Hugging Face's logo

shujatoor
/
phi3nedtuned-ner

PEFT
TensorBoard
Safetensors
trl
sft
Generated from Trainer
Model card Files Metrics Community
shujatoor commited on
Commit
8b2cbed
•
1 Parent(s): 9002da4

shujatoor/phi3nedtuned-ner

Browse files
Files changed (15) hide show
  1. README.md +61 -0
  2. adapter_config.json +31 -0
  3. adapter_model.safetensors +3 -0
  4. added_tokens.json +13 -0
  5. all_results.json +13 -0
  6. eval_results.json +8 -0
  7. runs/Apr30_09-20-43_imran-Precision-Tower-7910/events.out.tfevents.1714450876.imran-Precision-Tower-7910.1152753.0 +3 -0
  8. runs/Apr30_09-20-43_imran-Precision-Tower-7910/events.out.tfevents.1714458463.imran-Precision-Tower-7910.1152753.1 +3 -0
  9. special_tokens_map.json +24 -0
  10. tokenizer.json +0 -0
  11. tokenizer.model +3 -0
  12. tokenizer_config.json +129 -0
  13. train_results.json +8 -0
  14. trainer_state.json +2982 -0
  15. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ library_name: peft
4
+ tags:
5
+ - trl
6
+ - sft
7
+ - generated_from_trainer
8
+ base_model: microsoft/Phi-3-mini-4k-instruct
9
+ datasets:
10
+ - generator
11
+ model-index:
12
+ - name: checkpoint_dir
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
+ # checkpoint_dir
20
+
21
+ This model is a fine-tuned version of [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) on the generator dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 0.6568
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
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 0.0002
43
+ - train_batch_size: 1
44
+ - eval_batch_size: 1
45
+ - seed: 0
46
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
+ - lr_scheduler_type: cosine
48
+ - lr_scheduler_warmup_ratio: 0.2
49
+ - num_epochs: 1
50
+
51
+ ### Training results
52
+
53
+
54
+
55
+ ### Framework versions
56
+
57
+ - PEFT 0.10.1.dev0
58
+ - Transformers 4.41.0.dev0
59
+ - Pytorch 2.2.1+cu121
60
+ - Datasets 2.19.0
61
+ - Tokenizers 0.19.1
adapter_config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "microsoft/Phi-3-mini-4k-instruct",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.05,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 16,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "o_proj",
24
+ "gate_up_proj",
25
+ "qkv_proj",
26
+ "down_proj"
27
+ ],
28
+ "task_type": "CAUSAL_LM",
29
+ "use_dora": false,
30
+ "use_rslora": false
31
+ }
adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3fc636c1af587c55318e1bf4989e969194964b4508b3b645f69f428a82016be5
3
+ size 50366280
added_tokens.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "<|assistant|>": 32001,
3
+ "<|endoftext|>": 32000,
4
+ "<|end|>": 32007,
5
+ "<|placeholder1|>": 32002,
6
+ "<|placeholder2|>": 32003,
7
+ "<|placeholder3|>": 32004,
8
+ "<|placeholder4|>": 32005,
9
+ "<|placeholder5|>": 32008,
10
+ "<|placeholder6|>": 32009,
11
+ "<|system|>": 32006,
12
+ "<|user|>": 32010
13
+ }
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "eval_loss": 0.656758725643158,
4
+ "eval_runtime": 267.6779,
5
+ "eval_samples": 1638,
6
+ "eval_samples_per_second": 3.848,
7
+ "eval_steps_per_second": 3.848,
8
+ "total_flos": 9.688250026819584e+16,
9
+ "train_loss": 1.1022938394574842,
10
+ "train_runtime": 7319.3616,
11
+ "train_samples_per_second": 1.15,
12
+ "train_steps_per_second": 1.15
13
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "eval_loss": 0.656758725643158,
