chansung commited on
Commit
b26371f
1 Parent(s): 6e13180

Model save

Browse files
README.md ADDED
@@ -0,0 +1,85 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: google/gemma-7b
3
+ datasets:
4
+ - generator
5
+ library_name: peft
6
+ license: gemma
7
+ tags:
8
+ - trl
9
+ - sft
10
+ - generated_from_trainer
11
+ model-index:
12
+ - name: gemma7b-gpt4o_1k_summarize-ksaslora-wo-auxloss
13
+ results: []
14
+ ---
15
+
16
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
+ should probably proofread and complete it, then remove this comment. -->
18
+
19
+ # gemma7b-gpt4o_1k_summarize-ksaslora-wo-auxloss
20
+
21
+ This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the generator dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 2.4203
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+
40
+ The following `bitsandbytes` quantization config was used during training:
41
+ - quant_method: QuantizationMethod.BITS_AND_BYTES
42
+ - _load_in_8bit: True
43
+ - _load_in_4bit: False
44
+ - llm_int8_threshold: 6.0
45
+ - llm_int8_skip_modules: None
46
+ - llm_int8_enable_fp32_cpu_offload: False
47
+ - llm_int8_has_fp16_weight: False
48
+ - bnb_4bit_quant_type: fp4
49
+ - bnb_4bit_use_double_quant: False
50
+ - bnb_4bit_compute_dtype: float32
51
+ - bnb_4bit_quant_storage: uint8
52
+ - load_in_4bit: False
53
+ - load_in_8bit: True
54
+
55
+ ### Training hyperparameters
56
+
57
+ The following hyperparameters were used during training:
58
+ - learning_rate: 0.0003
59
+ - train_batch_size: 1
60
+ - eval_batch_size: 2
61
+ - seed: 42
62
+ - distributed_type: multi-GPU
63
+ - num_devices: 6
64
+ - gradient_accumulation_steps: 2
65
+ - total_train_batch_size: 12
66
+ - total_eval_batch_size: 12
67
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
68
+ - lr_scheduler_type: cosine
69
+ - lr_scheduler_warmup_ratio: 0.1
70
+ - num_epochs: 1
71
+
72
+ ### Training results
73
+
74
+ | Training Loss | Epoch | Step | Validation Loss |
75
+ |:-------------:|:-----:|:----:|:---------------:|
76
+ | 0.8382 | 1.0 | 1045 | 2.4203 |
77
+
78
+
79
+ ### Framework versions
80
+
81
+ - PEFT 0.6.3.dev0
82
+ - Transformers 4.45.0
83
+ - Pytorch 2.4.1+cu121
84
+ - Datasets 3.0.0
85
+ - Tokenizers 0.20.0
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 6.049215074407547e+17,
4
+ "train_loss": 1.5597172723432478,
5
+ "train_runtime": 7161.753,
6
+ "train_samples": 115376,
7
+ "train_samples_per_second": 1.751,
8
+ "train_steps_per_second": 0.146
9
+ }
runs/Sep25_22-15-13_user-HP-Z8-Fury-G5-Workstation-Desktop-PC/events.out.tfevents.1727317615.user-HP-Z8-Fury-G5-Workstation-Desktop-PC.316.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:adc2790756458a1c7d48918e2f00cd97cec9c078c64d62b4fcd2fc2f5d83cb46
3
- size 50989
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a69769da7ffd070d5fcc772be805ac3040e3f7cfb44f8b63012f2d6d9c3f9a1
3
+ size 51614
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 6.049215074407547e+17,
4
+ "train_loss": 1.5597172723432478,
5
+ "train_runtime": 7161.753,
6
+ "train_samples": 115376,
7
+ "train_samples_per_second": 1.751,
8
+ "train_steps_per_second": 0.146
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,1520 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 1045,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0009569377990430622,
13
+ "grad_norm": 948.5667114257812,
14
+ "learning_rate": 2.8571428571428573e-06,
15
+ "loss": 46.4048,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.004784688995215311,
20
+ "grad_norm": 579.7166137695312,
21
+ "learning_rate": 1.4285714285714284e-05,
22
+ "loss": 42.2625,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.009569377990430622,
27
+ "grad_norm": 96.71955871582031,
28
+ "learning_rate": 2.8571428571428567e-05,
29
+ "loss": 24.8784,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.014354066985645933,
34
+ "grad_norm": 14.768364906311035,
35
+ "learning_rate": 4.285714285714285e-05,
36
+ "loss": 20.4002,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.019138755980861243,
41
+ "grad_norm": 12.786796569824219,
42
+ "learning_rate": 5.7142857142857135e-05,
43
+ "loss": 18.6753,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.023923444976076555,
48
+ "grad_norm": 46.72217559814453,
49
+ "learning_rate": 7.142857142857142e-05,
50
+ "loss": 15.9881,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.028708133971291867,
55
+ "grad_norm": 16.096284866333008,
56
+ "learning_rate": 8.57142857142857e-05,
57
+ "loss": 7.4899,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.03349282296650718,
62
+ "grad_norm": 6.774743556976318,
63
+ "learning_rate": 9.999999999999999e-05,
64
+ "loss": 2.28,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.03827751196172249,
69
+ "grad_norm": 25.32003402709961,
70
+ "learning_rate": 0.00011428571428571427,
71
+ "loss": 1.816,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.0430622009569378,
76
+ "grad_norm": 4.912317276000977,
77
+ "learning_rate": 0.00012857142857142855,
78
+ "loss": 1.6463,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.04784688995215311,
83
+ "grad_norm": 3.605388879776001,
84
+ "learning_rate": 0.00014285714285714284,
85
+ "loss": 1.5122,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.05263157894736842,
