AmberYifan commited on
Commit
3b68dcb
·
verified ·
1 Parent(s): b81e228

Model save

Browse files
README.md ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: AmberYifan/Mistral-7B-v0.3-sft-ultrachat
3
+ library_name: transformers
4
+ model_name: Mistral-7B-v0.3-sft-ultrachat-hhrlhf
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for Mistral-7B-v0.3-sft-ultrachat-hhrlhf
13
+
14
+ This model is a fine-tuned version of [AmberYifan/Mistral-7B-v0.3-sft-ultrachat](https://huggingface.co/AmberYifan/Mistral-7B-v0.3-sft-ultrachat).
15
+ It has been trained using [TRL](https://github.com/huggingface/trl).
16
+
17
+ ## Quick start
18
+
19
+ ```python
20
+ from transformers import pipeline
21
+
22
+ question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
23
+ generator = pipeline("text-generation", model="AmberYifan/Mistral-7B-v0.3-sft-ultrachat-hhrlhf", device="cuda")
24
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
25
+ print(output["generated_text"])
26
+ ```
27
+
28
+ ## Training procedure
29
+
30
+
31
+
32
+ This model was trained with SFT.
33
+
34
+ ### Framework versions
35
+
36
+ - TRL: 0.12.2
37
+ - Transformers: 4.46.3
38
+ - Pytorch: 2.5.1+cu118
39
+ - Datasets: 3.2.0
40
+ - Tokenizers: 0.20.3
41
+
42
+ ## Citations
43
+
44
+
45
+
46
+ Cite TRL as:
47
+
48
+ ```bibtex
49
+ @misc{vonwerra2022trl,
50
+ title = {{TRL: Transformer Reinforcement Learning}},
51
+ author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
52
+ year = 2020,
53
+ journal = {GitHub repository},
54
+ publisher = {GitHub},
55
+ howpublished = {\url{https://github.com/huggingface/trl}}
56
+ }
57
+ ```
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 38150611599360.0,
4
+ "train_loss": 1.3537356438578816,
5
+ "train_runtime": 6160.8214,
6
+ "train_samples": 50000,
7
+ "train_samples_per_second": 8.116,
8
+ "train_steps_per_second": 0.254
9
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.46.3"
6
+ }
runs/Jan26_04-49-42_gilbreth-j001.rcac.purdue.edu/events.out.tfevents.1737885156.gilbreth-j001.rcac.purdue.edu.244816.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5c94ff15851c8055ea8f2a0654e720864d8bc71a2b364111f0bf2b2c29d05070
3
- size 72095
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c207160ad1a942bc83b94f9b5ab9f3c479b1cbca3c6d9171da7a71ec9ef58bb8
3
+ size 72449
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 38150611599360.0,
4
+ "train_loss": 1.3537356438578816,
5
+ "train_runtime": 6160.8214,
6
+ "train_samples": 50000,
7
+ "train_samples_per_second": 8.116,
8
+ "train_steps_per_second": 0.254
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,2241 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 1563,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0006397952655150352,
13
+ "grad_norm": 19.596042230632786,
14
+ "learning_rate": 1.2738853503184715e-07,
15
+ "loss": 1.9254,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.003198976327575176,
20
+ "grad_norm": 19.320989757503053,
21
+ "learning_rate": 6.369426751592357e-07,
22
+ "loss": 1.8985,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.006397952655150352,
27
+ "grad_norm": 10.48203671169689,
28
+ "learning_rate": 1.2738853503184715e-06,
29
+ "loss": 1.8034,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.009596928982725527,
34
+ "grad_norm": 8.35295489819408,
35
+ "learning_rate": 1.9108280254777074e-06,
36
+ "loss": 1.6227,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.012795905310300703,
41
+ "grad_norm": 6.366704733146609,
42
+ "learning_rate": 2.547770700636943e-06,
43
+ "loss": 1.5116,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.01599488163787588,
48
+ "grad_norm": 5.982203446460906,
49
+ "learning_rate": 3.1847133757961785e-06,
50
+ "loss": 1.4995,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.019193857965451054,
55
+ "grad_norm": 5.819962966473566,
56
+ "learning_rate": 3.821656050955415e-06,
57
+ "loss": 1.4278,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.022392834293026232,
62
+ "grad_norm": 6.029586520004324,
63
+ "learning_rate": 4.45859872611465e-06,
64
+ "loss": 1.4064,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.025591810620601407,
69
+ "grad_norm": 6.155880824723257,
70
+ "learning_rate": 5.095541401273886e-06,
71
+ "loss": 1.4693,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.028790786948176585,
76
+ "grad_norm": 6.183603254054322,
77
+ "learning_rate": 5.732484076433121e-06,
78
+ "loss": 1.4451,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.03198976327575176,
83
+ "grad_norm": 6.0739281893318,
84
+ "learning_rate": 6.369426751592357e-06,
85
+ "loss": 1.4503,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.035188739603326934,
90
+ "grad_norm": 5.666112734989155,
91
+ "learning_rate": 7.006369426751593e-06,
92
+ "loss": 1.4281,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.03838771593090211,
97
+ "grad_norm": 6.215484436208105,
98
+ "learning_rate": 7.64331210191083e-06,
99
+ "loss": 1.4641,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.04158669225847729,
104
+ "grad_norm": 5.91078945631238,
105
+ "learning_rate": 8.280254777070064e-06,
106
+ "loss": 1.4672,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.044785668586052464,
111
+ "grad_norm": 5.661599095084434,
112
+ "learning_rate": 8.9171974522293e-06,
113
+ "loss": 1.4572,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.04798464491362764,
118
+ "grad_norm": 6.399453722973696,
119
+ "learning_rate": 9.554140127388536e-06,
120
+ "loss": 1.4781,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.05118362124120281,
125
+ "grad_norm": 6.272431960129444,
126
+ "learning_rate": 1.0191082802547772e-05,
127
+ "loss": 1.4413,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.05438259756877799,
132
+ "grad_norm": 5.809472680223379,
133
+ "learning_rate": 1.0828025477707008e-05,
134
+ "loss": 1.4618,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.05758157389635317,
139
+ "grad_norm": 5.8795539576606854,
140
+ "learning_rate": 1.1464968152866242e-05,
141
+ "loss": 1.4612,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.060780550223928344,
146
+ "grad_norm": 6.353272916584155,
147
+ "learning_rate": 1.2101910828025478e-05,
148
+ "loss": 1.4663,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.06397952655150352,
153
+ "grad_norm": 5.8194943048004975,
154
+ "learning_rate": 1.2738853503184714e-05,
155
+ "loss": 1.4683,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.0671785028790787,
160
+ "grad_norm": 5.748599336291074,
161
+ "learning_rate": 1.337579617834395e-05,
162
+ "loss": 1.4635,
163
+ "step": 105
164
+ },
165
+ {
166
+ "epoch": 0.07037747920665387,
167
+ "grad_norm": 5.446443880371817,
168
+ "learning_rate": 1.4012738853503186e-05,
169
+ "loss": 1.4833,
170
+ "step": 110
171
+ },
172
+ {
173
+ "epoch": 0.07357645553422905,
174
+ "grad_norm": 5.846022334061706,
175
+ "learning_rate": 1.464968152866242e-05,
176
+ "loss": 1.4841,
177
+ "step": 115
178
+ },
179
+ {
180
+ "epoch": 0.07677543186180422,
181
+ "grad_norm": 5.8983508370893585,
182
+ "learning_rate": 1.528662420382166e-05,
183
+ "loss": 1.5009,
184
+ "step": 120
185
+ },
186
+ {
187
+ "epoch": 0.0799744081893794,
188
+ "grad_norm": 5.580182994348059,
189
+ "learning_rate": 1.5923566878980894e-05,
190
+ "loss": 1.5282,
191
+ "step": 125
192
+ },
193
+ {
194
+ "epoch": 0.08317338451695458,
195
+ "grad_norm": 7.00216131929732,
196
+ "learning_rate": 1.6560509554140128e-05,
197
+ "loss": 1.5138,
198
+ "step": 130
199
+ },
200
+ {
201
+ "epoch": 0.08637236084452975,
202