4
+ "eval_runtime": 267.6779,
5
+ "eval_samples": 1638,
6
+ "eval_samples_per_second": 3.848,
7
+ "eval_steps_per_second": 3.848
8
+ }
runs/Apr30_09-20-43_imran-Precision-Tower-7910/events.out.tfevents.1714450876.imran-Precision-Tower-7910.1152753.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6f13607caf902c715e35f9f984f959d69bad484b9c865f4fca23dbb47b1f91fb
3
+ size 94097
runs/Apr30_09-20-43_imran-Precision-Tower-7910/events.out.tfevents.1714458463.imran-Precision-Tower-7910.1152753.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:351dbbf9f68354de3f77f942ddc1d8674abeba8d5841ce6c3abe13ca4ca1e61e
3
+ size 359
special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|endoftext|>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<unk>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "added_tokens_decoder": {
5
+ "0": {
6
+ "content": "<unk>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "1": {
14
+ "content": "<s>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "2": {
22
+ "content": "</s>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "32000": {
30
+ "content": "<|endoftext|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "32001": {
38
+ "content": "<|assistant|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "32002": {
46
+ "content": "<|placeholder1|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "32003": {
54
+ "content": "<|placeholder2|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "32004": {
62
+ "content": "<|placeholder3|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "32005": {
70
+ "content": "<|placeholder4|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "32006": {
78
+ "content": "<|system|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "32007": {
86
+ "content": "<|end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "32008": {
94
+ "content": "<|placeholder5|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "32009": {
102
+ "content": "<|placeholder6|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "32010": {
110
+ "content": "<|user|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ }
117
+ },
118
+ "bos_token": "<s>",
119
+ "chat_template": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') %}{{'<|user|>' + '\n' + message['content'] + '<|end|>' + '\n' + '<|assistant|>' + '\n'}}{% elif (message['role'] == 'assistant') %}{{message['content'] + '<|end|>' + '\n'}}{% endif %}{% endfor %}",
120
+ "clean_up_tokenization_spaces": false,
121
+ "eos_token": "<|endoftext|>",
122
+ "model_max_length": 512,
123
+ "pad_token": "<unk>",
124
+ "padding_side": "left",
125
+ "sp_model_kwargs": {},
126
+ "tokenizer_class": "LlamaTokenizer",
127
+ "unk_token": "<unk>",
128
+ "use_default_system_prompt": false
129
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 9.688250026819584e+16,
4
+ "train_loss": 1.1022938394574842,
5
+ "train_runtime": 7319.3616,
6
+ "train_samples_per_second": 1.15,
7
+ "train_steps_per_second": 1.15
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,2982 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 8415,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0023767082590612004,
13
+ "grad_norm": 3.921875,
14
+ "learning_rate": 2.3767082590612003e-06,
15
+ "loss": 3.6948,
16
+ "step": 20
17
+ },
18
+ {
19
+ "epoch": 0.004753416518122401,
20
+ "grad_norm": 5.28125,