90
+ "grad_norm": 5.295993328094482,
91
+ "learning_rate": 0.00015714285714285713,
92
+ "loss": 1.419,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.05741626794258373,
97
+ "grad_norm": 9.990426063537598,
98
+ "learning_rate": 0.0001714285714285714,
99
+ "loss": 1.3221,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.06220095693779904,
104
+ "grad_norm": 28.818571090698242,
105
+ "learning_rate": 0.00018571428571428572,
106
+ "loss": 1.2352,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.06698564593301436,
111
+ "grad_norm": 5.248025417327881,
112
+ "learning_rate": 0.00019999999999999998,
113
+ "loss": 1.2013,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.07177033492822966,
118
+ "grad_norm": 4.098865509033203,
119
+ "learning_rate": 0.00021428571428571427,
120
+ "loss": 1.1641,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.07655502392344497,
125
+ "grad_norm": 1.8723022937774658,
126
+ "learning_rate": 0.00022857142857142854,
127
+ "loss": 1.1404,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.08133971291866028,
132
+ "grad_norm": 20.416040420532227,
133
+ "learning_rate": 0.00024285714285714283,
134
+ "loss": 1.1758,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.0861244019138756,
139
+ "grad_norm": 2.548046112060547,
140
+ "learning_rate": 0.0002571428571428571,
141
+ "loss": 1.1864,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.09090909090909091,
146
+ "grad_norm": 10.036776542663574,
147
+ "learning_rate": 0.0002714285714285714,
148
+ "loss": 1.1193,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.09569377990430622,
153
+ "grad_norm": 7.473161697387695,
154
+ "learning_rate": 0.0002857142857142857,
155
+ "loss": 1.1753,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.10047846889952153,
160
+ "grad_norm": 1.7201554775238037,
161
+ "learning_rate": 0.0003,
162
+ "loss": 1.1511,
163
+ "step": 105
164
+ },
165
+ {
166
+ "epoch": 0.10526315789473684,
167
+ "grad_norm": 22.742843627929688,
168
+ "learning_rate": 0.00029997905717787856,
169
+ "loss": 1.1194,
170
+ "step": 110
171
+ },
172
+ {
173
+ "epoch": 0.11004784688995216,
174
+ "grad_norm": 4.5689191818237305,
175
+ "learning_rate": 0.00029991623455953814,
176
+ "loss": 1.1401,
177
+ "step": 115
178
+ },
179
+ {
180
+ "epoch": 0.11483253588516747,
181
+ "grad_norm": 4.430986404418945,
182
+ "learning_rate": 0.00029981154968741785,
183
+ "loss": 1.1312,
184
+ "step": 120
185
+ },
186
+ {
187
+ "epoch": 0.11961722488038277,
188
+ "grad_norm": 9.391127586364746,
189
+ "learning_rate": 0.000299665031793473,
190
+ "loss": 1.0676,
191
+ "step": 125
192
+ },
193
+ {
194
+ "epoch": 0.12440191387559808,
195
+ "grad_norm": 10.183479309082031,
196
+ "learning_rate": 0.0002994767217910127,
197
+ "loss": 1.1184,
198
+ "step": 130
199
+ },
200
+ {
201
+ "epoch": 0.1291866028708134,
202
+ "grad_norm": 1.511924147605896,
203
+ "learning_rate": 0.00029924667226327557,
204
+ "loss": 1.0999,
205
+ "step": 135
206
+ },
207
+ {
208
+ "epoch": 0.1339712918660287,
209
+ "grad_norm": 8.494267463684082,
210
+ "learning_rate": 0.0002989749474487461,
211
+ "loss": 1.0913,
212
+ "step": 140
213
+ },
214
+ {
215
+ "epoch": 0.13875598086124402,
216
+ "grad_norm": 2.4349634647369385,
217
+ "learning_rate": 0.000298661623223217,
218
+ "loss": 1.0802,
219
+ "step": 145
220
+ },
221
+ {
222
+ "epoch": 0.14354066985645933,
223
+ "grad_norm": 1.914871096611023,
224
+ "learning_rate": 0.0002983067870786019,
225
+ "loss": 1.0589,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.14832535885167464,
230
+ "grad_norm": 8.084524154663086,
231
+ "learning_rate": 0.00029791053809850426,
232
+ "loss": 1.0661,
233
+ "step": 155
234
+ },
235
+ {
236
+ "epoch": 0.15311004784688995,
237
+ "grad_norm": 15.332684516906738,
238
+ "learning_rate": 0.0002974729869305495,
239
+ "loss": 1.0688,
240
+ "step": 160
241
+ },
242
+ {
243
+ "epoch": 0.15789473684210525,
244
+ "grad_norm": 30.3629207611084,
245
+ "learning_rate": 0.000296994255755488,
246
+ "loss": 1.0531,
247
+ "step": 165
248
+ },
249
+ {
250
+ "epoch": 0.16267942583732056,
251
+ "grad_norm": 6.318498611450195,
252
+ "learning_rate": 0.0002964744782530777,
253
+ "loss": 1.0664,
254
+ "step": 170
255
+ },
256
+ {
257
+ "epoch": 0.1674641148325359,
258
+ "grad_norm": 7.628419399261475,
259
+ "learning_rate": 0.0002959137995647556,
260
+ "loss": 1.1012,
261
+ "step": 175
262
+ },
263
+ {
264
+ "epoch": 0.1722488038277512,
265
+ "grad_norm": 6.398703098297119,
266
+ "learning_rate": 0.0002953123762531088,
267
+ "loss": 1.1012,
268
+ "step": 180
269
+ },
270
+ {
271
+ "epoch": 0.17703349282296652,
272
+ "grad_norm": 1.5758147239685059,
273
+ "learning_rate": 0.00029467037625815644,
274
+ "loss": 1.0655,
275
+ "step": 185
276
+ },
277
+ {
278
+ "epoch": 0.18181818181818182,
279
+ "grad_norm": 3.354144811630249,
280
+ "learning_rate": 0.0002939879788504546,
281
+ "loss": 1.0405,
282
+ "step": 190
283
+ },
284
+ {
285
+ "epoch": 0.18660287081339713,
286
+ "grad_norm": 7.777205467224121,
287
+ "learning_rate": 0.00029326537458103683,
288
+ "loss": 1.0575,
289
+ "step": 195
290
+ },
291
+ {
292
+ "epoch": 0.19138755980861244,
293
+ "grad_norm": 3.233966588973999,
294
+ "learning_rate": 0.0002925027652282056,
295