+ "grad_norm": 5.395228647689322,
203
+ "learning_rate": 1.7197452229299365e-05,
204
+ "loss": 1.5527,
205
+ "step": 135
206
+ },
207
+ {
208
+ "epoch": 0.08957133717210493,
209
+ "grad_norm": 6.4381395397364924,
210
+ "learning_rate": 1.78343949044586e-05,
211
+ "loss": 1.5202,
212
+ "step": 140
213
+ },
214
+ {
215
+ "epoch": 0.0927703134996801,
216
+ "grad_norm": 7.298084151887251,
217
+ "learning_rate": 1.8471337579617837e-05,
218
+ "loss": 1.5144,
219
+ "step": 145
220
+ },
221
+ {
222
+ "epoch": 0.09596928982725528,
223
+ "grad_norm": 5.820918003601176,
224
+ "learning_rate": 1.910828025477707e-05,
225
+ "loss": 1.4815,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.09916826615483046,
230
+ "grad_norm": 5.261025002970225,
231
+ "learning_rate": 1.9745222929936306e-05,
232
+ "loss": 1.5365,
233
+ "step": 155
234
+ },
235
+ {
236
+ "epoch": 0.10236724248240563,
237
+ "grad_norm": 5.822872418153333,
238
+ "learning_rate": 1.9999775332635076e-05,
239
+ "loss": 1.495,
240
+ "step": 160
241
+ },
242
+ {
243
+ "epoch": 0.10556621880998081,
244
+ "grad_norm": 5.990794687517629,
245
+ "learning_rate": 1.999840240196313e-05,
246
+ "loss": 1.5071,
247
+ "step": 165
248
+ },
249
+ {
250
+ "epoch": 0.10876519513755598,
251
+ "grad_norm": 5.613973413489411,
252
+ "learning_rate": 1.9995781526975738e-05,
253
+ "loss": 1.5319,
254
+ "step": 170
255
+ },
256
+ {
257
+ "epoch": 0.11196417146513116,
258
+ "grad_norm": 6.584077464363587,
259
+ "learning_rate": 1.9991913034795768e-05,
260
+ "loss": 1.5861,
261
+ "step": 175
262
+ },
263
+ {
264
+ "epoch": 0.11516314779270634,
265
+ "grad_norm": 5.848500499551572,
266
+ "learning_rate": 1.9986797408266636e-05,
267
+ "loss": 1.5,
268
+ "step": 180
269
+ },
270
+ {
271
+ "epoch": 0.1183621241202815,
272
+ "grad_norm": 5.564555003715003,
273
+ "learning_rate": 1.9980435285892056e-05,
274
+ "loss": 1.5481,
275
+ "step": 185
276
+ },
277
+ {
278
+ "epoch": 0.12156110044785669,
279
+ "grad_norm": 6.028564963206183,
280
+ "learning_rate": 1.9972827461756335e-05,
281
+ "loss": 1.578,
282
+ "step": 190
283
+ },
284
+ {
285
+ "epoch": 0.12476007677543186,
286
+ "grad_norm": 5.926547243421822,
287
+ "learning_rate": 1.9963974885425267e-05,
288
+ "loss": 1.5555,
289
+ "step": 195
290
+ },
291
+ {
292
+ "epoch": 0.12795905310300704,
293
+ "grad_norm": 5.966985587917129,
294
+ "learning_rate": 1.9953878661827603e-05,
295
+ "loss": 1.5288,
296
+ "step": 200
297
+ },
298
+ {
299
+ "epoch": 0.13115802943058222,
300
+ "grad_norm": 6.010709482959101,
301
+ "learning_rate": 1.994254005111715e-05,
302
+ "loss": 1.5387,
303
+ "step": 205
304
+ },
305
+ {
306
+ "epoch": 0.1343570057581574,
307
+ "grad_norm": 6.628579244426843,
308
+ "learning_rate": 1.992996046851548e-05,
309
+ "loss": 1.5874,
310
+ "step": 210
311
+ },
312
+ {
313
+ "epoch": 0.13755598208573255,
314
+ "grad_norm": 6.675308221300031,
315
+ "learning_rate": 1.9916141484135297e-05,
316
+ "loss": 1.5012,
317
+ "step": 215
318
+ },
319
+ {
320
+ "epoch": 0.14075495841330773,
321
+ "grad_norm": 6.371390174895763,
322
+ "learning_rate": 1.990108482278446e-05,
323
+ "loss": 1.5824,
324
+ "step": 220
325
+ },
326
+ {
327
+ "epoch": 0.14395393474088292,
328
+ "grad_norm": 5.64251865194712,
329
+ "learning_rate": 1.9884792363750684e-05,
330
+ "loss": 1.5324,
331
+ "step": 225
332
+ },
333
+ {
334
+ "epoch": 0.1471529110684581,
335
+ "grad_norm": 6.410849237786982,
336
+ "learning_rate": 1.9867266140567024e-05,
337
+ "loss": 1.5654,
338
+ "step": 230
339
+ },
340
+ {
341
+ "epoch": 0.15035188739603328,
342
+ "grad_norm": 7.066430464872344,
343
+ "learning_rate": 1.9848508340758014e-05,
344
+ "loss": 1.5612,
345
+ "step": 235
346
+ },
347
+ {
348
+ "epoch": 0.15355086372360843,
349
+ "grad_norm": 6.525361010219877,
350
+ "learning_rate": 1.9828521305566647e-05,
351
+ "loss": 1.5963,
352
+ "step": 240
353
+ },
354
+ {
355
+ "epoch": 0.15674984005118361,
356
+ "grad_norm": 27.981136788668703,
357
+ "learning_rate": 1.9807307529662175e-05,
358
+ "loss": 1.6072,
359
+ "step": 245
360
+ },
361
+ {
362
+ "epoch": 0.1599488163787588,
363
+ "grad_norm": 14.850012678039711,
364
+ "learning_rate": 1.9784869660828708e-05,
365
+ "loss": 1.6261,
366
+ "step": 250
367
+ },
368
+ {
369
+ "epoch": 0.16314779270633398,
370
+ "grad_norm": 9.017990600325932,
371
+ "learning_rate": 1.9761210499634754e-05,
372
+ "loss": 1.5862,
373
+ "step": 255
374
+ },
375
+ {
376
+ "epoch": 0.16634676903390916,
377
+ "grad_norm": 6.009881994757128,
378
+ "learning_rate": 1.973633299908365e-05,
379
+ "loss": 1.5476,
380
+ "step": 260
381
+ },
382
+ {
383
+ "epoch": 0.1695457453614843,
384
+ "grad_norm": 7.772193560607181,
385
+ "learning_rate": 1.9710240264245005e-05,
386
+ "loss": 1.5568,
387
+ "step": 265
388
+ },
389
+ {
390
+ "epoch": 0.1727447216890595,
391
+ "grad_norm": 6.0046356700151575,
392
+ "learning_rate": 1.9682935551867132e-05,
393
+ "loss": 1.5652,
394
+ "step": 270
395
+ },
396
+ {
397
+ "epoch": 0.17594369801663468,
398
+ "grad_norm": 7.259518897268176,
399
+ "learning_rate": 1.9654422269970545e-05,
400
+ "loss": 1.5152,
401
+ "step": 275
402
+ },
403
+ {
404
+ "epoch": 0.17914267434420986,
405
+ "grad_norm": 5.849933171807072,
406
+ "learning_rate": 1.9624703977422624e-05,
407
+ "loss": 1.5411,
408
+ "step": 280
409
+ },
410
+ {
411
+ "epoch": 0.18234165067178504,
412
+ "grad_norm": 6.900313411361811,
413
+ "learning_rate": 1.959378438349338e-05,
414
+ "loss": 1.5514,
415
+ "step": 285
416
+ },
417
+ {
418
+ "epoch": 0.1855406269993602,
419
+ "grad_norm": 5.976493878932652,
420
+ "learning_rate": 1.956166734739251e-05,
421
+ "loss": 1.5491,
422
+ "step": 290
423
+ },
424
+ {
425
+ "epoch": 0.18873960332693537,
426
+ "grad_norm": 6.839976739709417,
427
+ "learning_rate": 1.95283568777877e-05,
428
+ "loss": 1.5808,
429
+ "step": 295
430
+ },
431
+ {
432
+ "epoch": 0.19193857965451055,
433
+ "grad_norm": 23.36116832288107,
434
+ "learning_rate": 1.9493857132304295e-05,
435
+ "loss": 1.5103,
436
+ "step": 300
437
+ },
438
+ {
439
+ "epoch": 0.19513755598208574,
440
+ "grad_norm": 5.907471018056248,
441
+ "learning_rate": 1.9458172417006347e-05,
442
+ "loss": 1.5491,
443
+ "step": 305
444
+ },
445
+ {
446
+ "epoch": 0.19833653230966092,
447
+ "grad_norm": 5.462387821543604,
448
+ "learning_rate": 1.942130718585919e-05,
449
+ "loss": 1.4817,
450
+ "step": 310
451
+ },
452
+ {
453
+ "epoch": 0.20153550863723607,
454
+ "grad_norm": 5.445829839166847,
455
+ "learning_rate": 1.938326604017349e-05,
456
+ "loss": 1.5581,
457
+ "step": 315
458
+ },
459
+ {
460
+ "epoch": 0.20473448496481125,
461
+ "grad_norm": 6.17671143618823,
462
+ "learning_rate": 1.9344053728030952e-05,
463
+ "loss": 1.4874,
464
+ "step": 320
465
+ },
466
+ {
467
+ "epoch": 0.20793346129238643,
468
+ "grad_norm": 5.454424295381491,
469
+ "learning_rate": 1.9303675143691683e-05,
470
+ "loss": 1.5252,
471
+ "step": 325
472
+ },
473
+ {
474
+ "epoch": 0.21113243761996162,
475
+ "grad_norm": 5.820669131673871,
476
+ "learning_rate": 1.9262135326983326e-05,
477
+ "loss": 1.5198,
478
+ "step": 330
479
+ },
480
+ {
481
+ "epoch": 0.2143314139475368,
482
+ "grad_norm": 5.912753814527551,
483
+ "learning_rate": 1.921943946267201e-05,
484
+ "loss": 1.587,
485
+ "step": 335
486
+ },
487
+ {
488
+ "epoch": 0.21753039027511195,
489
+ "grad_norm": 7.487537515670172,
490
+ "learning_rate": 1.9175592879815217e-05,
491
+ "loss": 1.5657,
492
+ "step": 340
493
+ },
494
+ {
495
+ "epoch": 0.22072936660268713,
496
+ "grad_norm": 5.853172355021576,
497
+ "learning_rate": 1.9130601051096655e-05,
498
+ "loss": 1.4938,