21
+ "learning_rate": 4.753416518122401e-06,
22
+ "loss": 3.8478,
23
+ "step": 40
24
+ },
25
+ {
26
+ "epoch": 0.0071301247771836,
27
+ "grad_norm": 7.21875,
28
+ "learning_rate": 7.130124777183601e-06,
29
+ "loss": 3.6042,
30
+ "step": 60
31
+ },
32
+ {
33
+ "epoch": 0.009506833036244802,
34
+ "grad_norm": 2.625,
35
+ "learning_rate": 9.506833036244801e-06,
36
+ "loss": 3.5338,
37
+ "step": 80
38
+ },
39
+ {
40
+ "epoch": 0.011883541295306001,
41
+ "grad_norm": 1.515625,
42
+ "learning_rate": 1.1883541295306002e-05,
43
+ "loss": 3.2504,
44
+ "step": 100
45
+ },
46
+ {
47
+ "epoch": 0.0142602495543672,
48
+ "grad_norm": 1.3203125,
49
+ "learning_rate": 1.4260249554367203e-05,
50
+ "loss": 2.956,
51
+ "step": 120
52
+ },
53
+ {
54
+ "epoch": 0.016636957813428402,
55
+ "grad_norm": 1.6015625,
56
+ "learning_rate": 1.66369578134284e-05,
57
+ "loss": 2.9073,
58
+ "step": 140
59
+ },
60
+ {
61
+ "epoch": 0.019013666072489603,
62
+ "grad_norm": 1.0703125,
63
+ "learning_rate": 1.9013666072489602e-05,
64
+ "loss": 2.6436,
65
+ "step": 160
66
+ },
67
+ {
68
+ "epoch": 0.0213903743315508,
69
+ "grad_norm": 1.0234375,
70
+ "learning_rate": 2.1390374331550803e-05,
71
+ "loss": 2.6018,
72
+ "step": 180
73
+ },
74
+ {
75
+ "epoch": 0.023767082590612002,
76
+ "grad_norm": 1.0234375,
77
+ "learning_rate": 2.3767082590612004e-05,
78
+ "loss": 2.4289,
79
+ "step": 200
80
+ },
81
+ {
82
+ "epoch": 0.026143790849673203,
83
+ "grad_norm": 1.109375,
84
+ "learning_rate": 2.6143790849673204e-05,
85
+ "loss": 2.5239,
86
+ "step": 220
87
+ },
88
+ {
89
+ "epoch": 0.0285204991087344,
90
+ "grad_norm": 1.5625,
91
+ "learning_rate": 2.8520499108734405e-05,
92
+ "loss": 2.4617,
93
+ "step": 240
94
+ },
95
+ {
96
+ "epoch": 0.030897207367795602,
97
+ "grad_norm": 1.1171875,
98
+ "learning_rate": 3.08972073677956e-05,
99
+ "loss": 2.396,
100
+ "step": 260
101
+ },
102
+ {
103
+ "epoch": 0.033273915626856804,
104
+ "grad_norm": 1.03125,
105
+ "learning_rate": 3.32739156268568e-05,
106
+ "loss": 2.2682,
107
+ "step": 280
108
+ },
109
+ {
110
+ "epoch": 0.035650623885918005,
111
+ "grad_norm": 1.265625,
112
+ "learning_rate": 3.5650623885918004e-05,
113
+ "loss": 2.3627,
114
+ "step": 300
115
+ },
116
+ {
117
+ "epoch": 0.038027332144979206,
118
+ "grad_norm": 1.1484375,
119
+ "learning_rate": 3.8027332144979205e-05,
120
+ "loss": 2.3072,
121
+ "step": 320
122
+ },
123
+ {
124
+ "epoch": 0.04040404040404041,
125
+ "grad_norm": 1.5859375,
126
+ "learning_rate": 4.0404040404040405e-05,
127
+ "loss": 2.2124,
128
+ "step": 340
129
+ },
130
+ {
131
+ "epoch": 0.0427807486631016,
132
+ "grad_norm": 1.0703125,
133
+ "learning_rate": 4.2780748663101606e-05,
134
+ "loss": 2.2619,
135
+ "step": 360
136
+ },
137
+ {
138
+ "epoch": 0.0451574569221628,
139
+ "grad_norm": 1.46875,
140
+ "learning_rate": 4.515745692216281e-05,
141
+ "loss": 2.1928,
142
+ "step": 380
143
+ },
144
+ {
145
+ "epoch": 0.047534165181224004,
146
+ "grad_norm": 1.3515625,
147
+ "learning_rate": 4.753416518122401e-05,
148
+ "loss": 2.0428,
149
+ "step": 400
150
+ },
151
+ {
152
+ "epoch": 0.049910873440285206,
153
+ "grad_norm": 1.90625,
154
+ "learning_rate": 4.991087344028521e-05,
155
+ "loss": 1.9929,
156
+ "step": 420
157
+ },
158
+ {
159
+ "epoch": 0.05228758169934641,
160
+ "grad_norm": 1.3828125,
161
+ "learning_rate": 5.228758169934641e-05,
162
+ "loss": 2.146,
163