+ "loss": 1.0519,
296
+ "step": 200
297
+ },
298
+ {
299
+ "epoch": 0.19617224880382775,
300
+ "grad_norm": 5.466601848602295,
301
+ "learning_rate": 0.00029170036374118777,
302
+ "loss": 1.0377,
303
+ "step": 205
304
+ },
305
+ {
306
+ "epoch": 0.20095693779904306,
307
+ "grad_norm": 2.6383893489837646,
308
+ "learning_rate": 0.0002908583941806717,
309
+ "loss": 1.0435,
310
+ "step": 210
311
+ },
312
+ {
313
+ "epoch": 0.20574162679425836,
314
+ "grad_norm": 1.6238977909088135,
315
+ "learning_rate": 0.00028997709165624034,
316
+ "loss": 1.0098,
317
+ "step": 215
318
+ },
319
+ {
320
+ "epoch": 0.21052631578947367,
321
+ "grad_norm": 4.616154193878174,
322
+ "learning_rate": 0.0002890567022607206,
323
+ "loss": 1.007,
324
+ "step": 220
325
+ },
326
+ {
327
+ "epoch": 0.215311004784689,
328
+ "grad_norm": 3.3014893531799316,
329
+ "learning_rate": 0.0002880974830014643,
330
+ "loss": 1.0311,
331
+ "step": 225
332
+ },
333
+ {
334
+ "epoch": 0.22009569377990432,
335
+ "grad_norm": 1.7747435569763184,
336
+ "learning_rate": 0.0002870997017285824,
337
+ "loss": 0.9898,
338
+ "step": 230
339
+ },
340
+ {
341
+ "epoch": 0.22488038277511962,
342
+ "grad_norm": 2.502115488052368,
343
+ "learning_rate": 0.0002860636370601511,
344
+ "loss": 1.0031,
345
+ "step": 235
346
+ },
347
+ {
348
+ "epoch": 0.22966507177033493,
349
+ "grad_norm": 2.33097767829895,
350
+ "learning_rate": 0.00028498957830441117,
351
+ "loss": 1.0175,
352
+ "step": 240
353
+ },
354
+ {
355
+ "epoch": 0.23444976076555024,
356
+ "grad_norm": 2.667079448699951,
357
+ "learning_rate": 0.00028387782537898215,
358
+ "loss": 1.0102,
359
+ "step": 245
360
+ },
361
+ {
362
+ "epoch": 0.23923444976076555,
363
+ "grad_norm": 5.493870735168457,
364
+ "learning_rate": 0.0002827286887271143,
365
+ "loss": 0.9815,
366
+ "step": 250
367
+ },
368
+ {
369
+ "epoch": 0.24401913875598086,
370
+ "grad_norm": 1.950860619544983,
371
+ "learning_rate": 0.0002815424892310007,
372
+ "loss": 0.9865,
373
+ "step": 255
374
+ },
375
+ {
376
+ "epoch": 0.24880382775119617,
377
+ "grad_norm": 1.8500947952270508,
378
+ "learning_rate": 0.00028031955812217544,
379
+ "loss": 0.9819,
380
+ "step": 260
381
+ },
382
+ {
383
+ "epoch": 0.2535885167464115,
384
+ "grad_norm": 4.860908508300781,
385
+ "learning_rate": 0.0002790602368890209,
386
+ "loss": 0.9728,
387
+ "step": 265
388
+ },
389
+ {
390
+ "epoch": 0.2583732057416268,
391
+ "grad_norm": 2.0686490535736084,
392
+ "learning_rate": 0.0002777648771814114,
393
+ "loss": 1.02,
394
+ "step": 270
395
+ },
396
+ {
397
+ "epoch": 0.2631578947368421,
398
+ "grad_norm": 3.445263624191284,
399
+ "learning_rate": 0.00027643384071251954,
400
+ "loss": 1.0279,
401
+ "step": 275
402
+ },
403
+ {
404
+ "epoch": 0.2679425837320574,
405
+ "grad_norm": 1.4115147590637207,
406
+ "learning_rate": 0.000275067499157812,
407
+ "loss": 0.9813,
408
+ "step": 280
409
+ },
410
+ {
411
+ "epoch": 0.2727272727272727,
412
+ "grad_norm": 1.2224127054214478,
413
+ "learning_rate": 0.00027366623405126404,
414
+ "loss": 0.9903,
415
+ "step": 285
416
+ },
417
+ {
418
+ "epoch": 0.27751196172248804,
419
+ "grad_norm": 1.8986269235610962,
420
+ "learning_rate": 0.0002722304366788205,
421
+ "loss": 1.0094,
422
+ "step": 290
423
+ },
424
+ {
425
+ "epoch": 0.2822966507177033,
426
+ "grad_norm": 4.729682445526123,
427
+ "learning_rate": 0.0002707605079691344,
428
+ "loss": 1.0046,
429
+ "step": 295
430
+ },
431
+ {
432
+ "epoch": 0.28708133971291866,
433
+ "grad_norm": 2.3022851943969727,
434
+ "learning_rate": 0.0002692568583816124,
435
+ "loss": 1.0305,
436
+ "step": 300
437
+ },
438
+ {
439
+ "epoch": 0.291866028708134,
440
+ "grad_norm": 2.677311658859253,
441
+ "learning_rate": 0.0002677199077917991,
442
+ "loss": 1.0308,
443
+ "step": 305
444
+ },
445
+ {
446
+ "epoch": 0.2966507177033493,
447
+ "grad_norm": 1.3505252599716187,
448
+ "learning_rate": 0.000266150085374132,
449
+ "loss": 0.9911,
450
+ "step": 310
451
+ },
452
+ {
453
+ "epoch": 0.3014354066985646,
454
+ "grad_norm": 1.311340093612671,
455
+ "learning_rate": 0.00026454782948209983,
456
+ "loss": 0.9897,
457
+ "step": 315
458
+ },
459
+ {
460
+ "epoch": 0.3062200956937799,
461
+ "grad_norm": 1.3043484687805176,
462
+ "learning_rate": 0.00026291358752583764,
463
+ "loss": 0.9659,
464
+ "step": 320
465
+ },
466
+ {
467
+ "epoch": 0.31100478468899523,
468
+ "grad_norm": 1.1022688150405884,
469
+ "learning_rate": 0.0002612478158471936,
470
+ "loss": 0.9852,
471
+ "step": 325
472
+ },
473
+ {
474
+ "epoch": 0.3157894736842105,
475
+ "grad_norm": 1.8880759477615356,
476
+ "learning_rate": 0.0002595509795923004,
477
+ "loss": 0.9707,
478
+ "step": 330
479
+ },
480
+ {
481
+ "epoch": 0.32057416267942584,
482
+ "grad_norm": 4.946929454803467,
483
+ "learning_rate": 0.0002578235525816894,
484
+ "loss": 1.0037,
485
+ "step": 335
486
+ },
487
+ {
488
+ "epoch": 0.3253588516746411,
489
+ "grad_norm": 1.5874648094177246,
490
+ "learning_rate": 0.00025606601717798207,
491
+ "loss": 1.013,
492
+ "step": 340
493
+ },
494
+ {
495
+ "epoch": 0.33014354066985646,
496
+ "grad_norm": 1.1188175678253174,
497
+ "learning_rate": 0.0002542788641511963,
498
+ "loss": 0.9814,
499
+ "step": 345
500
+ },
501
+ {
502
+ "epoch": 0.3349282296650718,
503