499
+ "step": 345
500
+ },
501
+ {
502
+ "epoch": 0.22392834293026231,
503
+ "grad_norm": 5.709149350487184,
504
+ "learning_rate": 1.9084469592143154e-05,
505
+ "loss": 1.5328,
506
+ "step": 350
507
+ },
508
+ {
509
+ "epoch": 0.2271273192578375,
510
+ "grad_norm": 7.575317890582474,
511
+ "learning_rate": 1.9037204260823788e-05,
512
+ "loss": 1.4863,
513
+ "step": 355
514
+ },
515
+ {
516
+ "epoch": 0.23032629558541268,
517
+ "grad_norm": 6.140971076057151,
518
+ "learning_rate": 1.89888109565312e-05,
519
+ "loss": 1.5559,
520
+ "step": 360
521
+ },
522
+ {
523
+ "epoch": 0.23352527191298783,
524
+ "grad_norm": 6.7321671777819745,
525
+ "learning_rate": 1.893929571944527e-05,
526
+ "loss": 1.4776,
527
+ "step": 365
528
+ },
529
+ {
530
+ "epoch": 0.236724248240563,
531
+ "grad_norm": 5.673506383692647,
532
+ "learning_rate": 1.8888664729779205e-05,
533
+ "loss": 1.5505,
534
+ "step": 370
535
+ },
536
+ {
537
+ "epoch": 0.2399232245681382,
538
+ "grad_norm": 6.7549167620013755,
539
+ "learning_rate": 1.883692430700818e-05,
540
+ "loss": 1.5136,
541
+ "step": 375
542
+ },
543
+ {
544
+ "epoch": 0.24312220089571338,
545
+ "grad_norm": 5.657108216640427,
546
+ "learning_rate": 1.8784080909080568e-05,
547
+ "loss": 1.4313,
548
+ "step": 380
549
+ },
550
+ {
551
+ "epoch": 0.24632117722328856,
552
+ "grad_norm": 5.706894107517054,
553
+ "learning_rate": 1.8730141131611882e-05,
554
+ "loss": 1.5071,
555
+ "step": 385
556
+ },
557
+ {
558
+ "epoch": 0.2495201535508637,
559
+ "grad_norm": 5.492551858258093,
560
+ "learning_rate": 1.867511170706157e-05,
561
+ "loss": 1.4764,
562
+ "step": 390
563
+ },
564
+ {
565
+ "epoch": 0.2527191298784389,
566
+ "grad_norm": 17.304144834459898,
567
+ "learning_rate": 1.861899950389269e-05,
568
+ "loss": 1.5008,
569
+ "step": 395
570
+ },
571
+ {
572
+ "epoch": 0.2559181062060141,
573
+ "grad_norm": 6.5102974464825705,
574
+ "learning_rate": 1.856181152571463e-05,
575
+ "loss": 1.5829,
576
+ "step": 400
577
+ },
578
+ {
579
+ "epoch": 0.2591170825335892,
580
+ "grad_norm": 5.948826356074738,
581
+ "learning_rate": 1.850355491040897e-05,
582
+ "loss": 1.5518,
583
+ "step": 405
584
+ },
585
+ {
586
+ "epoch": 0.26231605886116444,
587
+ "grad_norm": 5.971106058355866,
588
+ "learning_rate": 1.8444236929238556e-05,
589
+ "loss": 1.5757,
590
+ "step": 410
591
+ },
592
+ {
593
+ "epoch": 0.2655150351887396,
594
+ "grad_norm": 5.8679997341744885,
595
+ "learning_rate": 1.8383864985939944e-05,
596
+ "loss": 1.5092,
597
+ "step": 415
598
+ },
599
+ {
600
+ "epoch": 0.2687140115163148,
601
+ "grad_norm": 5.6678931253995275,
602
+ "learning_rate": 1.8322446615799317e-05,
603
+ "loss": 1.5914,
604
+ "step": 420
605
+ },
606
+ {
607
+ "epoch": 0.27191298784388995,
608
+ "grad_norm": 5.73166539396345,
609
+ "learning_rate": 1.825998948471197e-05,
610
+ "loss": 1.5175,
611
+ "step": 425
612
+ },
613
+ {
614
+ "epoch": 0.2751119641714651,
615
+ "grad_norm": 6.505274488122578,
616
+ "learning_rate": 1.819650138822548e-05,
617
+ "loss": 1.5308,
618
+ "step": 430
619
+ },
620
+ {
621
+ "epoch": 0.2783109404990403,
622
+ "grad_norm": 6.1831675736655365,
623
+ "learning_rate": 1.8131990250566733e-05,
624
+ "loss": 1.4676,
625
+ "step": 435
626
+ },
627
+ {
628
+ "epoch": 0.28150991682661547,
629
+ "grad_norm": 5.898782759058662,
630
+ "learning_rate": 1.8066464123652857e-05,
631
+ "loss": 1.4624,
632
+ "step": 440
633
+ },
634
+ {
635
+ "epoch": 0.2847088931541907,
636
+ "grad_norm": 6.0743511266401455,
637
+ "learning_rate": 1.7999931186086225e-05,
638
+ "loss": 1.497,
639
+ "step": 445
640
+ },
641
+ {
642
+ "epoch": 0.28790786948176583,
643
+ "grad_norm": 6.439702879767435,
644
+ "learning_rate": 1.793239974213364e-05,
645
+ "loss": 1.4987,
646
+ "step": 450
647
+ },
648
+ {
649
+ "epoch": 0.291106845809341,
650
+ "grad_norm": 5.652845631220745,
651
+ "learning_rate": 1.786387822068987e-05,
652
+ "loss": 1.4731,
653
+ "step": 455
654
+ },
655
+ {
656
+ "epoch": 0.2943058221369162,
657
+ "grad_norm": 6.32893452010987,
658
+ "learning_rate": 1.7794375174225565e-05,
659
+ "loss": 1.5686,
660
+ "step": 460
661
+ },
662
+ {
663
+ "epoch": 0.29750479846449135,
664
+ "grad_norm": 6.832606404525833,
665
+ "learning_rate": 1.7723899277719815e-05,
666
+ "loss": 1.5616,
667
+ "step": 465
668
+ },
669
+ {
670
+ "epoch": 0.30070377479206656,
671
+ "grad_norm": 6.376151858944639,
672
+ "learning_rate": 1.7652459327577377e-05,
673
+ "loss": 1.5418,
674
+ "step": 470
675
+ },
676
+ {
677
+ "epoch": 0.3039027511196417,
678
+ "grad_norm": 5.100702734889225,
679
+ "learning_rate": 1.7580064240530746e-05,
680
+ "loss": 1.4835,
681
+ "step": 475
682
+ },
683
+ {
684
+ "epoch": 0.30710172744721687,
685
+ "grad_norm": 5.99984161717874,
686
+ "learning_rate": 1.7506723052527243e-05,
687
+ "loss": 1.5465,
688
+ "step": 480
689
+ },
690
+ {
691
+ "epoch": 0.3103007037747921,
692
+ "grad_norm": 6.053485182766263,
693
+ "learning_rate": 1.7432444917601183e-05,
694
+ "loss": 1.5017,
695
+ "step": 485
696
+ },
697
+ {
698
+ "epoch": 0.31349968010236723,
699
+ "grad_norm": 5.6631171211389875,
700
+ "learning_rate": 1.735723910673132e-05,
701
+ "loss": 1.4774,
702
+ "step": 490
703
+ },
704
+ {
705
+ "epoch": 0.31669865642994244,
706
+ "grad_norm": 5.584804357150484,
707
+ "learning_rate": 1.7281115006683687e-05,
708
+ "loss": 1.4851,
709
+ "step": 495
710
+ },
711
+ {
712
+ "epoch": 0.3198976327575176,
713
+ "grad_norm": 6.342021402984585,
714
+ "learning_rate": 1.7204082118840035e-05,
715
+ "loss": 1.4825,
716
+ "step": 500
717
+ },
718
+ {
719
+ "epoch": 0.32309660908509275,
720
+ "grad_norm": 5.579391027670954,
721
+ "learning_rate": 1.712615005801185e-05,
722
+ "loss": 1.472,
723
+ "step": 505
724
+ },
725
+ {
726
+ "epoch": 0.32629558541266795,
727
+ "grad_norm": 5.394518012610248,
728
+ "learning_rate": 1.704732855124037e-05,
729
+ "loss": 1.4966,
730
+ "step": 510
731
+ },
732
+ {
733
+ "epoch": 0.3294945617402431,
734
+ "grad_norm": 5.980215145783222,
735
+ "learning_rate": 1.6967627436582445e-05,
736
+ "loss": 1.4075,
737
+ "step": 515
738
+ },
739
+ {
740
+ "epoch": 0.3326935380678183,
741
+ "grad_norm": 5.808270410776529,
742
+ "learning_rate": 1.6887056661882644e-05,
743
+ "loss": 1.4771,
744
+ "step": 520
745
+ },
746
+ {
747
+ "epoch": 0.33589251439539347,
748
+ "grad_norm": 7.152355179367295,
749
+ "learning_rate": 1.6805626283531592e-05,
750
+ "loss": 1.4722,
751
+ "step": 525
752
+ },
753
+ {
754
+ "epoch": 0.3390914907229686,
755
+ "grad_norm": 6.939335206870156,
756
+ "learning_rate": 1.6723346465210815e-05,
757
+ "loss": 1.5168,
758
+ "step": 530
759
+ },
760
+ {
761
+ "epoch": 0.34229046705054383,
762
+ "grad_norm": 5.552572033728716,
763
+ "learning_rate": 1.6640227476624163e-05,
764
+ "loss": 1.5027,
765
+ "step": 535
766
+ },
767
+ {
768
+ "epoch": 0.345489443378119,
769
+ "grad_norm": 5.492220368766092,
770
+ "learning_rate": 1.655627969221598e-05,
771
+ "loss": 1.4868,
772
+ "step": 540
773
+ },
774
+ {
775
+ "epoch": 0.3486884197056942,
776
+ "grad_norm": 5.994401409437794,
777
+ "learning_rate": 1.6471513589876247e-05,
778
+ "loss": 1.4977,
779
+ "step": 545
780
+ },
781
+ {
782
+ "epoch": 0.35188739603326935,
783
+ "grad_norm": 5.997814866621087,
784
+ "learning_rate": 1.638593974963278e-05,
785
+ "loss": 1.4191,
786
+ "step": 550
787
+ },
788
+ {
789
+ "epoch": 0.3550863723608445,
790
+ "grad_norm": 5.829666218564288,
791
+ "learning_rate": 1.6299568852330703e-05,
792
+ "loss": 1.4315,
793
+ "step": 555
794
+ },
795
+ {
796
+ "epoch": 0.3582853486884197,
797
+ "grad_norm": 5.61328215827393,