+ "step": 440
164
+ },
165
+ {
166
+ "epoch": 0.05466428995840761,
167
+ "grad_norm": 1.5859375,
168
+ "learning_rate": 5.466428995840761e-05,
169
+ "loss": 2.0289,
170
+ "step": 460
171
+ },
172
+ {
173
+ "epoch": 0.0570409982174688,
174
+ "grad_norm": 1.3125,
175
+ "learning_rate": 5.704099821746881e-05,
176
+ "loss": 1.9554,
177
+ "step": 480
178
+ },
179
+ {
180
+ "epoch": 0.059417706476530004,
181
+ "grad_norm": 1.6484375,
182
+ "learning_rate": 5.941770647653001e-05,
183
+ "loss": 1.9871,
184
+ "step": 500
185
+ },
186
+ {
187
+ "epoch": 0.061794414735591205,
188
+ "grad_norm": 1.9140625,
189
+ "learning_rate": 6.17944147355912e-05,
190
+ "loss": 2.0099,
191
+ "step": 520
192
+ },
193
+ {
194
+ "epoch": 0.06417112299465241,
195
+ "grad_norm": 2.140625,
196
+ "learning_rate": 6.41711229946524e-05,
197
+ "loss": 1.9199,
198
+ "step": 540
199
+ },
200
+ {
201
+ "epoch": 0.06654783125371361,
202
+ "grad_norm": 2.203125,
203
+ "learning_rate": 6.65478312537136e-05,
204
+ "loss": 2.0074,
205
+ "step": 560
206
+ },
207
+ {
208
+ "epoch": 0.0689245395127748,
209
+ "grad_norm": 1.2421875,
210
+ "learning_rate": 6.892453951277481e-05,
211
+ "loss": 1.8052,
212
+ "step": 580
213
+ },
214
+ {
215
+ "epoch": 0.07130124777183601,
216
+ "grad_norm": 2.25,
217
+ "learning_rate": 7.130124777183601e-05,
218
+ "loss": 2.0366,
219
+ "step": 600
220
+ },
221
+ {
222
+ "epoch": 0.0736779560308972,
223
+ "grad_norm": 2.4375,
224
+ "learning_rate": 7.367795603089721e-05,
225
+ "loss": 1.8426,
226
+ "step": 620
227
+ },
228
+ {
229
+ "epoch": 0.07605466428995841,
230
+ "grad_norm": 3.75,
231
+ "learning_rate": 7.605466428995841e-05,
232
+ "loss": 1.6836,
233
+ "step": 640
234
+ },
235
+ {
236
+ "epoch": 0.0784313725490196,
237
+ "grad_norm": 3.09375,
238
+ "learning_rate": 7.843137254901961e-05,
239
+ "loss": 2.0483,
240
+ "step": 660
241
+ },
242
+ {
243
+ "epoch": 0.08080808080808081,
244
+ "grad_norm": 1.6953125,
245
+ "learning_rate": 8.080808080808081e-05,
246
+ "loss": 1.902,
247
+ "step": 680
248
+ },
249
+ {
250
+ "epoch": 0.08318478906714201,
251
+ "grad_norm": 2.109375,
252
+ "learning_rate": 8.318478906714201e-05,
253
+ "loss": 1.9887,
254
+ "step": 700
255
+ },
256
+ {
257
+ "epoch": 0.0855614973262032,
258
+ "grad_norm": 2.984375,
259
+ "learning_rate": 8.556149732620321e-05,
260
+ "loss": 1.6542,
261
+ "step": 720
262
+ },
263
+ {
264
+ "epoch": 0.08793820558526441,
265
+ "grad_norm": 3.3125,
266
+ "learning_rate": 8.793820558526441e-05,
267
+ "loss": 1.6718,
268
+ "step": 740
269
+ },
270
+ {
271
+ "epoch": 0.0903149138443256,
272
+ "grad_norm": 2.9375,
273
+ "learning_rate": 9.031491384432561e-05,
274
+ "loss": 1.706,
275
+ "step": 760
276
+ },
277
+ {
278
+ "epoch": 0.09269162210338681,
279
+ "grad_norm": 3.59375,
280
+ "learning_rate": 9.269162210338681e-05,
281
+ "loss": 1.629,
282
+ "step": 780
283
+ },
284
+ {
285
+ "epoch": 0.09506833036244801,
286
+ "grad_norm": 2.40625,
287
+ "learning_rate": 9.506833036244802e-05,
288
+ "loss": 1.6748,
289
+ "step": 800
290
+ },
291
+ {
292
+ "epoch": 0.0974450386215092,
293
+ "grad_norm": 2.578125,
294
+ "learning_rate": 9.744503862150922e-05,
295
+ "loss": 1.8269,
296
+ "step": 820
297
+ },
298
+ {
299
+ "epoch": 0.09982174688057041,
300
+ "grad_norm": 2.328125,
301
+ "learning_rate": 9.982174688057042e-05,
302
+ "loss": 1.8727,
303
+ "step": 840
304
+ },
305
+ {
306