+ "grad_norm": 1.3674702644348145,
504
+ "learning_rate": 0.0002524625925417046,
505
+ "loss": 0.9924,
506
+ "step": 350
507
+ },
508
+ {
509
+ "epoch": 0.3397129186602871,
510
+ "grad_norm": 1.3748054504394531,
511
+ "learning_rate": 0.0002506177095208835,
512
+ "loss": 0.9952,
513
+ "step": 355
514
+ },
515
+ {
516
+ "epoch": 0.3444976076555024,
517
+ "grad_norm": 1.4951032400131226,
518
+ "learning_rate": 0.00024874473024949224,
519
+ "loss": 0.9859,
520
+ "step": 360
521
+ },
522
+ {
523
+ "epoch": 0.3492822966507177,
524
+ "grad_norm": 1.5338605642318726,
525
+ "learning_rate": 0.0002468441777338203,
526
+ "loss": 0.9986,
527
+ "step": 365
528
+ },
529
+ {
530
+ "epoch": 0.35406698564593303,
531
+ "grad_norm": 2.292693614959717,
532
+ "learning_rate": 0.00024491658267964474,
533
+ "loss": 0.9815,
534
+ "step": 370
535
+ },
536
+ {
537
+ "epoch": 0.3588516746411483,
538
+ "grad_norm": 3.6766607761383057,
539
+ "learning_rate": 0.00024296248334403672,
540
+ "loss": 0.9696,
541
+ "step": 375
542
+ },
543
+ {
544
+ "epoch": 0.36363636363636365,
545
+ "grad_norm": 1.8351837396621704,
546
+ "learning_rate": 0.00024098242538506007,
547
+ "loss": 0.9806,
548
+ "step": 380
549
+ },
550
+ {
551
+ "epoch": 0.3684210526315789,
552
+ "grad_norm": 3.7850887775421143,
553
+ "learning_rate": 0.00023897696170940326,
554
+ "loss": 0.9922,
555
+ "step": 385
556
+ },
557
+ {
558
+ "epoch": 0.37320574162679426,
559
+ "grad_norm": 1.6278636455535889,
560
+ "learning_rate": 0.0002369466523179866,
561
+ "loss": 0.9524,
562
+ "step": 390
563
+ },
564
+ {
565
+ "epoch": 0.37799043062200954,
566
+ "grad_norm": 2.634073495864868,
567
+ "learning_rate": 0.0002348920641495893,
568
+ "loss": 0.9721,
569
+ "step": 395
570
+ },
571
+ {
572
+ "epoch": 0.3827751196172249,
573
+ "grad_norm": 1.0519402027130127,
574
+ "learning_rate": 0.0002328137709225385,
575
+ "loss": 0.9433,
576
+ "step": 400
577
+ },
578
+ {
579
+ "epoch": 0.3875598086124402,
580
+ "grad_norm": 1.6269092559814453,
581
+ "learning_rate": 0.00023071235297450588,
582
+ "loss": 0.965,
583
+ "step": 405
584
+ },
585
+ {
586
+ "epoch": 0.3923444976076555,
587
+ "grad_norm": 1.0608420372009277,
588
+ "learning_rate": 0.0002285883971004553,
589
+ "loss": 0.9458,
590
+ "step": 410
591
+ },
592
+ {
593
+ "epoch": 0.39712918660287083,
594
+ "grad_norm": 1.1753120422363281,
595
+ "learning_rate": 0.00022644249638878762,
596
+ "loss": 0.9708,
597
+ "step": 415
598
+ },
599
+ {
600
+ "epoch": 0.4019138755980861,
601
+ "grad_norm": 1.1925132274627686,
602
+ "learning_rate": 0.000224275250055728,
603
+ "loss": 0.9419,
604
+ "step": 420
605
+ },
606
+ {
607
+ "epoch": 0.40669856459330145,
608
+ "grad_norm": 2.7282474040985107,
609
+ "learning_rate": 0.00022208726327800255,
610
+ "loss": 0.9776,
611
+ "step": 425
612
+ },
613
+ {
614
+ "epoch": 0.41148325358851673,
615
+ "grad_norm": 1.2128546237945557,
616
+ "learning_rate": 0.0002198791470238497,
617
+ "loss": 0.9726,
618
+ "step": 430
619
+ },
620
+ {
621
+ "epoch": 0.41626794258373206,
622
+ "grad_norm": 1.216680884361267,
623
+ "learning_rate": 0.00021765151788241525,
624
+ "loss": 0.965,
625
+ "step": 435
626
+ },
627
+ {
628
+ "epoch": 0.42105263157894735,
629
+ "grad_norm": 1.2095612287521362,
630
+ "learning_rate": 0.0002154049978915774,
631
+ "loss": 0.9734,
632
+ "step": 440
633
+ },
634
+ {
635
+ "epoch": 0.4258373205741627,
636
+ "grad_norm": 2.2863378524780273,
637
+ "learning_rate": 0.00021314021436425024,
638
+ "loss": 0.9668,
639
+ "step": 445
640
+ },
641
+ {
642
+ "epoch": 0.430622009569378,
643
+ "grad_norm": 1.6475805044174194,
644
+ "learning_rate": 0.00021085779971321454,
645
+ "loss": 0.953,
646
+ "step": 450
647
+ },
648
+ {
649
+ "epoch": 0.4354066985645933,
650
+ "grad_norm": 2.581031560897827,
651
+ "learning_rate": 0.0002085583912745242,
652
+ "loss": 0.9481,
653
+ "step": 455
654
+ },
655
+ {
656
+ "epoch": 0.44019138755980863,
657
+ "grad_norm": 1.6886945962905884,
658
+ "learning_rate": 0.00020624263112953815,
659
+ "loss": 0.9521,
660
+ "step": 460
661
+ },
662
+ {
663
+ "epoch": 0.4449760765550239,
664
+ "grad_norm": 1.6784495115280151,
665
+ "learning_rate": 0.0002039111659256269,
666
+ "loss": 0.9448,
667
+ "step": 465
668
+ },
669
+ {
670
+ "epoch": 0.44976076555023925,
671
+ "grad_norm": 1.741603970527649,
672
+ "learning_rate": 0.00020156464669560448,
673
+ "loss": 0.9396,
674
+ "step": 470
675
+ },
676
+ {
677
+ "epoch": 0.45454545454545453,
678
+ "grad_norm": 1.4577537775039673,
679
+ "learning_rate": 0.00019920372867593533,
680
+ "loss": 0.9381,
681
+ "step": 475
682
+ },
683
+ {
684
+ "epoch": 0.45933014354066987,
685
+ "grad_norm": 1.2143363952636719,
686
+ "learning_rate": 0.00019682907112376796,
687
+ "loss": 0.942,
688
+ "step": 480
689
+ },
690
+ {
691
+ "epoch": 0.46411483253588515,
692
+ "grad_norm": 1.133268117904663,
693
+ "learning_rate": 0.0001944413371328451,
694
+ "loss": 0.9307,
695
+ "step": 485
696
+ },
697
+ {
698
+ "epoch": 0.4688995215311005,
699
+ "grad_norm": 1.3096511363983154,
700
+ "learning_rate": 0.0001920411934483434,
701
+ "loss": 0.9489,
702
+ "step": 490
703
+ },
704
+ {
705
+ "epoch": 0.47368421052631576,
706
+ "grad_norm": 1.1326684951782227,
707
+ "learning_rate": 0.00018962931028069293,
708
+ "loss": 0.9481,
709