798
+ "learning_rate": 1.6212411678299306e-05,
799
+ "loss": 1.5058,
800
+ "step": 560
801
+ },
802
+ {
803
+ "epoch": 0.36148432501599487,
804
+ "grad_norm": 5.327552350734913,
805
+ "learning_rate": 1.612447910600652e-05,
806
+ "loss": 1.4737,
807
+ "step": 565
808
+ },
809
+ {
810
+ "epoch": 0.3646833013435701,
811
+ "grad_norm": 5.69563503987059,
812
+ "learning_rate": 1.6035782110701125e-05,
813
+ "loss": 1.4831,
814
+ "step": 570
815
+ },
816
+ {
817
+ "epoch": 0.36788227767114523,
818
+ "grad_norm": 5.10594196156975,
819
+ "learning_rate": 1.594633176304287e-05,
820
+ "loss": 1.5169,
821
+ "step": 575
822
+ },
823
+ {
824
+ "epoch": 0.3710812539987204,
825
+ "grad_norm": 6.293737941786795,
826
+ "learning_rate": 1.5856139227720714e-05,
827
+ "loss": 1.5339,
828
+ "step": 580
829
+ },
830
+ {
831
+ "epoch": 0.3742802303262956,
832
+ "grad_norm": 5.995452546156939,
833
+ "learning_rate": 1.5765215762059304e-05,
834
+ "loss": 1.4663,
835
+ "step": 585
836
+ },
837
+ {
838
+ "epoch": 0.37747920665387075,
839
+ "grad_norm": 6.0346764414373775,
840
+ "learning_rate": 1.5673572714613886e-05,
841
+ "loss": 1.4116,
842
+ "step": 590
843
+ },
844
+ {
845
+ "epoch": 0.38067818298144596,
846
+ "grad_norm": 6.020240628226962,
847
+ "learning_rate": 1.558122152375387e-05,
848
+ "loss": 1.5372,
849
+ "step": 595
850
+ },
851
+ {
852
+ "epoch": 0.3838771593090211,
853
+ "grad_norm": 6.736733472792626,
854
+ "learning_rate": 1.548817371623513e-05,
855
+ "loss": 1.4495,
856
+ "step": 600
857
+ },
858
+ {
859
+ "epoch": 0.38707613563659626,
860
+ "grad_norm": 22.11767334193536,
861
+ "learning_rate": 1.539444090576132e-05,
862
+ "loss": 1.4788,
863
+ "step": 605
864
+ },
865
+ {
866
+ "epoch": 0.3902751119641715,
867
+ "grad_norm": 5.323640782312709,
868
+ "learning_rate": 1.5300034791534297e-05,
869
+ "loss": 1.4732,
870
+ "step": 610
871
+ },
872
+ {
873
+ "epoch": 0.3934740882917466,
874
+ "grad_norm": 5.3327689435484835,
875
+ "learning_rate": 1.520496715679391e-05,
876
+ "loss": 1.5138,
877
+ "step": 615
878
+ },
879
+ {
880
+ "epoch": 0.39667306461932184,
881
+ "grad_norm": 5.7610778542310985,
882
+ "learning_rate": 1.5109249867347276e-05,
883
+ "loss": 1.4399,
884
+ "step": 620
885
+ },
886
+ {
887
+ "epoch": 0.399872040946897,
888
+ "grad_norm": 5.399701474557813,
889
+ "learning_rate": 1.5012894870087751e-05,
890
+ "loss": 1.3912,
891
+ "step": 625
892
+ },
893
+ {
894
+ "epoch": 0.40307101727447214,
895
+ "grad_norm": 5.480600811287772,
896
+ "learning_rate": 1.4915914191503792e-05,
897
+ "loss": 1.4894,
898
+ "step": 630
899
+ },
900
+ {
901
+ "epoch": 0.40626999360204735,
902
+ "grad_norm": 5.441970253275154,
903
+ "learning_rate": 1.4818319936177885e-05,
904
+ "loss": 1.4936,
905
+ "step": 635
906
+ },
907
+ {
908
+ "epoch": 0.4094689699296225,
909
+ "grad_norm": 6.296326244418185,
910
+ "learning_rate": 1.4720124285275703e-05,
911
+ "loss": 1.4706,
912
+ "step": 640
913
+ },
914
+ {
915
+ "epoch": 0.4126679462571977,
916
+ "grad_norm": 5.893181005202456,
917
+ "learning_rate": 1.4621339495025731e-05,
918
+ "loss": 1.4504,
919
+ "step": 645
920
+ },
921
+ {
922
+ "epoch": 0.41586692258477287,
923
+ "grad_norm": 5.276496357891099,
924
+ "learning_rate": 1.4521977895189518e-05,
925
+ "loss": 1.5316,
926
+ "step": 650
927
+ },
928
+ {
929
+ "epoch": 0.419065898912348,
930
+ "grad_norm": 5.355395453864716,
931
+ "learning_rate": 1.4422051887522735e-05,
932
+ "loss": 1.4182,
933
+ "step": 655
934
+ },
935
+ {
936
+ "epoch": 0.42226487523992323,
937
+ "grad_norm": 5.750131591812706,
938
+ "learning_rate": 1.4321573944227254e-05,
939
+ "loss": 1.4162,
940
+ "step": 660
941
+ },
942
+ {
943
+ "epoch": 0.4254638515674984,
944
+ "grad_norm": 5.6704431690976715,
945
+ "learning_rate": 1.4220556606394465e-05,
946
+ "loss": 1.4115,
947
+ "step": 665
948
+ },
949
+ {
950
+ "epoch": 0.4286628278950736,
951
+ "grad_norm": 5.594312503139955,
952
+ "learning_rate": 1.4119012482439929e-05,
953
+ "loss": 1.4828,
954
+ "step": 670
955
+ },
956
+ {
957
+ "epoch": 0.43186180422264875,
958
+ "grad_norm": 5.501849792689381,
959
+ "learning_rate": 1.4016954246529697e-05,
960
+ "loss": 1.4409,
961
+ "step": 675
962
+ },
963
+ {
964
+ "epoch": 0.4350607805502239,
965
+ "grad_norm": 5.694559255173223,
966
+ "learning_rate": 1.3914394636998374e-05,
967
+ "loss": 1.4415,
968
+ "step": 680
969
+ },
970
+ {
971
+ "epoch": 0.4382597568777991,
972
+ "grad_norm": 5.7273106418478745,
973
+ "learning_rate": 1.3811346454759211e-05,
974
+ "loss": 1.4229,
975
+ "step": 685
976
+ },
977
+ {
978
+ "epoch": 0.44145873320537427,
979
+ "grad_norm": 5.072530356298814,
980
+ "learning_rate": 1.3707822561706336e-05,
981
+ "loss": 1.4428,
982
+ "step": 690
983
+ },
984
+ {
985
+ "epoch": 0.4446577095329495,
986
+ "grad_norm": 5.158739577606318,
987
+ "learning_rate": 1.3603835879109442e-05,
988
+ "loss": 1.409,
989
+ "step": 695
990
+ },
991
+ {
992
+ "epoch": 0.44785668586052463,
993
+ "grad_norm": 6.061233482330587,
994
+ "learning_rate": 1.349939938600099e-05,
995
+ "loss": 1.3806,
996
+ "step": 700
997
+ },
998
+ {
999
+ "epoch": 0.4510556621880998,
1000
+ "grad_norm": 5.707331290817792,
1001
+ "learning_rate": 1.3394526117556277e-05,
1002
+ "loss": 1.3426,
1003
+ "step": 705
1004
+ },
1005
+ {
1006
+ "epoch": 0.454254638515675,
1007
+ "grad_norm": 5.398255971340936,
1008
+ "learning_rate": 1.3289229163466421e-05,
1009
+ "loss": 1.3656,
1010
+ "step": 710
1011
+ },
1012
+ {
1013
+ "epoch": 0.45745361484325014,
1014
+ "grad_norm": 5.2746408411360965,
1015
+ "learning_rate": 1.3183521666304611e-05,
1016
+ "loss": 1.3791,
1017
+ "step": 715
1018
+ },
1019
+ {
1020
+ "epoch": 0.46065259117082535,
1021
+ "grad_norm": 6.106490023435743,
1022
+ "learning_rate": 1.3077416819885707e-05,
1023
+ "loss": 1.4025,
1024
+ "step": 720
1025
+ },
1026
+ {
1027
+ "epoch": 0.4638515674984005,
1028
+ "grad_norm": 9.033703923623596,
1029
+ "learning_rate": 1.297092786761946e-05,
1030
+ "loss": 1.3972,
1031
+ "step": 725
1032
+ },
1033
+ {
1034
+ "epoch": 0.46705054382597566,
1035
+ "grad_norm": 5.203446436783156,
1036
+ "learning_rate": 1.2864068100857565e-05,
1037
+ "loss": 1.4624,
1038
+ "step": 730
1039
+ },
1040
+ {
1041
+ "epoch": 0.47024952015355087,
1042
+ "grad_norm": 5.783521080834399,
1043
+ "learning_rate": 1.2756850857234686e-05,
1044
+ "loss": 1.3224,
1045
+ "step": 735
1046
+ },
1047
+ {
1048
+ "epoch": 0.473448496481126,
1049
+ "grad_norm": 5.1601982559203075,
1050
+ "learning_rate": 1.2649289519003739e-05,
1051
+ "loss": 1.4001,
1052
+ "step": 740
1053
+ },
1054
+ {
1055
+ "epoch": 0.47664747280870123,
1056
+ "grad_norm": 6.162375731265669,
1057
+ "learning_rate": 1.2541397511365584e-05,
1058
+ "loss": 1.3971,
1059
+ "step": 745
1060
+ },
1061
+ {
1062
+ "epoch": 0.4798464491362764,
1063
+ "grad_norm": 5.44449446975744,
1064
+ "learning_rate": 1.2433188300793399e-05,
1065
+ "loss": 1.3678,
1066
+ "step": 750
1067
+ },
1068
+ {
1069
+ "epoch": 0.48304542546385154,
1070
+ "grad_norm": 5.742001125711013,
1071
+ "learning_rate": 1.2324675393351818e-05,
1072
+ "loss": 1.3902,
1073
+ "step": 755
1074
+ },
1075
+ {
1076
+ "epoch": 0.48624440179142675,
1077
+ "grad_norm": 6.588679903273224,
1078
+ "learning_rate": 1.221587233301123e-05,
1079
+ "loss": 1.3967,
1080
+ "step": 760
1081
+ },
1082
+ {
1083
+ "epoch": 0.4894433781190019,
1084
+ "grad_norm": 4.749871970824734,
1085
+ "learning_rate": 1.2106792699957264e-05,
1086
+ "loss": 1.3323,
1087
+ "step": 765
1088
+ },
1089
+ {
1090
+ "epoch": 0.4926423544465771,
1091
+ "grad_norm": 5.241600068506743,