+ "epoch": 0.1021984551396316,
307
+ "grad_norm": 1.8671875,
308
+ "learning_rate": 0.00010219845513963162,
309
+ "loss": 1.7455,
310
+ "step": 860
311
+ },
312
+ {
313
+ "epoch": 0.10457516339869281,
314
+ "grad_norm": 2.140625,
315
+ "learning_rate": 0.00010457516339869282,
316
+ "loss": 1.7395,
317
+ "step": 880
318
+ },
319
+ {
320
+ "epoch": 0.10695187165775401,
321
+ "grad_norm": 1.609375,
322
+ "learning_rate": 0.00010695187165775402,
323
+ "loss": 1.7935,
324
+ "step": 900
325
+ },
326
+ {
327
+ "epoch": 0.10932857991681522,
328
+ "grad_norm": 1.6796875,
329
+ "learning_rate": 0.00010932857991681522,
330
+ "loss": 1.8009,
331
+ "step": 920
332
+ },
333
+ {
334
+ "epoch": 0.11170528817587641,
335
+ "grad_norm": 1.8515625,
336
+ "learning_rate": 0.00011170528817587642,
337
+ "loss": 1.5582,
338
+ "step": 940
339
+ },
340
+ {
341
+ "epoch": 0.1140819964349376,
342
+ "grad_norm": 2.0,
343
+ "learning_rate": 0.00011408199643493762,
344
+ "loss": 1.6402,
345
+ "step": 960
346
+ },
347
+ {
348
+ "epoch": 0.11645870469399881,
349
+ "grad_norm": 1.9921875,
350
+ "learning_rate": 0.00011645870469399882,
351
+ "loss": 1.665,
352
+ "step": 980
353
+ },
354
+ {
355
+ "epoch": 0.11883541295306001,
356
+ "grad_norm": 1.78125,
357
+ "learning_rate": 0.00011883541295306002,
358
+ "loss": 1.8407,
359
+ "step": 1000
360
+ },
361
+ {
362
+ "epoch": 0.12121212121212122,
363
+ "grad_norm": 1.9296875,
364
+ "learning_rate": 0.00012121212121212122,
365
+ "loss": 1.7665,
366
+ "step": 1020
367
+ },
368
+ {
369
+ "epoch": 0.12358882947118241,
370
+ "grad_norm": 2.5625,
371
+ "learning_rate": 0.0001235888294711824,
372
+ "loss": 1.8593,
373
+ "step": 1040
374
+ },
375
+ {
376
+ "epoch": 0.1259655377302436,
377
+ "grad_norm": 2.109375,
378
+ "learning_rate": 0.00012596553773024362,
379
+ "loss": 1.7738,
380
+ "step": 1060
381
+ },
382
+ {
383
+ "epoch": 0.12834224598930483,
384
+ "grad_norm": 2.078125,
385
+ "learning_rate": 0.0001283422459893048,
386
+ "loss": 1.8227,
387
+ "step": 1080
388
+ },
389
+ {
390
+ "epoch": 0.13071895424836602,
391
+ "grad_norm": 3.53125,
392
+ "learning_rate": 0.00013071895424836603,
393
+ "loss": 1.7159,
394
+ "step": 1100
395
+ },
396
+ {
397
+ "epoch": 0.13309566250742721,
398
+ "grad_norm": 1.890625,
399
+ "learning_rate": 0.0001330956625074272,
400
+ "loss": 1.6953,
401
+ "step": 1120
402
+ },
403
+ {
404
+ "epoch": 0.1354723707664884,
405
+ "grad_norm": 1.8671875,
406
+ "learning_rate": 0.00013547237076648843,
407
+ "loss": 1.5567,
408
+ "step": 1140
409
+ },
410
+ {
411
+ "epoch": 0.1378490790255496,
412
+ "grad_norm": 1.515625,
413
+ "learning_rate": 0.00013784907902554961,
414
+ "loss": 1.7637,
415
+ "step": 1160
416
+ },
417
+ {
418
+ "epoch": 0.14022578728461083,
419
+ "grad_norm": 2.375,
420
+ "learning_rate": 0.00014022578728461083,
421
+ "loss": 1.5884,
422
+ "step": 1180
423
+ },
424
+ {
425
+ "epoch": 0.14260249554367202,
426
+ "grad_norm": 2.046875,
427
+ "learning_rate": 0.00014260249554367202,
428
+ "loss": 1.5542,
429
+ "step": 1200
430
+ },
431
+ {
432
+ "epoch": 0.14497920380273321,
433
+ "grad_norm": 2.515625,
434
+ "learning_rate": 0.00014497920380273323,
435
+ "loss": 1.5124,
436
+ "step": 1220
437
+ },
438
+ {
439
+ "epoch": 0.1473559120617944,
440
+ "grad_norm": 1.609375,
441
+ "learning_rate": 0.00014735591206179442,
442
+ "loss": 1.5822,
443
+ "step": 1240
444
+ },
445