+ "step": 495
710
+ },
711
+ {
712
+ "epoch": 0.4784688995215311,
713
+ "grad_norm": 1.435521125793457,
714
+ "learning_rate": 0.00018720636111842898,
715
+ "loss": 0.941,
716
+ "step": 500
717
+ },
718
+ {
719
+ "epoch": 0.48325358851674644,
720
+ "grad_norm": 1.5249536037445068,
721
+ "learning_rate": 0.00018477302254012924,
722
+ "loss": 0.9466,
723
+ "step": 505
724
+ },
725
+ {
726
+ "epoch": 0.4880382775119617,
727
+ "grad_norm": 1.5078822374343872,
728
+ "learning_rate": 0.000182329974025487,
729
+ "loss": 0.9407,
730
+ "step": 510
731
+ },
732
+ {
733
+ "epoch": 0.49282296650717705,
734
+ "grad_norm": 1.1270455121994019,
735
+ "learning_rate": 0.0001798778977655754,
736
+ "loss": 0.9342,
737
+ "step": 515
738
+ },
739
+ {
740
+ "epoch": 0.49760765550239233,
741
+ "grad_norm": 1.8892922401428223,
742
+ "learning_rate": 0.00017741747847235353,
743
+ "loss": 0.9215,
744
+ "step": 520
745
+ },
746
+ {
747
+ "epoch": 0.5023923444976076,
748
+ "grad_norm": 1.1376285552978516,
749
+ "learning_rate": 0.0001749494031874695,
750
+ "loss": 0.941,
751
+ "step": 525
752
+ },
753
+ {
754
+ "epoch": 0.507177033492823,
755
+ "grad_norm": 1.3463551998138428,
756
+ "learning_rate": 0.0001724743610904122,
757
+ "loss": 0.9494,
758
+ "step": 530
759
+ },
760
+ {
761
+ "epoch": 0.5119617224880383,
762
+ "grad_norm": 2.9610071182250977,
763
+ "learning_rate": 0.00016999304330606683,
764
+ "loss": 0.9342,
765
+ "step": 535
766
+ },
767
+ {
768
+ "epoch": 0.5167464114832536,
769
+ "grad_norm": 1.1553057432174683,
770
+ "learning_rate": 0.00016750614271172643,
771
+ "loss": 0.9421,
772
+ "step": 540
773
+ },
774
+ {
775
+ "epoch": 0.5215311004784688,
776
+ "grad_norm": 1.025996446609497,
777
+ "learning_rate": 0.00016501435374361475,
778
+ "loss": 0.9165,
779
+ "step": 545
780
+ },
781
+ {
782
+ "epoch": 0.5263157894736842,
783
+ "grad_norm": 1.1247053146362305,
784
+ "learning_rate": 0.00016251837220297347,
785
+ "loss": 0.9382,
786
+ "step": 550
787
+ },
788
+ {
789
+ "epoch": 0.5311004784688995,
790
+ "grad_norm": 1.027742624282837,
791
+ "learning_rate": 0.00016001889506176824,
792
+ "loss": 0.9461,
793
+ "step": 555
794
+ },
795
+ {
796
+ "epoch": 0.5358851674641149,
797
+ "grad_norm": 1.2423522472381592,
798
+ "learning_rate": 0.00015751662026806788,
799
+ "loss": 0.9411,
800
+ "step": 560
801
+ },
802
+ {
803
+ "epoch": 0.5406698564593302,
804
+ "grad_norm": 1.5147805213928223,
805
+ "learning_rate": 0.00015501224655115118,
806
+ "loss": 0.9334,
807
+ "step": 565
808
+ },
809
+ {
810
+ "epoch": 0.5454545454545454,
811
+ "grad_norm": 1.2580386400222778,
812
+ "learning_rate": 0.00015250647322639512,
813
+ "loss": 0.9258,
814
+ "step": 570
815
+ },
816
+ {
817
+ "epoch": 0.5502392344497608,
818
+ "grad_norm": 1.1148898601531982,
819
+ "learning_rate": 0.00015,
820
+ "loss": 0.92,
821
+ "step": 575
822
+ },
823
+ {
824
+ "epoch": 0.5550239234449761,
825
+ "grad_norm": 1.6557159423828125,
826
+ "learning_rate": 0.00014749352677360483,
827
+ "loss": 0.9124,
828
+ "step": 580
829
+ },
830
+ {
831
+ "epoch": 0.5598086124401914,
832
+ "grad_norm": 1.3142054080963135,
833
+ "learning_rate": 0.00014498775344884884,
834
+ "loss": 0.9292,
835
+ "step": 585
836
+ },
837
+ {
838
+ "epoch": 0.5645933014354066,
839
+ "grad_norm": 1.3109759092330933,
840
+ "learning_rate": 0.00014248337973193212,
841
+ "loss": 0.9075,
842
+ "step": 590
843
+ },
844
+ {
845
+ "epoch": 0.569377990430622,
846
+ "grad_norm": 1.8519434928894043,
847
+ "learning_rate": 0.00013998110493823176,
848
+ "loss": 0.9334,
849
+ "step": 595
850
+ },
851
+ {
852
+ "epoch": 0.5741626794258373,
853
+ "grad_norm": 1.467532992362976,
854
+ "learning_rate": 0.00013748162779702653,
855
+ "loss": 0.9038,
856
+ "step": 600
857
+ },
858
+ {
859
+ "epoch": 0.5789473684210527,
860
+ "grad_norm": 1.2827484607696533,
861
+ "learning_rate": 0.00013498564625638522,
862
+ "loss": 0.9209,
863
+ "step": 605
864
+ },
865
+ {
866
+ "epoch": 0.583732057416268,
867
+ "grad_norm": 2.0187387466430664,
868
+ "learning_rate": 0.00013249385728827357,
869
+ "loss": 0.909,
870
+ "step": 610
871
+ },
872
+ {
873
+ "epoch": 0.5885167464114832,
874
+ "grad_norm": 1.0019465684890747,
875
+ "learning_rate": 0.0001300069566939332,
876
+ "loss": 0.9324,
877
+ "step": 615
878
+ },
879
+ {
880
+ "epoch": 0.5933014354066986,
881
+ "grad_norm": 1.1702934503555298,
882
+ "learning_rate": 0.00012752563890958778,
883
+ "loss": 0.8818,
884
+ "step": 620
885
+ },
886
+ {
887
+ "epoch": 0.5980861244019139,
888
+ "grad_norm": 1.1036059856414795,
889
+ "learning_rate": 0.0001250505968125305,
890
+ "loss": 0.9082,
891
+ "step": 625
892
+ },
893
+ {
894
+ "epoch": 0.6028708133971292,
895
+ "grad_norm": 1.3649473190307617,
896
+ "learning_rate": 0.00012258252152764647,
897
+ "loss": 0.9077,
898
+ "step": 630
899
+ },
900
+ {
901
+ "epoch": 0.6076555023923444,
902
+ "grad_norm": 1.2819594144821167,
903
+ "learning_rate": 0.0001201221022344246,
904
+ "loss": 0.899,
905
+ "step": 635
906
+ },
907
+ {
908
+ "epoch": 0.6124401913875598,
909
+ "grad_norm": 1.1259446144104004,
910
+ "learning_rate": 0.00011767002597451296,
911
+ "loss": 0.8888,
912
+ "step": 640
913
+ },
914
+ {
915
+ "epoch": 0.6172248803827751,
916
+ "grad_norm": 1.0908528566360474,
917
