449
+ },
1450
+ {
1451
+ "epoch": 1.0494026477236036,
1452
+ "grad_norm": 0.6352260112762451,
1453
+ "learning_rate": 0.0005376274108393491,
1454
+ "loss": 3.7876,
1455
+ "step": 9750
1456
+ },
1457
+ {
1458
+ "epoch": 1.0547841997632117,
1459
+ "grad_norm": 0.565929114818573,
1460
+ "learning_rate": 0.0005373041698092877,
1461
+ "loss": 3.7864,
1462
+ "step": 9800
1463
+ },
1464
+ {
1465
+ "epoch": 1.06016575180282,
1466
+ "grad_norm": 0.5776049494743347,
1467
+ "learning_rate": 0.0005369809287792263,
1468
+ "loss": 3.7535,
1469
+ "step": 9850
1470
+ },
1471
+ {
1472
+ "epoch": 1.0655473038424281,
1473
+ "grad_norm": 0.5360944867134094,
1474
+ "learning_rate": 0.000536657687749165,
1475
+ "loss": 3.7739,
1476
+ "step": 9900
1477
+ },
1478
+ {
1479
+ "epoch": 1.0709288558820365,
1480
+ "grad_norm": 0.5402177572250366,
1481
+ "learning_rate": 0.0005363344467191035,
1482
+ "loss": 3.7569,
1483
+ "step": 9950
1484
+ },
1485
+ {
1486
+ "epoch": 1.0763104079216446,
1487
+ "grad_norm": 0.5348372459411621,
1488
+ "learning_rate": 0.000536011205689042,
1489
+ "loss": 3.7597,
1490
+ "step": 10000
1491
+ },
1492
+ {
1493
+ "epoch": 1.0763104079216446,
1494
+ "eval_accuracy": 0.3438087031180445,
1495
+ "eval_loss": 3.7523183822631836,
1496
+ "eval_runtime": 185.331,
1497
+ "eval_samples_per_second": 97.183,
1498
+ "eval_steps_per_second": 6.076,
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": 8.3609792151552e+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 sha256:097b2f4878fa458807831f34b125c9126dbb86ed90dd7b8414ba94f45393ff0f
3
+ size 5304