+ {
446
+ "epoch": 0.1497326203208556,
447
+ "grad_norm": 1.7890625,
448
+ "learning_rate": 0.00014973262032085563,
449
+ "loss": 1.6216,
450
+ "step": 1260
451
+ },
452
+ {
453
+ "epoch": 0.15210932857991682,
454
+ "grad_norm": 2.34375,
455
+ "learning_rate": 0.00015210932857991682,
456
+ "loss": 1.7366,
457
+ "step": 1280
458
+ },
459
+ {
460
+ "epoch": 0.15448603683897802,
461
+ "grad_norm": 1.8828125,
462
+ "learning_rate": 0.00015448603683897803,
463
+ "loss": 1.3673,
464
+ "step": 1300
465
+ },
466
+ {
467
+ "epoch": 0.1568627450980392,
468
+ "grad_norm": 1.8046875,
469
+ "learning_rate": 0.00015686274509803922,
470
+ "loss": 1.4071,
471
+ "step": 1320
472
+ },
473
+ {
474
+ "epoch": 0.1592394533571004,
475
+ "grad_norm": 2.59375,
476
+ "learning_rate": 0.00015923945335710043,
477
+ "loss": 1.6515,
478
+ "step": 1340
479
+ },
480
+ {
481
+ "epoch": 0.16161616161616163,
482
+ "grad_norm": 2.28125,
483
+ "learning_rate": 0.00016161616161616162,
484
+ "loss": 1.7604,
485
+ "step": 1360
486
+ },
487
+ {
488
+ "epoch": 0.16399286987522282,
489
+ "grad_norm": 1.9140625,
490
+ "learning_rate": 0.0001639928698752228,
491
+ "loss": 1.5925,
492
+ "step": 1380
493
+ },
494
+ {
495
+ "epoch": 0.16636957813428402,
496
+ "grad_norm": 3.40625,
497
+ "learning_rate": 0.00016636957813428402,
498
+ "loss": 1.5874,
499
+ "step": 1400
500
+ },
501
+ {
502
+ "epoch": 0.1687462863933452,
503
+ "grad_norm": 2.65625,
504
+ "learning_rate": 0.0001687462863933452,
505
+ "loss": 1.6116,
506
+ "step": 1420
507
+ },
508
+ {
509
+ "epoch": 0.1711229946524064,
510
+ "grad_norm": 1.8671875,
511
+ "learning_rate": 0.00017112299465240642,
512
+ "loss": 1.544,
513
+ "step": 1440
514
+ },
515
+ {
516
+ "epoch": 0.17349970291146763,
517
+ "grad_norm": 2.15625,
518
+ "learning_rate": 0.0001734997029114676,
519
+ "loss": 1.5015,
520
+ "step": 1460
521
+ },
522
+ {
523
+ "epoch": 0.17587641117052882,
524
+ "grad_norm": 1.8515625,
525
+ "learning_rate": 0.00017587641117052883,
526
+ "loss": 1.5858,
527
+ "step": 1480
528
+ },
529
+ {
530
+ "epoch": 0.17825311942959002,
531
+ "grad_norm": 1.90625,
532
+ "learning_rate": 0.00017825311942959,
533
+ "loss": 1.5469,
534
+ "step": 1500
535
+ },
536
+ {
537
+ "epoch": 0.1806298276886512,
538
+ "grad_norm": 2.421875,
539
+ "learning_rate": 0.00018062982768865123,
540
+ "loss": 1.5092,
541
+ "step": 1520
542
+ },
543
+ {
544
+ "epoch": 0.1830065359477124,
545
+ "grad_norm": 1.9921875,
546
+ "learning_rate": 0.00018300653594771241,
547
+ "loss": 1.3362,
548
+ "step": 1540
549
+ },
550
+ {
551
+ "epoch": 0.18538324420677363,
552
+ "grad_norm": 1.6015625,
553
+ "learning_rate": 0.00018538324420677363,
554
+ "loss": 1.5088,
555
+ "step": 1560
556
+ },
557
+ {
558
+ "epoch": 0.18775995246583482,
559
+ "grad_norm": 1.9765625,
560
+ "learning_rate": 0.00018775995246583482,
561
+ "loss": 1.5518,
562
+ "step": 1580
563
+ },
564
+ {
565
+ "epoch": 0.19013666072489602,
566
+ "grad_norm": 1.8046875,
567
+ "learning_rate": 0.00019013666072489603,
568
+ "loss": 1.4344,
569
+ "step": 1600
570
+ },
571
+ {
572
+ "epoch": 0.1925133689839572,
573
+ "grad_norm": 1.71875,
574
+ "learning_rate": 0.00019251336898395722,
575
+ "loss": 1.5411,
576
+ "step": 1620
577
+ },
578
+ {
579
+ "epoch": 0.1948900772430184,
580
+ "grad_norm": 1.515625,
581
+ "learning_rate": 0.00019489007724301843,
582
+ "loss": 1.5832,
583
+ "step": 1640