+ "learning_rate": 0.00011522697745987075,
918
+ "loss": 0.8954,
919
+ "step": 645
920
+ },
921
+ {
922
+ "epoch": 0.6220095693779905,
923
+ "grad_norm": 1.0630327463150024,
924
+ "learning_rate": 0.000112793638881571,
925
+ "loss": 0.8974,
926
+ "step": 650
927
+ },
928
+ {
929
+ "epoch": 0.6267942583732058,
930
+ "grad_norm": 1.2224838733673096,
931
+ "learning_rate": 0.00011037068971930709,
932
+ "loss": 0.8945,
933
+ "step": 655
934
+ },
935
+ {
936
+ "epoch": 0.631578947368421,
937
+ "grad_norm": 1.2856760025024414,
938
+ "learning_rate": 0.00010795880655165656,
939
+ "loss": 0.8994,
940
+ "step": 660
941
+ },
942
+ {
943
+ "epoch": 0.6363636363636364,
944
+ "grad_norm": 1.0856343507766724,
945
+ "learning_rate": 0.0001055586628671549,
946
+ "loss": 0.8883,
947
+ "step": 665
948
+ },
949
+ {
950
+ "epoch": 0.6411483253588517,
951
+ "grad_norm": 1.063491940498352,
952
+ "learning_rate": 0.00010317092887623205,
953
+ "loss": 0.8866,
954
+ "step": 670
955
+ },
956
+ {
957
+ "epoch": 0.645933014354067,
958
+ "grad_norm": 1.2210334539413452,
959
+ "learning_rate": 0.00010079627132406462,
960
+ "loss": 0.8902,
961
+ "step": 675
962
+ },
963
+ {
964
+ "epoch": 0.6507177033492823,
965
+ "grad_norm": 0.9613766074180603,
966
+ "learning_rate": 9.843535330439555e-05,
967
+ "loss": 0.8748,
968
+ "step": 680
969
+ },
970
+ {
971
+ "epoch": 0.6555023923444976,
972
+ "grad_norm": 0.9286556839942932,
973
+ "learning_rate": 9.608883407437309e-05,
974
+ "loss": 0.8824,
975
+ "step": 685
976
+ },
977
+ {
978
+ "epoch": 0.6602870813397129,
979
+ "grad_norm": 1.13881254196167,
980
+ "learning_rate": 9.375736887046185e-05,
981
+ "loss": 0.8779,
982
+ "step": 690
983
+ },
984
+ {
985
+ "epoch": 0.6650717703349283,
986
+ "grad_norm": 1.128664255142212,
987
+ "learning_rate": 9.144160872547578e-05,
988
+ "loss": 0.8623,
989
+ "step": 695
990
+ },
991
+ {
992
+ "epoch": 0.6698564593301436,
993
+ "grad_norm": 1.3374972343444824,
994
+ "learning_rate": 8.914220028678546e-05,
995
+ "loss": 0.8831,
996
+ "step": 700
997
+ },
998
+ {
999
+ "epoch": 0.6746411483253588,
1000
+ "grad_norm": 1.5959270000457764,
1001
+ "learning_rate": 8.685978563574976e-05,
1002
+ "loss": 0.8956,
1003
+ "step": 705
1004
+ },
1005
+ {
1006
+ "epoch": 0.6794258373205742,
1007
+ "grad_norm": 1.582465410232544,
1008
+ "learning_rate": 8.459500210842261e-05,
1009
+ "loss": 0.8804,
1010
+ "step": 710
1011
+ },
1012
+ {
1013
+ "epoch": 0.6842105263157895,
1014
+ "grad_norm": 1.3153728246688843,
1015
+ "learning_rate": 8.23484821175847e-05,
1016
+ "loss": 0.8776,
1017
+ "step": 715
1018
+ },
1019
+ {
1020
+ "epoch": 0.6889952153110048,
1021
+ "grad_norm": 1.1769953966140747,
1022
+ "learning_rate": 8.012085297615027e-05,
1023
+ "loss": 0.8777,
1024
+ "step": 720
1025
+ },
1026
+ {
1027
+ "epoch": 0.69377990430622,
1028
+ "grad_norm": 1.1776398420333862,
1029
+ "learning_rate": 7.791273672199742e-05,
1030
+ "loss": 0.8866,
1031
+ "step": 725
1032
+ },
1033
+ {
1034
+ "epoch": 0.6985645933014354,
1035
+ "grad_norm": 1.0515143871307373,
1036
+ "learning_rate": 7.572474994427197e-05,
1037
+ "loss": 0.8887,
1038
+ "step": 730
1039
+ },
1040
+ {
1041
+ "epoch": 0.7033492822966507,
1042
+ "grad_norm": 1.0670396089553833,
1043
+ "learning_rate": 7.355750361121237e-05,
1044
+ "loss": 0.8631,
1045
+ "step": 735
1046
+ },
1047
+ {
1048
+ "epoch": 0.7081339712918661,
1049
+ "grad_norm": 1.3679234981536865,
1050
+ "learning_rate": 7.141160289954471e-05,
1051
+ "loss": 0.8701,
1052
+ "step": 740
1053
+ },
1054
+ {
1055
+ "epoch": 0.7129186602870813,
1056
+ "grad_norm": 1.0791184902191162,
1057
+ "learning_rate": 6.92876470254941e-05,
1058
+ "loss": 0.879,
1059
+ "step": 745
1060
+ },
1061
+ {
1062
+ "epoch": 0.7177033492822966,
1063
+ "grad_norm": 1.0610390901565552,
1064
+ "learning_rate": 6.718622907746151e-05,
1065
+ "loss": 0.8617,
1066
+ "step": 750
1067
+ },
1068
+ {
1069
+ "epoch": 0.722488038277512,
1070
+ "grad_norm": 1.062545657157898,
1071
+ "learning_rate": 6.51079358504107e-05,
1072
+ "loss": 0.8743,
1073
+ "step": 755
1074
+ },
1075
+ {
1076
+ "epoch": 0.7272727272727273,
1077
+ "grad_norm": 0.8820713758468628,
1078
+ "learning_rate": 6.305334768201336e-05,
1079
+ "loss": 0.8553,
1080
+ "step": 760
1081
+ },
1082
+ {
1083
+ "epoch": 0.7320574162679426,
1084
+ "grad_norm": 0.9072192311286926,
1085
+ "learning_rate": 6.1023038290596715e-05,
1086
+ "loss": 0.8613,
1087
+ "step": 765
1088
+ },
1089
+ {
1090
+ "epoch": 0.7368421052631579,
1091
+ "grad_norm": 1.0919402837753296,
1092
+ "learning_rate": 5.901757461493988e-05,
1093
+ "loss": 0.8679,
1094
+ "step": 770
1095
+ },
1096
+ {
1097
+ "epoch": 0.7416267942583732,
1098
+ "grad_norm": 1.0981807708740234,
1099
+ "learning_rate": 5.703751665596328e-05,
1100
+ "loss": 0.8665,
1101
+ "step": 775
1102
+ },
1103
+ {
1104
+ "epoch": 0.7464114832535885,
1105
+ "grad_norm": 1.0239495038986206,
1106
+ "learning_rate": 5.50834173203552e-05,
1107
+ "loss": 0.8819,
1108
+ "step": 780
1109
+ },
1110
+ {
1111
+ "epoch": 0.7511961722488039,
1112
+ "grad_norm": 1.0707833766937256,
1113
+ "learning_rate": 5.315582226617963e-05,
1114
+ "loss": 0.8571,
1115
+ "step": 785
1116
+ },
1117
+ {
1118
+ "epoch": 0.7559808612440191,
1119
+ "grad_norm": 1.0176221132278442,
1120