1092
+ "learning_rate": 1.1997450108895807e-05,
1093
+ "loss": 1.3341,
1094
+ "step": 770
1095
+ },
1096
+ {
1097
+ "epoch": 0.49584133077415227,
1098
+ "grad_norm": 6.669804856668056,
1099
+ "learning_rate": 1.1887858207353678e-05,
1100
+ "loss": 1.4422,
1101
+ "step": 775
1102
+ },
1103
+ {
1104
+ "epoch": 0.4990403071017274,
1105
+ "grad_norm": 6.267377222844033,
1106
+ "learning_rate": 1.1778030673975227e-05,
1107
+ "loss": 1.3734,
1108
+ "step": 780
1109
+ },
1110
+ {
1111
+ "epoch": 0.5022392834293026,
1112
+ "grad_norm": 5.387079839641074,
1113
+ "learning_rate": 1.166798121681505e-05,
1114
+ "loss": 1.3142,
1115
+ "step": 785
1116
+ },
1117
+ {
1118
+ "epoch": 0.5054382597568778,
1119
+ "grad_norm": 5.257949564475725,
1120
+ "learning_rate": 1.1557723571627016e-05,
1121
+ "loss": 1.412,
1122
+ "step": 790
1123
+ },
1124
+ {
1125
+ "epoch": 0.508637236084453,
1126
+ "grad_norm": 6.286120326860924,
1127
+ "learning_rate": 1.1447271500149849e-05,
1128
+ "loss": 1.3923,
1129
+ "step": 795
1130
+ },
1131
+ {
1132
+ "epoch": 0.5118362124120281,
1133
+ "grad_norm": 4.8250523003825325,
1134
+ "learning_rate": 1.1336638788389473e-05,
1135
+ "loss": 1.3933,
1136
+ "step": 800
1137
+ },
1138
+ {
1139
+ "epoch": 0.5150351887396033,
1140
+ "grad_norm": 5.309455908621953,
1141
+ "learning_rate": 1.122583924489832e-05,
1142
+ "loss": 1.3419,
1143
+ "step": 805
1144
+ },
1145
+ {
1146
+ "epoch": 0.5182341650671785,
1147
+ "grad_norm": 5.4051345865376845,
1148
+ "learning_rate": 1.1114886699051803e-05,
1149
+ "loss": 1.3372,
1150
+ "step": 810
1151
+ },
1152
+ {
1153
+ "epoch": 0.5214331413947537,
1154
+ "grad_norm": 5.4776811813134385,
1155
+ "learning_rate": 1.1003794999322246e-05,
1156
+ "loss": 1.4053,
1157
+ "step": 815
1158
+ },
1159
+ {
1160
+ "epoch": 0.5246321177223289,
1161
+ "grad_norm": 6.028813751150013,
1162
+ "learning_rate": 1.089257801155037e-05,
1163
+ "loss": 1.2717,
1164
+ "step": 820
1165
+ },
1166
+ {
1167
+ "epoch": 0.527831094049904,
1168
+ "grad_norm": 5.6893885646734725,
1169
+ "learning_rate": 1.0781249617214642e-05,
1170
+ "loss": 1.3647,
1171
+ "step": 825
1172
+ },
1173
+ {
1174
+ "epoch": 0.5310300703774792,
1175
+ "grad_norm": 7.845081962784706,
1176
+ "learning_rate": 1.0669823711698668e-05,
1177
+ "loss": 1.3118,
1178
+ "step": 830
1179
+ },
1180
+ {
1181
+ "epoch": 0.5342290467050543,
1182
+ "grad_norm": 5.974302837219022,
1183
+ "learning_rate": 1.0558314202556866e-05,
1184
+ "loss": 1.3196,
1185
+ "step": 835
1186
+ },
1187
+ {
1188
+ "epoch": 0.5374280230326296,
1189
+ "grad_norm": 6.2079916851397945,
1190
+ "learning_rate": 1.0446735007778563e-05,
1191
+ "loss": 1.3387,
1192
+ "step": 840
1193
+ },
1194
+ {
1195
+ "epoch": 0.5406269993602048,
1196
+ "grad_norm": 5.710214267091366,
1197
+ "learning_rate": 1.0335100054050877e-05,
1198
+ "loss": 1.3401,
1199
+ "step": 845
1200
+ },
1201
+ {
1202
+ "epoch": 0.5438259756877799,
1203
+ "grad_norm": 5.245166732528631,
1204
+ "learning_rate": 1.0223423275020431e-05,
1205
+ "loss": 1.3537,
1206
+ "step": 850
1207
+ },
1208
+ {
1209
+ "epoch": 0.5470249520153551,
1210
+ "grad_norm": 5.47319375038005,
1211
+ "learning_rate": 1.0111718609554254e-05,
1212
+ "loss": 1.3702,
1213
+ "step": 855
1214
+ },
1215
+ {
1216
+ "epoch": 0.5502239283429302,
1217
+ "grad_norm": 5.563044035160415,
1218
+ "learning_rate": 1e-05,
1219
+ "loss": 1.3042,
1220
+ "step": 860
1221
+ },
1222
+ {
1223
+ "epoch": 0.5534229046705055,
1224
+ "grad_norm": 5.32197020170691,
1225
+ "learning_rate": 9.888281390445747e-06,
1226
+ "loss": 1.3702,
1227
+ "step": 865
1228
+ },
1229
+ {
1230
+ "epoch": 0.5566218809980806,
1231
+ "grad_norm": 4.9805710370306935,
1232
+ "learning_rate": 9.776576724979572e-06,
1233
+ "loss": 1.3279,
1234
+ "step": 870
1235
+ },
1236
+ {
1237
+ "epoch": 0.5598208573256558,
1238
+ "grad_norm": 5.517527208389846,
1239
+ "learning_rate": 9.664899945949128e-06,
1240
+ "loss": 1.3328,
1241
+ "step": 875
1242
+ },
1243
+ {
1244
+ "epoch": 0.5630198336532309,
1245
+ "grad_norm": 5.7898575142466955,
1246
+ "learning_rate": 9.553264992221442e-06,
1247
+ "loss": 1.3181,
1248
+ "step": 880
1249
+ },
1250
+ {
1251
+ "epoch": 0.5662188099808061,
1252
+ "grad_norm": 5.573656915945259,
1253
+ "learning_rate": 9.441685797443138e-06,
1254
+ "loss": 1.2587,
1255
+ "step": 885
1256
+ },
1257
+ {
1258
+ "epoch": 0.5694177863083814,
1259
+ "grad_norm": 5.276092647196823,
1260
+ "learning_rate": 9.330176288301332e-06,
1261
+ "loss": 1.2959,
1262
+ "step": 890
1263
+ },
1264
+ {
1265
+ "epoch": 0.5726167626359565,
1266
+ "grad_norm": 5.343314007371001,
1267
+ "learning_rate": 9.21875038278536e-06,
1268
+ "loss": 1.352,
1269
+ "step": 895
1270
+ },
1271
+ {
1272
+ "epoch": 0.5758157389635317,
1273
+ "grad_norm": 5.6426664531262425,
1274
+ "learning_rate": 9.107421988449632e-06,
1275
+ "loss": 1.292,
1276
+ "step": 900
1277
+ },
1278
+ {
1279
+ "epoch": 0.5790147152911068,
1280
+ "grad_norm": 5.640737666551658,
1281
+ "learning_rate": 8.996205000677758e-06,
1282
+ "loss": 1.3715,
1283
+ "step": 905
1284
+ },
1285
+ {
1286
+ "epoch": 0.582213691618682,
1287
+ "grad_norm": 5.9612820213670945,
1288
+ "learning_rate": 8.885113300948199e-06,
1289
+ "loss": 1.2402,
1290
+ "step": 910
1291
+ },
1292
+ {
1293
+ "epoch": 0.5854126679462572,
1294
+ "grad_norm": 5.163441313443216,
1295
+ "learning_rate": 8.774160755101685e-06,
1296
+ "loss": 1.2863,
1297
+ "step": 915
1298
+ },
1299
+ {
1300
+ "epoch": 0.5886116442738324,
1301
+ "grad_norm": 6.145271777985554,
1302
+ "learning_rate": 8.663361211610529e-06,
1303
+ "loss": 1.3555,
1304
+ "step": 920
1305
+ },
1306
+ {
1307
+ "epoch": 0.5918106206014075,
1308
+ "grad_norm": 5.4418317157557405,
1309
+ "learning_rate": 8.552728499850153e-06,
1310
+ "loss": 1.3657,
1311
+ "step": 925
1312
+ },
1313
+ {
1314
+ "epoch": 0.5950095969289827,
1315
+ "grad_norm": 5.467010402521484,
1316
+ "learning_rate": 8.442276428372986e-06,
1317
+ "loss": 1.2939,
1318
+ "step": 930
1319
+ },
1320
+ {
1321
+ "epoch": 0.5982085732565579,
1322
+ "grad_norm": 5.375591925949063,
1323
+ "learning_rate": 8.332018783184952e-06,
1324
+ "loss": 1.3256,
1325
+ "step": 935
1326
+ },
1327
+ {
1328
+ "epoch": 0.6014075495841331,
1329
+ "grad_norm": 5.669641606911128,
1330
+ "learning_rate": 8.221969326024776e-06,
1331
+ "loss": 1.2655,
1332
+ "step": 940
1333
+ },
1334
+ {
1335
+ "epoch": 0.6046065259117083,
1336
+ "grad_norm": 5.107615949778542,
1337
+ "learning_rate": 8.112141792646326e-06,
1338
+ "loss": 1.2487,
1339
+ "step": 945
1340
+ },
1341
+ {
1342
+ "epoch": 0.6078055022392834,
1343
+ "grad_norm": 5.917321765317,
1344
+ "learning_rate": 8.002549891104196e-06,
1345
+ "loss": 1.3936,
1346
+ "step": 950
1347
+ },
1348
+ {
1349
+ "epoch": 0.6110044785668586,
1350
+ "grad_norm": 5.405584578362852,
1351
+ "learning_rate": 7.89320730004274e-06,
1352
+ "loss": 1.3401,
1353
+ "step": 955
1354
+ },
1355
+ {
1356
+ "epoch": 0.6142034548944337,
1357
+ "grad_norm": 5.217914222418895,
1358
+ "learning_rate": 7.784127666988774e-06,
1359
+ "loss": 1.2734,
1360
+ "step": 960
1361
+ },
1362
+ {
1363
+ "epoch": 0.617402431222009,
1364
+ "grad_norm": 5.41366943915775,
1365
+ "learning_rate": 7.675324606648187e-06,
1366
+ "loss": 1.3577,
1367
+ "step": 965
1368
+ },
1369
+ {
1370
+ "epoch": 0.6206014075495841,
1371
+ "grad_norm": 5.613307742444461,
1372
+ "learning_rate": 7.566811699206604e-06,
1373
+ "loss": 1.2305,
1374
+ "step": 970
1375
+ },
1376
+ {
1377
+ "epoch": 0.6238003838771593,
1378
+ "grad_norm": 6.161058161659451,
1379
+ "learning_rate": 7.458602488634416e-06,