584
+ },
585
+ {
586
+ "epoch": 0.19726678550207963,
587
+ "grad_norm": 1.71875,
588
+ "learning_rate": 0.00019726678550207962,
589
+ "loss": 1.5468,
590
+ "step": 1660
591
+ },
592
+ {
593
+ "epoch": 0.19964349376114082,
594
+ "grad_norm": 2.078125,
595
+ "learning_rate": 0.00019964349376114083,
596
+ "loss": 1.3053,
597
+ "step": 1680
598
+ },
599
+ {
600
+ "epoch": 0.20202020202020202,
601
+ "grad_norm": 2.0625,
602
+ "learning_rate": 0.0001999968531423333,
603
+ "loss": 1.5303,
604
+ "step": 1700
605
+ },
606
+ {
607
+ "epoch": 0.2043969102792632,
608
+ "grad_norm": 1.4921875,
609
+ "learning_rate": 0.0001999850935511802,
610
+ "loss": 1.6331,
611
+ "step": 1720
612
+ },
613
+ {
614
+ "epoch": 0.20677361853832443,
615
+ "grad_norm": 1.8203125,
616
+ "learning_rate": 0.0001999646242513938,
617
+ "loss": 1.5621,
618
+ "step": 1740
619
+ },
620
+ {
621
+ "epoch": 0.20915032679738563,
622
+ "grad_norm": 1.578125,
623
+ "learning_rate": 0.00019993544702605638,
624
+ "loss": 1.4932,
625
+ "step": 1760
626
+ },
627
+ {
628
+ "epoch": 0.21152703505644682,
629
+ "grad_norm": 2.09375,
630
+ "learning_rate": 0.00019989756441679795,
631
+ "loss": 1.3936,
632
+ "step": 1780
633
+ },
634
+ {
635
+ "epoch": 0.21390374331550802,
636
+ "grad_norm": 1.4453125,
637
+ "learning_rate": 0.00019985097972357547,
638
+ "loss": 1.3227,
639
+ "step": 1800
640
+ },
641
+ {
642
+ "epoch": 0.2162804515745692,
643
+ "grad_norm": 1.671875,
644
+ "learning_rate": 0.0001997956970043848,
645
+ "loss": 1.4253,
646
+ "step": 1820
647
+ },
648
+ {
649
+ "epoch": 0.21865715983363043,
650
+ "grad_norm": 1.5078125,
651
+ "learning_rate": 0.00019973172107490758,
652
+ "loss": 1.3946,
653
+ "step": 1840
654
+ },
655
+ {
656
+ "epoch": 0.22103386809269163,
657
+ "grad_norm": 1.625,
658
+ "learning_rate": 0.00019965905750809158,
659
+ "loss": 1.2779,
660
+ "step": 1860
661
+ },
662
+ {
663
+ "epoch": 0.22341057635175282,
664
+ "grad_norm": 1.4296875,
665
+ "learning_rate": 0.00019957771263366526,
666
+ "loss": 1.5051,
667
+ "step": 1880
668
+ },
669
+ {
670
+ "epoch": 0.22578728461081402,
671
+ "grad_norm": 1.9765625,
672
+ "learning_rate": 0.00019948769353758653,
673
+ "loss": 1.3954,
674
+ "step": 1900
675
+ },
676
+ {
677
+ "epoch": 0.2281639928698752,
678
+ "grad_norm": 2.25,
679
+ "learning_rate": 0.00019938900806142518,
680
+ "loss": 1.2049,
681
+ "step": 1920
682
+ },
683
+ {
684
+ "epoch": 0.23054070112893643,
685
+ "grad_norm": 2.5625,
686
+ "learning_rate": 0.00019928166480168011,
687
+ "loss": 1.5335,
688
+ "step": 1940
689
+ },
690
+ {
691
+ "epoch": 0.23291740938799763,
692
+ "grad_norm": 2.03125,
693
+ "learning_rate": 0.00019916567310903035,
694
+ "loss": 1.3868,
695
+ "step": 1960
696
+ },
697
+ {
698
+ "epoch": 0.23529411764705882,
699
+ "grad_norm": 1.46875,
700
+ "learning_rate": 0.0001990410430875205,
701
+ "loss": 1.3677,
702
+ "step": 1980
703
+ },
704
+ {
705
+ "epoch": 0.23767082590612001,
706
+ "grad_norm": 2.484375,
707
+ "learning_rate": 0.0001989077855936807,
708
+ "loss": 1.2757,
709
+ "step": 2000
710
+ },
711
+ {
712
+ "epoch": 0.24004753416518124,
713
+ "grad_norm": 2.21875,
714
+ "learning_rate": 0.0001987659122355806,
715
+ "loss": 1.4893,
716
+ "step": 2020
717
+ },
718
+ {
719
+ "epoch": 0.24242424242424243,
720
+ "grad_norm": 1.28125,
721
+ "learning_rate": 0.00019861543537181867,
722
+ "loss": 1.2123,
723