+ "learning_rate": 5.125526975050776e-05,
1121
+ "loss": 0.8718,
1122
+ "step": 790
1123
+ },
1124
+ {
1125
+ "epoch": 0.7607655502392344,
1126
+ "grad_norm": 0.9414092898368835,
1127
+ "learning_rate": 4.938229047911651e-05,
1128
+ "loss": 0.8547,
1129
+ "step": 795
1130
+ },
1131
+ {
1132
+ "epoch": 0.7655502392344498,
1133
+ "grad_norm": 0.8967716097831726,
1134
+ "learning_rate": 4.753740745829537e-05,
1135
+ "loss": 0.8385,
1136
+ "step": 800
1137
+ },
1138
+ {
1139
+ "epoch": 0.7703349282296651,
1140
+ "grad_norm": 1.160251498222351,
1141
+ "learning_rate": 4.5721135848803653e-05,
1142
+ "loss": 0.8471,
1143
+ "step": 805
1144
+ },
1145
+ {
1146
+ "epoch": 0.7751196172248804,
1147
+ "grad_norm": 0.9440346360206604,
1148
+ "learning_rate": 4.3933982822017876e-05,
1149
+ "loss": 0.8526,
1150
+ "step": 810
1151
+ },
1152
+ {
1153
+ "epoch": 0.7799043062200957,
1154
+ "grad_norm": 0.9394163489341736,
1155
+ "learning_rate": 4.217644741831062e-05,
1156
+ "loss": 0.8661,
1157
+ "step": 815
1158
+ },
1159
+ {
1160
+ "epoch": 0.784688995215311,
1161
+ "grad_norm": 0.9960381388664246,
1162
+ "learning_rate": 4.044902040769963e-05,
1163
+ "loss": 0.8848,
1164
+ "step": 820
1165
+ },
1166
+ {
1167
+ "epoch": 0.7894736842105263,
1168
+ "grad_norm": 0.9392317533493042,
1169
+ "learning_rate": 3.875218415280636e-05,
1170
+ "loss": 0.8562,
1171
+ "step": 825
1172
+ },
1173
+ {
1174
+ "epoch": 0.7942583732057417,
1175
+ "grad_norm": 1.0181559324264526,
1176
+ "learning_rate": 3.708641247416229e-05,
1177
+ "loss": 0.854,
1178
+ "step": 830
1179
+ },
1180
+ {
1181
+ "epoch": 0.7990430622009569,
1182
+ "grad_norm": 1.1440843343734741,
1183
+ "learning_rate": 3.54521705179002e-05,
1184
+ "loss": 0.8597,
1185
+ "step": 835
1186
+ },
1187
+ {
1188
+ "epoch": 0.8038277511961722,
1189
+ "grad_norm": 1.0670745372772217,
1190
+ "learning_rate": 3.3849914625868004e-05,
1191
+ "loss": 0.8847,
1192
+ "step": 840
1193
+ },
1194
+ {
1195
+ "epoch": 0.8086124401913876,
1196
+ "grad_norm": 0.9774266481399536,
1197
+ "learning_rate": 3.228009220820085e-05,
1198
+ "loss": 0.8673,
1199
+ "step": 845
1200
+ },
1201
+ {
1202
+ "epoch": 0.8133971291866029,
1203
+ "grad_norm": 0.9691421389579773,
1204
+ "learning_rate": 3.074314161838754e-05,
1205
+ "loss": 0.8315,
1206
+ "step": 850
1207
+ },
1208
+ {
1209
+ "epoch": 0.8181818181818182,
1210
+ "grad_norm": 0.9607205986976624,
1211
+ "learning_rate": 2.923949203086557e-05,
1212
+ "loss": 0.862,
1213
+ "step": 855
1214
+ },
1215
+ {
1216
+ "epoch": 0.8229665071770335,
1217
+ "grad_norm": 1.0411772727966309,
1218
+ "learning_rate": 2.77695633211795e-05,
1219
+ "loss": 0.8659,
1220
+ "step": 860
1221
+ },
1222
+ {
1223
+ "epoch": 0.8277511961722488,
1224
+ "grad_norm": 1.090226173400879,
1225
+ "learning_rate": 2.6333765948735986e-05,
1226
+ "loss": 0.8491,
1227
+ "step": 865
1228
+ },
1229
+ {
1230
+ "epoch": 0.8325358851674641,
1231
+ "grad_norm": 0.8712325096130371,
1232
+ "learning_rate": 2.493250084218795e-05,
1233
+ "loss": 0.8626,
1234
+ "step": 870
1235
+ },
1236
+ {
1237
+ "epoch": 0.8373205741626795,
1238
+ "grad_norm": 0.8832561373710632,
1239
+ "learning_rate": 2.3566159287480447e-05,
1240
+ "loss": 0.8479,
1241
+ "step": 875
1242
+ },
1243
+ {
1244
+ "epoch": 0.8421052631578947,
1245
+ "grad_norm": 0.9606965184211731,
1246
+ "learning_rate": 2.2235122818588565e-05,
1247
+ "loss": 0.8625,
1248
+ "step": 880
1249
+ },
1250
+ {
1251
+ "epoch": 0.84688995215311,
1252
+ "grad_norm": 1.056420087814331,
1253
+ "learning_rate": 2.0939763110979125e-05,
1254
+ "loss": 0.863,
1255
+ "step": 885
1256
+ },
1257
+ {
1258
+ "epoch": 0.8516746411483254,
1259
+ "grad_norm": 0.9405905604362488,
1260
+ "learning_rate": 1.9680441877824527e-05,
1261
+ "loss": 0.853,
1262
+ "step": 890
1263
+ },
1264
+ {
1265
+ "epoch": 0.8564593301435407,
1266
+ "grad_norm": 0.8905633687973022,
1267
+ "learning_rate": 1.8457510768999273e-05,
1268
+ "loss": 0.8431,
1269
+ "step": 895
1270
+ },
1271
+ {
1272
+ "epoch": 0.861244019138756,
1273
+ "grad_norm": 0.981576144695282,
1274
+ "learning_rate": 1.727131127288572e-05,
1275
+ "loss": 0.8477,
1276
+ "step": 900
1277
+ },
1278
+ {
1279
+ "epoch": 0.8660287081339713,
1280
+ "grad_norm": 1.0129060745239258,
1281
+ "learning_rate": 1.612217462101783e-05,
1282
+ "loss": 0.8384,
1283
+ "step": 905
1284
+ },
1285
+ {
1286
+ "epoch": 0.8708133971291866,
1287
+ "grad_norm": 0.9811695218086243,
1288
+ "learning_rate": 1.5010421695588825e-05,
1289
+ "loss": 0.8573,
1290
+ "step": 910
1291
+ },
1292
+ {
1293
+ "epoch": 0.8755980861244019,
1294
+ "grad_norm": 0.9269863963127136,
1295
+ "learning_rate": 1.3936362939848893e-05,
1296
+ "loss": 0.8561,
1297
+ "step": 915
1298
+ },
1299
+ {
1300
+ "epoch": 0.8803827751196173,
1301
+ "grad_norm": 0.8773042559623718,
1302
+ "learning_rate": 1.2900298271417592e-05,
1303
+ "loss": 0.8203,
1304
+ "step": 920
1305
+ },
1306
+ {
1307
+ "epoch": 0.8851674641148325,
1308
+ "grad_norm": 0.9804287552833557,
1309
+ "learning_rate": 1.1902516998535666e-05,
1310
+ "loss": 0.8382,
1311
+ "step": 925
1312
+ },
1313
+ {
1314
+ "epoch": 0.8899521531100478,
1315
+ "grad_norm": 0.8934711217880249,
1316
+ "learning_rate": 1.0943297739279395e-05,
1317
+ "loss": 0.8416,
1318
+ "step": 930
1319
+ },
1320
+ {
1321