1380
+ "loss": 1.2949,
1381
+ "step": 975
1382
+ },
1383
+ {
1384
+ "epoch": 0.6269993602047345,
1385
+ "grad_norm": 5.8815820932534075,
1386
+ "learning_rate": 7.350710480996266e-06,
1387
+ "loss": 1.3216,
1388
+ "step": 980
1389
+ },
1390
+ {
1391
+ "epoch": 0.6301983365323096,
1392
+ "grad_norm": 5.8377871027181625,
1393
+ "learning_rate": 7.243149142765317e-06,
1394
+ "loss": 1.2566,
1395
+ "step": 985
1396
+ },
1397
+ {
1398
+ "epoch": 0.6333973128598849,
1399
+ "grad_norm": 6.007065423720387,
1400
+ "learning_rate": 7.135931899142438e-06,
1401
+ "loss": 1.1938,
1402
+ "step": 990
1403
+ },
1404
+ {
1405
+ "epoch": 0.63659628918746,
1406
+ "grad_norm": 5.665836143013,
1407
+ "learning_rate": 7.029072132380543e-06,
1408
+ "loss": 1.2106,
1409
+ "step": 995
1410
+ },
1411
+ {
1412
+ "epoch": 0.6397952655150352,
1413
+ "grad_norm": 5.976229117153174,
1414
+ "learning_rate": 6.922583180114299e-06,
1415
+ "loss": 1.2839,
1416
+ "step": 1000
1417
+ },
1418
+ {
1419
+ "epoch": 0.6429942418426103,
1420
+ "grad_norm": 6.016028155992671,
1421
+ "learning_rate": 6.816478333695392e-06,
1422
+ "loss": 1.2725,
1423
+ "step": 1005
1424
+ },
1425
+ {
1426
+ "epoch": 0.6461932181701855,
1427
+ "grad_norm": 5.83046161596007,
1428
+ "learning_rate": 6.710770836533584e-06,
1429
+ "loss": 1.2883,
1430
+ "step": 1010
1431
+ },
1432
+ {
1433
+ "epoch": 0.6493921944977608,
1434
+ "grad_norm": 5.2835178908360625,
1435
+ "learning_rate": 6.605473882443725e-06,
1436
+ "loss": 1.286,
1437
+ "step": 1015
1438
+ },
1439
+ {
1440
+ "epoch": 0.6525911708253359,
1441
+ "grad_norm": 5.405038918816778,
1442
+ "learning_rate": 6.5006006139990115e-06,
1443
+ "loss": 1.2575,
1444
+ "step": 1020
1445
+ },
1446
+ {
1447
+ "epoch": 0.6557901471529111,
1448
+ "grad_norm": 5.345367266134731,
1449
+ "learning_rate": 6.396164120890562e-06,
1450
+ "loss": 1.3175,
1451
+ "step": 1025
1452
+ },
1453
+ {
1454
+ "epoch": 0.6589891234804862,
1455
+ "grad_norm": 5.327195492441683,
1456
+ "learning_rate": 6.292177438293665e-06,
1457
+ "loss": 1.2368,
1458
+ "step": 1030
1459
+ },
1460
+ {
1461
+ "epoch": 0.6621880998080614,
1462
+ "grad_norm": 5.429158143917414,
1463
+ "learning_rate": 6.188653545240793e-06,
1464
+ "loss": 1.2618,
1465
+ "step": 1035
1466
+ },
1467
+ {
1468
+ "epoch": 0.6653870761356366,
1469
+ "grad_norm": 9.442946173046769,
1470
+ "learning_rate": 6.085605363001628e-06,
1471
+ "loss": 1.2635,
1472
+ "step": 1040
1473
+ },
1474
+ {
1475
+ "epoch": 0.6685860524632118,
1476
+ "grad_norm": 5.461550264996468,
1477
+ "learning_rate": 5.983045753470308e-06,
1478
+ "loss": 1.2292,
1479
+ "step": 1045
1480
+ },
1481
+ {
1482
+ "epoch": 0.6717850287907869,
1483
+ "grad_norm": 5.433865692907909,
1484
+ "learning_rate": 5.880987517560075e-06,
1485
+ "loss": 1.2533,
1486
+ "step": 1050
1487
+ },
1488
+ {
1489
+ "epoch": 0.6749840051183621,
1490
+ "grad_norm": 5.360876312233277,
1491
+ "learning_rate": 5.779443393605536e-06,
1492
+ "loss": 1.2302,
1493
+ "step": 1055
1494
+ },
1495
+ {
1496
+ "epoch": 0.6781829814459372,
1497
+ "grad_norm": 5.597034291592108,
1498
+ "learning_rate": 5.678426055772746e-06,
1499
+ "loss": 1.2621,
1500
+ "step": 1060
1501
+ },
1502
+ {
1503
+ "epoch": 0.6813819577735125,
1504
+ "grad_norm": 5.364563467233611,
1505
+ "learning_rate": 5.577948112477271e-06,
1506
+ "loss": 1.2081,
1507
+ "step": 1065
1508
+ },
1509
+ {
1510
+ "epoch": 0.6845809341010877,
1511
+ "grad_norm": 5.479671671356431,
1512
+ "learning_rate": 5.478022104810483e-06,
1513
+ "loss": 1.2343,
1514
+ "step": 1070
1515
+ },
1516
+ {
1517
+ "epoch": 0.6877799104286628,
1518
+ "grad_norm": 5.42208058407067,
1519
+ "learning_rate": 5.378660504974271e-06,
1520
+ "loss": 1.1894,
1521
+ "step": 1075
1522
+ },
1523
+ {
1524
+ "epoch": 0.690978886756238,
1525
+ "grad_norm": 5.318236870909474,
1526
+ "learning_rate": 5.2798757147242986e-06,
1527
+ "loss": 1.2133,
1528
+ "step": 1080
1529
+ },
1530
+ {
1531
+ "epoch": 0.6941778630838131,
1532
+ "grad_norm": 7.743258878956998,
1533
+ "learning_rate": 5.1816800638221176e-06,
1534
+ "loss": 1.2861,
1535
+ "step": 1085
1536
+ },
1537
+ {
1538
+ "epoch": 0.6973768394113884,
1539
+ "grad_norm": 5.071086020945074,
1540
+ "learning_rate": 5.084085808496211e-06,
1541
+ "loss": 1.1672,
1542
+ "step": 1090
1543
+ },
1544
+ {
1545
+ "epoch": 0.7005758157389635,
1546
+ "grad_norm": 5.938216173933617,
1547
+ "learning_rate": 4.987105129912252e-06,
1548
+ "loss": 1.2525,
1549
+ "step": 1095
1550
+ },
1551
+ {
1552
+ "epoch": 0.7037747920665387,
1553
+ "grad_norm": 5.14560931343335,
1554
+ "learning_rate": 4.890750132652724e-06,
1555
+ "loss": 1.2527,
1556
+ "step": 1100
1557
+ },
1558
+ {
1559
+ "epoch": 0.7069737683941139,
1560
+ "grad_norm": 5.824953341068154,
1561
+ "learning_rate": 4.795032843206091e-06,
1562
+ "loss": 1.2699,
1563
+ "step": 1105
1564
+ },
1565
+ {
1566
+ "epoch": 0.710172744721689,
1567
+ "grad_norm": 5.823852748558207,
1568
+ "learning_rate": 4.699965208465702e-06,
1569
+ "loss": 1.2413,
1570
+ "step": 1110
1571
+ },
1572
+ {
1573
+ "epoch": 0.7133717210492643,
1574
+ "grad_norm": 6.159332363176924,
1575
+ "learning_rate": 4.605559094238681e-06,
1576
+ "loss": 1.2449,
1577
+ "step": 1115
1578
+ },
1579
+ {
1580
+ "epoch": 0.7165706973768394,
1581
+ "grad_norm": 5.491076270333848,
1582
+ "learning_rate": 4.511826283764872e-06,
1583
+ "loss": 1.2375,
1584
+ "step": 1120
1585
+ },
1586
+ {
1587
+ "epoch": 0.7197696737044146,
1588
+ "grad_norm": 5.998578981936084,
1589
+ "learning_rate": 4.418778476246132e-06,
1590
+ "loss": 1.2537,
1591
+ "step": 1125
1592
+ },
1593
+ {
1594
+ "epoch": 0.7229686500319897,
1595
+ "grad_norm": 5.387463063273167,
1596
+ "learning_rate": 4.326427285386118e-06,
1597
+ "loss": 1.1868,
1598
+ "step": 1130
1599
+ },
1600
+ {
1601
+ "epoch": 0.7261676263595649,
1602
+ "grad_norm": 5.769559852193052,
1603
+ "learning_rate": 4.234784237940705e-06,
1604
+ "loss": 1.2264,
1605
+ "step": 1135
1606
+ },
1607
+ {
1608
+ "epoch": 0.7293666026871402,
1609
+ "grad_norm": 5.68912164929325,
1610
+ "learning_rate": 4.143860772279287e-06,
1611
+ "loss": 1.1669,
1612
+ "step": 1140
1613
+ },
1614
+ {
1615
+ "epoch": 0.7325655790147153,
1616
+ "grad_norm": 6.169259924681037,
1617
+ "learning_rate": 4.053668236957135e-06,
1618
+ "loss": 1.188,
1619
+ "step": 1145
1620
+ },
1621
+ {
1622
+ "epoch": 0.7357645553422905,
1623
+ "grad_norm": 5.775273351173123,
1624
+ "learning_rate": 3.964217889298882e-06,
1625
+ "loss": 1.2167,
1626
+ "step": 1150
1627
+ },
1628
+ {
1629
+ "epoch": 0.7389635316698656,
1630
+ "grad_norm": 6.66491466045373,
1631
+ "learning_rate": 3.875520893993482e-06,
1632
+ "loss": 1.2407,
1633
+ "step": 1155
1634
+ },
1635
+ {
1636
+ "epoch": 0.7421625079974408,
1637
+ "grad_norm": 5.4584893738231814,
1638
+ "learning_rate": 3.787588321700697e-06,
1639
+ "loss": 1.1773,
1640
+ "step": 1160
1641
+ },
1642
+ {
1643
+ "epoch": 0.745361484325016,
1644
+ "grad_norm": 5.495714202965497,
1645
+ "learning_rate": 3.7004311476692977e-06,
1646
+ "loss": 1.263,
1647
+ "step": 1165
1648
+ },
1649
+ {
1650
+ "epoch": 0.7485604606525912,
1651
+ "grad_norm": 5.367093577033402,
1652
+ "learning_rate": 3.6140602503672217e-06,
1653
+ "loss": 1.1816,
1654
+ "step": 1170
1655
+ },
1656
+ {
1657
+ "epoch": 0.7517594369801663,
1658
+ "grad_norm": 5.929562727552283,
1659
+ "learning_rate": 3.528486410123758e-06,
1660
+ "loss": 1.287,
1661
+ "step": 1175
1662
+ },
1663
+ {
1664
+ "epoch": 0.7549584133077415,
1665
+ "grad_norm": 5.16200286713459,
1666
+ "learning_rate": 3.443720307784022e-06,
1667
+ "loss": 1.1518,
1668
+ "step": 1180
1669
+ },
1670
+ {
1671
+ "epoch": 0.7581573896353166,
1672
+ "grad_norm": 5.617358144040829,
1673
+ "learning_rate": 3.359772523375837e-06,
1674
+ "loss": 1.126,
1675
+ "step": 1185
1676
+ },
1677
+ {
1678
+ "epoch": 0.7613563659628919,
1679
+ "grad_norm": 5.274480247725789,
1680
+ "learning_rate": 3.276653534789185e-06,
1681
+ "loss": 1.1226,
1682
+ "step": 1190
1683
+ },
1684
+ {
1685
+ "epoch": 0.7645553422904671,
1686
+ "grad_norm": 5.087961244651565,
1687
+ "learning_rate": 3.1943737164684094e-06,
1688
+ "loss": 1.1843,
1689
+ "step": 1195
1690
+ },
1691
+ {
1692
+ "epoch": 0.7677543186180422,
1693
+ "grad_norm": 5.165537670382836,
1694
+ "learning_rate": 3.11294333811736e-06,
1695
+ "loss": 1.2111,
1696
+ "step": 1200
1697
+ },
1698
+ {
1699
+ "epoch": 0.7709532949456174,
1700
+ "grad_norm": 5.141713986869319,
1701
+ "learning_rate": 3.032372563417556e-06,
1702
+ "loss": 1.1842,
1703
+ "step": 1205
1704
+ },
1705
+ {
1706
+ "epoch": 0.7741522712731925,
1707
+ "grad_norm": 5.0964117361663455,
1708
+ "learning_rate": 2.952671448759631e-06,
1709
+ "loss": 1.2247,
1710
+ "step": 1210
1711
+ },
1712
+ {
1713
+ "epoch": 0.7773512476007678,
1714
+ "grad_norm": 5.3243216832528715,
1715
+ "learning_rate": 2.8738499419881517e-06,
1716
+ "loss": 1.1762,
1717
+ "step": 1215
1718
+ },
1719
+ {
1720
+ "epoch": 0.780550223928343,
1721
+ "grad_norm": 4.851797186739384,
1722
+ "learning_rate": 2.795917881159973e-06,
1723
+ "loss": 1.1502,
1724
+ "step": 1220
1725
+ },
1726
+ {
1727
+ "epoch": 0.7837492002559181,
1728
+ "grad_norm": 5.496612395138414,
1729
+ "learning_rate": 2.718884993316311e-06,
1730
+ "loss": 1.1772,
1731
+ "step": 1225
1732
+ },
1733
+ {
1734
+ "epoch": 0.7869481765834933,
1735
+ "grad_norm": 5.442736560179264,
1736
+ "learning_rate": 2.642760893268684e-06,
1737
+ "loss": 1.1553,
1738
+ "step": 1230
1739
+ },
1740
+ {
1741
+ "epoch": 0.7901471529110684,
1742
+ "grad_norm": 5.17508160986582,
1743
+ "learning_rate": 2.5675550823988184e-06,
1744
+ "loss": 1.156,
1745
+ "step": 1235
1746
+ },
1747
+ {
1748
+ "epoch": 0.7933461292386437,
1749
+ "grad_norm": 5.549015221247355,
1750
+ "learning_rate": 2.493276947472756e-06,
1751
+ "loss": 1.1863,
1752
+ "step": 1240
1753
+ },
1754
+ {
1755
+ "epoch": 0.7965451055662188,
1756
+ "grad_norm": 5.482924269402519,
1757
+ "learning_rate": 2.4199357594692564e-06,
1758
+ "loss": 1.1136,
1759
+ "step": 1245
1760
+ },
1761
+ {
1762
+ "epoch": 0.799744081893794,
1763
+ "grad_norm": 6.019191778587459,
1764
+ "learning_rate": 2.347540672422627e-06,
1765
+ "loss": 1.1587,
1766
+ "step": 1250
1767
+ },
1768
+ {
1769
+ "epoch": 0.8029430582213691,
1770
+ "grad_norm": 5.49048264574569,
1771
+ "learning_rate": 2.2761007222801866e-06,
1772
+ "loss": 1.1607,
1773
+ "step": 1255
1774
+ },
1775
+ {
1776
+ "epoch": 0.8061420345489443,
1777
+ "grad_norm": 5.16885210872782,
1778
+ "learning_rate": 2.2056248257744383e-06,
1779
+ "loss": 1.1408,
1780
+ "step": 1260
1781
+ },
1782
+ {
1783
+ "epoch": 0.8093410108765196,
1784
+ "grad_norm": 5.432012432577872,
1785
+ "learning_rate": 2.136121779310132e-06,
1786
+ "loss": 1.2092,
1787
+ "step": 1265
1788
+ },
1789
+ {
1790
+ "epoch": 0.8125399872040947,
1791
+ "grad_norm": 5.406927709627139,
1792
+ "learning_rate": 2.067600257866358e-06,
1793
+ "loss": 1.1709,
1794
+ "step": 1270
1795
+ },
1796
+ {
1797
+ "epoch": 0.8157389635316699,
1798
+ "grad_norm": 5.141882407100849,
1799
+ "learning_rate": 2.000068813913777e-06,
1800
+ "loss": 1.0669,
1801
+ "step": 1275
1802
+ },
1803
+ {
1804
+ "epoch": 0.818937939859245,
1805
+ "grad_norm": 5.499076743202356,
1806
+ "learning_rate": 1.933535876347141e-06,
1807
+ "loss": 1.1749,
1808
+ "step": 1280
1809
+ },
1810
+ {
1811
+ "epoch": 0.8221369161868202,
1812
+ "grad_norm": 5.475323870607504,
1813
+ "learning_rate": 1.8680097494332682e-06,
1814
+ "loss": 1.0934,
1815
+ "step": 1285
1816
+ },
1817
+ {
1818
+ "epoch": 0.8253358925143954,
1819
+ "grad_norm": 5.391123200327997,
1820
+ "learning_rate": 1.8034986117745245e-06,
1821
+ "loss": 1.1402,
1822
+ "step": 1290
1823
+ },
1824
+ {
1825
+ "epoch": 0.8285348688419706,
1826
+ "grad_norm": 5.348163069842104,
1827
+ "learning_rate": 1.7400105152880331e-06,
1828
+ "loss": 1.1806,
1829
+ "step": 1295
1830
+ },
1831
+ {
1832
+ "epoch": 0.8317338451695457,
1833
+ "grad_norm": 5.699253380097164,
1834
+ "learning_rate": 1.6775533842006853e-06,
1835
+ "loss": 1.0805,
1836
+ "step": 1300
1837
+ },
1838
+ {
1839
+ "epoch": 0.8349328214971209,
1840
+ "grad_norm": 6.0098839122086005,
1841
+ "learning_rate": 1.6161350140600606e-06,
1842
+ "loss": 1.0664,
1843
+ "step": 1305
1844
+ },
1845
+ {
1846
+ "epoch": 0.838131797824696,
1847
+ "grad_norm": 5.119178144887819,
1848
+ "learning_rate": 1.555763070761448e-06,
1849
+ "loss": 1.16,
1850
+ "step": 1310
1851
+ },
1852
+ {
1853
+ "epoch": 0.8413307741522713,
1854
+ "grad_norm": 5.82644856476954,
1855
+ "learning_rate": 1.496445089591031e-06,
1856
+ "loss": 1.2315,
1857
+ "step": 1315
1858
+ },
1859
+ {
1860
+ "epoch": 0.8445297504798465,
1861
+ "grad_norm": 5.272508975962535,
1862
+ "learning_rate": 1.4381884742853723e-06,
1863
+ "loss": 1.1552,
1864
+ "step": 1320
1865
+ },
1866
+ {
1867
+ "epoch": 0.8477287268074216,
1868
+ "grad_norm": 5.636426941111563,
1869
+ "learning_rate": 1.381000496107313e-06,
1870
+ "loss": 1.1839,
1871
+ "step": 1325
1872
+ },
1873
+ {
1874
+ "epoch": 0.8509277031349968,
1875
+ "grad_norm": 5.298864794368035,
1876
+ "learning_rate": 1.3248882929384321e-06,
1877
+ "loss": 1.1189,
1878
+ "step": 1330
1879
+ },
1880
+ {
1881
+ "epoch": 0.8541266794625719,
1882
+ "grad_norm": 6.334207459793715,
1883
+ "learning_rate": 1.2698588683881185e-06,
1884
+ "loss": 1.1156,
1885
+ "step": 1335
1886
+ },
1887
+ {
1888
+ "epoch": 0.8573256557901472,
1889
+ "grad_norm": 5.085204432451678,
1890
+ "learning_rate": 1.215919090919434e-06,
1891
+ "loss": 1.1514,
1892
+ "step": 1340
1893
+ },
1894
+ {
1895
+ "epoch": 0.8605246321177223,
1896
+ "grad_norm": 5.519657129601866,
1897
+ "learning_rate": 1.1630756929918218e-06,
1898
+ "loss": 1.1383,
1899
+ "step": 1345
1900
+ },
1901
+ {
1902
+ "epoch": 0.8637236084452975,
1903
+ "grad_norm": 5.64198104430077,
1904
+ "learning_rate": 1.111335270220798e-06,
1905
+ "loss": 1.2014,
1906
+ "step": 1350
1907
+ },
1908
+ {
1909
+ "epoch": 0.8669225847728727,
1910
+ "grad_norm": 5.639338014653528,
1911
+ "learning_rate": 1.060704280554733e-06,
1912
+ "loss": 1.1162,
1913
+ "step": 1355
1914
+ },
1915
+ {
1916
+ "epoch": 0.8701215611004478,
1917
+ "grad_norm": 6.2590061059224755,
1918
+ "learning_rate": 1.0111890434688009e-06,
1919
+ "loss": 1.1873,
1920
+ "step": 1360
1921
+ },
1922
+ {
1923
+ "epoch": 0.8733205374280231,
1924
+ "grad_norm": 5.764812667567324,
1925
+ "learning_rate": 9.62795739176212e-07,
1926
+ "loss": 1.1753,
1927
+ "step": 1365
1928
+ },
1929
+ {
1930
+ "epoch": 0.8765195137555982,
1931
+ "grad_norm": 5.468694265932519,
1932
+ "learning_rate": 9.155304078568495e-07,
1933
+ "loss": 1.1844,
1934
+ "step": 1370
1935
+ },
1936
+ {
1937
+ "epoch": 0.8797184900831734,
1938
+ "grad_norm": 6.144081430233088,
1939
+ "learning_rate": 8.693989489033494e-07,
1940
+ "loss": 1.223,
1941
+ "step": 1375
1942
+ },
1943
+ {
1944
+ "epoch": 0.8829174664107485,
1945
+ "grad_norm": 5.329648133115913,
1946
+ "learning_rate": 8.244071201847826e-07,
1947
+ "loss": 1.0919,
1948
+ "step": 1380
1949
+ },
1950
+ {
1951
+ "epoch": 0.8861164427383237,
1952
+ "grad_norm": 5.392201352239966,
1953
+ "learning_rate": 7.805605373279934e-07,