+ "epoch": 0.8947368421052632,
1322
+ "grad_norm": 1.0209717750549316,
1323
+ "learning_rate": 1.0022908343759616e-05,
1324
+ "loss": 0.8595,
1325
+ "step": 935
1326
+ },
1327
+ {
1328
+ "epoch": 0.8995215311004785,
1329
+ "grad_norm": 0.8881580829620361,
1330
+ "learning_rate": 9.141605819328318e-06,
1331
+ "loss": 0.8332,
1332
+ "step": 940
1333
+ },
1334
+ {
1335
+ "epoch": 0.9043062200956937,
1336
+ "grad_norm": 0.8664154410362244,
1337
+ "learning_rate": 8.299636258812197e-06,
1338
+ "loss": 0.8437,
1339
+ "step": 945
1340
+ },
1341
+ {
1342
+ "epoch": 0.9090909090909091,
1343
+ "grad_norm": 0.8912498950958252,
1344
+ "learning_rate": 7.4972347717943916e-06,
1345
+ "loss": 0.8485,
1346
+ "step": 950
1347
+ },
1348
+ {
1349
+ "epoch": 0.9138755980861244,
1350
+ "grad_norm": 0.9088477492332458,
1351
+ "learning_rate": 6.734625418963135e-06,
1352
+ "loss": 0.8472,
1353
+ "step": 955
1354
+ },
1355
+ {
1356
+ "epoch": 0.9186602870813397,
1357
+ "grad_norm": 0.9696795344352722,
1358
+ "learning_rate": 6.012021149545382e-06,
1359
+ "loss": 0.8548,
1360
+ "step": 960
1361
+ },
1362
+ {
1363
+ "epoch": 0.9234449760765551,
1364
+ "grad_norm": 0.9158973097801208,
1365
+ "learning_rate": 5.329623741843531e-06,
1366
+ "loss": 0.8333,
1367
+ "step": 965
1368
+ },
1369
+ {
1370
+ "epoch": 0.9282296650717703,
1371
+ "grad_norm": 0.9025459289550781,
1372
+ "learning_rate": 4.6876237468912e-06,
1373
+ "loss": 0.8422,
1374
+ "step": 970
1375
+ },
1376
+ {
1377
+ "epoch": 0.9330143540669856,
1378
+ "grad_norm": 0.8703954815864563,
1379
+ "learning_rate": 4.086200435244441e-06,
1380
+ "loss": 0.8265,
1381
+ "step": 975
1382
+ },
1383
+ {
1384
+ "epoch": 0.937799043062201,
1385
+ "grad_norm": 0.9033318161964417,
1386
+ "learning_rate": 3.5255217469222942e-06,
1387
+ "loss": 0.8203,
1388
+ "step": 980
1389
+ },
1390
+ {
1391
+ "epoch": 0.9425837320574163,
1392
+ "grad_norm": 1.0247541666030884,
1393
+ "learning_rate": 3.0057442445119872e-06,
1394
+ "loss": 0.8296,
1395
+ "step": 985
1396
+ },
1397
+ {
1398
+ "epoch": 0.9473684210526315,
1399
+ "grad_norm": 0.9569628238677979,
1400
+ "learning_rate": 2.5270130694504733e-06,
1401
+ "loss": 0.8644,
1402
+ "step": 990
1403
+ },
1404
+ {
1405
+ "epoch": 0.9521531100478469,
1406
+ "grad_norm": 0.8948321342468262,
1407
+ "learning_rate": 2.089461901495715e-06,
1408
+ "loss": 0.8278,
1409
+ "step": 995
1410
+ },
1411
+ {
1412
+ "epoch": 0.9569377990430622,
1413
+ "grad_norm": 0.969680666923523,
1414
+ "learning_rate": 1.6932129213980793e-06,
1415
+ "loss": 0.8452,
1416
+ "step": 1000
1417
+ },
1418
+ {
1419
+ "epoch": 0.9617224880382775,
1420
+ "grad_norm": 0.9193556904792786,
1421
+ "learning_rate": 1.3383767767829956e-06,
1422
+ "loss": 0.8341,
1423
+ "step": 1005
1424
+ },
1425
+ {
1426
+ "epoch": 0.9665071770334929,
1427
+ "grad_norm": 0.9120749235153198,
1428
+ "learning_rate": 1.0250525512538855e-06,
1429
+ "loss": 0.8359,
1430
+ "step": 1010
1431
+ },
1432
+ {
1433
+ "epoch": 0.9712918660287081,
1434
+ "grad_norm": 0.8678485155105591,
1435
+ "learning_rate": 7.533277367243795e-07,
1436
+ "loss": 0.8187,
1437
+ "step": 1015
1438
+ },
1439
+ {
1440
+ "epoch": 0.9760765550239234,
1441
+ "grad_norm": 1.054465889930725,
1442
+ "learning_rate": 5.232782089872601e-07,
1443
+ "loss": 0.8487,
1444
+ "step": 1020
1445
+ },
1446
+ {
1447
+ "epoch": 0.9808612440191388,
1448
+ "grad_norm": 0.8861187696456909,
1449
+ "learning_rate": 3.349682065270254e-07,
1450
+ "loss": 0.8504,
1451
+ "step": 1025
1452
+ },
1453
+ {
1454
+ "epoch": 0.9856459330143541,
1455
+ "grad_norm": 0.9375335574150085,
1456
+ "learning_rate": 1.8845031258213905e-07,
1457
+ "loss": 0.8459,
1458
+ "step": 1030
1459
+ },
1460
+ {
1461
+ "epoch": 0.9904306220095693,
1462
+ "grad_norm": 0.9127489924430847,
1463
+ "learning_rate": 8.376544046180822e-08,
1464
+ "loss": 0.8447,
1465
+ "step": 1035
1466
+ },
1467
+ {
1468
+ "epoch": 0.9952153110047847,
1469
+ "grad_norm": 0.8907783031463623,
1470
+ "learning_rate": 2.0942822121439164e-08,
1471
+ "loss": 0.8246,
1472
+ "step": 1040
1473
+ },
1474
+ {
1475
+ "epoch": 1.0,
1476
+ "grad_norm": 0.9073578715324402,
1477
+ "learning_rate": 0.0,
1478
+ "loss": 0.8382,
1479
+ "step": 1045
1480
+ },
1481
+ {
1482
+ "epoch": 1.0,
1483
+ "eval_loss": 2.4202542304992676,
1484
+ "eval_runtime": 1.1602,
1485
+ "eval_samples_per_second": 8.619,
1486
+ "eval_steps_per_second": 0.862,
1487
+ "step": 1045
1488
+ },
1489
+ {
1490
+ "epoch": 1.0,
1491
+ "step": 1045,
1492
+ "total_flos": 6.049215074407547e+17,
1493
+ "train_loss": 1.5597172723432478,
1494
+ "train_runtime": 7161.753,
1495
+ "train_samples_per_second": 1.751,
1496
+ "train_steps_per_second": 0.146
1497
+ }
1498
+ ],
1499
+ "logging_steps": 5,
1500
+ "max_steps": 1045,
1501
+ "num_input_tokens_seen": 0,
1502
+ "num_train_epochs": 1,
1503
+ "save_steps": 100,
1504
+ "stateful_callbacks": {
1505
+ "TrainerControl": {
1506
+ "args": {
1507
+ "should_epoch_stop": false,
1508
+ "should_evaluate": false,
1509
+ "should_log": false,
1510
+ "should_save": true,
1511
+ "should_training_stop": true
1512
+ },
1513
+ "attributes": {}
1514
+ }
1515
+ },
1516
+ "total_flos": 6.049215074407547e+17,
1517
+ "train_batch_size": 1,
1518
+ "trial_name": null,
1519
+ "trial_params": null
1520
+ }