1954
+ "loss": 1.1712,
1955
+ "step": 1385
1956
+ },
1957
+ {
1958
+ "epoch": 0.889315419065899,
1959
+ "grad_norm": 6.36606850342023,
1960
+ "learning_rate": 7.378646730166783e-07,
1961
+ "loss": 1.0977,
1962
+ "step": 1390
1963
+ },
1964
+ {
1965
+ "epoch": 0.8925143953934741,
1966
+ "grad_norm": 5.577850513924324,
1967
+ "learning_rate": 6.963248563083203e-07,
1968
+ "loss": 1.1191,
1969
+ "step": 1395
1970
+ },
1971
+ {
1972
+ "epoch": 0.8957133717210493,
1973
+ "grad_norm": 5.78598403542479,
1974
+ "learning_rate": 6.559462719690501e-07,
1975
+ "loss": 1.1134,
1976
+ "step": 1400
1977
+ },
1978
+ {
1979
+ "epoch": 0.8989123480486244,
1980
+ "grad_norm": 5.3934506662058626,
1981
+ "learning_rate": 6.167339598265109e-07,
1982
+ "loss": 1.1181,
1983
+ "step": 1405
1984
+ },
1985
+ {
1986
+ "epoch": 0.9021113243761996,
1987
+ "grad_norm": 5.236856333816423,
1988
+ "learning_rate": 5.78692814140811e-07,
1989
+ "loss": 1.1359,
1990
+ "step": 1410
1991
+ },
1992
+ {
1993
+ "epoch": 0.9053103007037748,
1994
+ "grad_norm": 5.224339666624607,
1995
+ "learning_rate": 5.418275829936537e-07,
1996
+ "loss": 1.079,
1997
+ "step": 1415
1998
+ },
1999
+ {
2000
+ "epoch": 0.90850927703135,
2001
+ "grad_norm": 5.997958861887249,
2002
+ "learning_rate": 5.06142867695708e-07,
2003
+ "loss": 1.1306,
2004
+ "step": 1420
2005
+ },
2006
+ {
2007
+ "epoch": 0.9117082533589251,
2008
+ "grad_norm": 7.257431339690822,
2009
+ "learning_rate": 4.716431222122997e-07,
2010
+ "loss": 1.202,
2011
+ "step": 1425
2012
+ },
2013
+ {
2014
+ "epoch": 0.9149072296865003,
2015
+ "grad_norm": 5.409318506244401,
2016
+ "learning_rate": 4.3833265260749157e-07,
2017
+ "loss": 1.1247,
2018
+ "step": 1430
2019
+ },
2020
+ {
2021
+ "epoch": 0.9181062060140754,
2022
+ "grad_norm": 5.66599799044067,
2023
+ "learning_rate": 4.062156165066211e-07,
2024
+ "loss": 1.2072,
2025
+ "step": 1435
2026
+ },
2027
+ {
2028
+ "epoch": 0.9213051823416507,
2029
+ "grad_norm": 5.1071218474615785,
2030
+ "learning_rate": 3.752960225773772e-07,
2031
+ "loss": 1.1526,
2032
+ "step": 1440
2033
+ },
2034
+ {
2035
+ "epoch": 0.9245041586692259,
2036
+ "grad_norm": 5.612505718218031,
2037
+ "learning_rate": 3.4557773002945607e-07,
2038
+ "loss": 1.1799,
2039
+ "step": 1445
2040
+ },
2041
+ {
2042
+ "epoch": 0.927703134996801,
2043
+ "grad_norm": 5.569758076359398,
2044
+ "learning_rate": 3.170644481328711e-07,
2045
+ "loss": 1.1338,
2046
+ "step": 1450
2047
+ },
2048
+ {
2049
+ "epoch": 0.9309021113243762,
2050
+ "grad_norm": 5.150978620117567,
2051
+ "learning_rate": 2.8975973575499526e-07,
2052
+ "loss": 1.0794,
2053
+ "step": 1455
2054
+ },
2055
+ {
2056
+ "epoch": 0.9341010876519513,
2057
+ "grad_norm": 5.318105176276369,
2058
+ "learning_rate": 2.636670009163522e-07,
2059
+ "loss": 1.1089,
2060
+ "step": 1460
2061
+ },
2062
+ {
2063
+ "epoch": 0.9373000639795266,
2064
+ "grad_norm": 5.272467269890987,
2065
+ "learning_rate": 2.3878950036524963e-07,
2066
+ "loss": 1.0252,
2067
+ "step": 1465
2068
+ },
2069
+ {
2070
+ "epoch": 0.9404990403071017,
2071
+ "grad_norm": 5.374330955578694,
2072
+ "learning_rate": 2.1513033917129334e-07,
2073
+ "loss": 1.1165,
2074
+ "step": 1470
2075
+ },
2076
+ {
2077
+ "epoch": 0.9436980166346769,
2078
+ "grad_norm": 5.901734563944482,
2079
+ "learning_rate": 1.9269247033782744e-07,
2080
+ "loss": 1.0798,
2081
+ "step": 1475
2082
+ },
2083
+ {
2084
+ "epoch": 0.946896992962252,
2085
+ "grad_norm": 5.725259152435928,
2086
+ "learning_rate": 1.7147869443335463e-07,
2087
+ "loss": 1.0879,
2088
+ "step": 1480
2089
+ },
2090
+ {
2091
+ "epoch": 0.9500959692898272,
2092
+ "grad_norm": 6.531843733416842,
2093
+ "learning_rate": 1.5149165924199016e-07,
2094
+ "loss": 1.1457,
2095
+ "step": 1485
2096
+ },
2097
+ {
2098
+ "epoch": 0.9532949456174025,
2099
+ "grad_norm": 5.482485131721133,
2100
+ "learning_rate": 1.3273385943297746e-07,
2101
+ "loss": 1.0818,
2102
+ "step": 1490
2103
+ },
2104
+ {
2105
+ "epoch": 0.9564939219449776,
2106
+ "grad_norm": 5.710529768741672,
2107
+ "learning_rate": 1.1520763624931597e-07,
2108
+ "loss": 1.0723,
2109
+ "step": 1495
2110
+ },
2111
+ {
2112
+ "epoch": 0.9596928982725528,
2113
+ "grad_norm": 6.2511499196111195,
2114
+ "learning_rate": 9.891517721554499e-08,
2115
+ "loss": 1.2099,
2116
+ "step": 1500
2117
+ },
2118
+ {
2119
+ "epoch": 0.9628918746001279,
2120
+ "grad_norm": 6.188073654950074,
2121
+ "learning_rate": 8.385851586470318e-08,
2122
+ "loss": 1.1173,
2123
+ "step": 1505
2124
+ },
2125
+ {
2126
+ "epoch": 0.9660908509277031,
2127
+ "grad_norm": 5.170348272609539,
2128
+ "learning_rate": 7.003953148452036e-08,
2129
+ "loss": 1.1465,
2130
+ "step": 1510
2131
+ },
2132
+ {
2133
+ "epoch": 0.9692898272552783,
2134
+ "grad_norm": 5.49895966704058,
2135
+ "learning_rate": 5.745994888285311e-08,
2136
+ "loss": 1.0866,
2137
+ "step": 1515
2138
+ },
2139
+ {
2140
+ "epoch": 0.9724888035828535,
2141
+ "grad_norm": 5.862183299505906,
2142
+ "learning_rate": 4.612133817239905e-08,
2143
+ "loss": 1.1478,
2144
+ "step": 1520
2145
+ },
2146
+ {
2147
+ "epoch": 0.9756877799104287,
2148
+ "grad_norm": 5.522072674954621,
2149
+ "learning_rate": 3.602511457473479e-08,
2150
+ "loss": 1.0727,
2151
+ "step": 1525
2152
+ },
2153
+ {
2154
+ "epoch": 0.9788867562380038,
2155
+ "grad_norm": 5.9700737220751146,
2156
+ "learning_rate": 2.7172538243666057e-08,
2157
+ "loss": 1.2276,
2158
+ "step": 1530
2159
+ },
2160
+ {
2161
+ "epoch": 0.982085732565579,
2162
+ "grad_norm": 5.812591223408153,
2163
+ "learning_rate": 1.9564714107945804e-08,
2164
+ "loss": 1.2139,
2165
+ "step": 1535
2166
+ },
2167
+ {
2168
+ "epoch": 0.9852847088931542,
2169
+ "grad_norm": 5.401482512458415,
2170
+ "learning_rate": 1.3202591733365577e-08,
2171
+ "loss": 1.1062,
2172
+ "step": 1540
2173
+ },
2174
+ {
2175
+ "epoch": 0.9884836852207294,
2176
+ "grad_norm": 5.685685800297117,
2177
+ "learning_rate": 8.086965204233688e-09,
2178
+ "loss": 1.1616,
2179
+ "step": 1545
2180
+ },
2181
+ {
2182
+ "epoch": 0.9916826615483045,
2183
+ "grad_norm": 5.825452202788891,
2184
+ "learning_rate": 4.218473024261149e-09,
2185
+ "loss": 1.1341,
2186
+ "step": 1550
2187
+ },
2188
+ {
2189
+ "epoch": 0.9948816378758797,
2190
+ "grad_norm": 5.673198190032165,
2191
+ "learning_rate": 1.5975980368709843e-09,
2192
+ "loss": 1.1724,
2193
+ "step": 1555
2194
+ },
2195
+ {
2196
+ "epoch": 0.9980806142034548,
2197
+ "grad_norm": 5.554210422341384,
2198
+ "learning_rate": 2.2466736492643416e-10,
2199
+ "loss": 1.0668,
2200
+ "step": 1560
2201
+ },
2202
+ {
2203
+ "epoch": 1.0,
2204
+ "eval_loss": 1.142279028892517,
2205
+ "eval_runtime": 20.0059,
2206
+ "eval_samples_per_second": 24.993,
2207
+ "eval_steps_per_second": 0.8,
2208
+ "step": 1563
2209
+ },
2210
+ {
2211
+ "epoch": 1.0,
2212
+ "step": 1563,
2213
+ "total_flos": 38150611599360.0,
2214
+ "train_loss": 1.3537356438578816,
2215
+ "train_runtime": 6160.8214,
2216
+ "train_samples_per_second": 8.116,
2217
+ "train_steps_per_second": 0.254
2218
+ }
2219
+ ],
2220
+ "logging_steps": 5,
2221
+ "max_steps": 1563,
2222
+ "num_input_tokens_seen": 0,
2223
+ "num_train_epochs": 1,
2224
+ "save_steps": 500,
2225
+ "stateful_callbacks": {
2226
+ "TrainerControl": {
2227
+ "args": {
2228
+ "should_epoch_stop": false,
2229
+ "should_evaluate": false,
2230
+ "should_log": false,
2231
+ "should_save": true,
2232
+ "should_training_stop": true
2233
+ },
2234
+ "attributes": {}
2235
+ }
2236
+ },
2237
+ "total_flos": 38150611599360.0,
2238
+ "train_batch_size": 8,
2239
+ "trial_name": null,
2240
+ "trial_params": null
2241
+ }