+ "step": 2040
724
+ },
725
+ {
726
+ "epoch": 0.24480095068330363,
727
+ "grad_norm": 2.078125,
728
+ "learning_rate": 0.00019845636811044521,
729
+ "loss": 1.4714,
730
+ "step": 2060
731
+ },
732
+ {
733
+ "epoch": 0.24717765894236482,
734
+ "grad_norm": 2.453125,
735
+ "learning_rate": 0.0001982887243078207,
736
+ "loss": 1.463,
737
+ "step": 2080
738
+ },
739
+ {
740
+ "epoch": 0.24955436720142601,
741
+ "grad_norm": 1.953125,
742
+ "learning_rate": 0.00019811251856740873,
743
+ "loss": 1.5086,
744
+ "step": 2100
745
+ },
746
+ {
747
+ "epoch": 0.2519310754604872,
748
+ "grad_norm": 1.734375,
749
+ "learning_rate": 0.000197927766238504,
750
+ "loss": 1.3604,
751
+ "step": 2120
752
+ },
753
+ {
754
+ "epoch": 0.25430778371954843,
755
+ "grad_norm": 1.96875,
756
+ "learning_rate": 0.00019773448341489495,
757
+ "loss": 1.2226,
758
+ "step": 2140
759
+ },
760
+ {
761
+ "epoch": 0.25668449197860965,
762
+ "grad_norm": 1.609375,
763
+ "learning_rate": 0.00019753268693346225,
764
+ "loss": 1.2003,
765
+ "step": 2160
766
+ },
767
+ {
768
+ "epoch": 0.2590612002376708,
769
+ "grad_norm": 1.9296875,
770
+ "learning_rate": 0.0001973223943727117,
771
+ "loss": 1.192,
772
+ "step": 2180
773
+ },
774
+ {
775
+ "epoch": 0.26143790849673204,
776
+ "grad_norm": 1.9765625,
777
+ "learning_rate": 0.00019710362405124334,
778
+ "loss": 1.2616,
779
+ "step": 2200
780
+ },
781
+ {
782
+ "epoch": 0.2638146167557932,
783
+ "grad_norm": 1.2265625,
784
+ "learning_rate": 0.0001968763950261554,
785
+ "loss": 1.2493,
786
+ "step": 2220
787
+ },
788
+ {
789
+ "epoch": 0.26619132501485443,
790
+ "grad_norm": 2.234375,
791
+ "learning_rate": 0.0001966407270913846,
792
+ "loss": 1.278,
793
+ "step": 2240
794
+ },
795
+ {
796
+ "epoch": 0.26856803327391565,
797
+ "grad_norm": 1.5703125,
798
+ "learning_rate": 0.00019639664077598142,
799
+ "loss": 1.3797,
800
+ "step": 2260
801
+ },
802
+ {
803
+ "epoch": 0.2709447415329768,
804
+ "grad_norm": 2.078125,
805
+ "learning_rate": 0.0001961441573423223,
806
+ "loss": 1.1243,
807
+ "step": 2280
808
+ },
809
+ {
810
+ "epoch": 0.27332144979203804,
811
+ "grad_norm": 2.0,
812
+ "learning_rate": 0.0001958832987842571,
813
+ "loss": 1.324,
814
+ "step": 2300
815
+ },
816
+ {
817
+ "epoch": 0.2756981580510992,
818
+ "grad_norm": 2.375,
819
+ "learning_rate": 0.00019561408782519345,
820
+ "loss": 1.3572,
821
+ "step": 2320
822
+ },
823
+ {
824
+ "epoch": 0.27807486631016043,
825
+ "grad_norm": 2.265625,
826
+ "learning_rate": 0.0001953365479161172,
827
+ "loss": 1.3404,
828
+ "step": 2340
829
+ },
830
+ {
831
+ "epoch": 0.28045157456922165,
832
+ "grad_norm": 2.984375,
833
+ "learning_rate": 0.00019505070323354965,
834
+ "loss": 1.3237,
835
+ "step": 2360
836
+ },
837
+ {
838
+ "epoch": 0.2828282828282828,
839
+ "grad_norm": 1.8203125,
840
+ "learning_rate": 0.0001947565786774415,
841
+ "loss": 1.2894,
842
+ "step": 2380
843
+ },
844
+ {
845
+ "epoch": 0.28520499108734404,
846
+ "grad_norm": 1.7578125,
847
+ "learning_rate": 0.0001944541998690038,
848
+ "loss": 1.3445,
849
+ "step": 2400
850
+ },
851
+ {
852
+ "epoch": 0.2875816993464052,
853
+ "grad_norm": 2.015625,
854
+ "learning_rate": 0.0001941435931484761,
855
+ "loss": 1.3859,
856
+ "step": 2420
857
+ },
858
+ {
859
+ "epoch": 0.28995840760546643,
860
+ "grad_norm": 1.7890625,
861
+ "learning_rate": 0.00019382478557283204,
862
+ "loss": 1.2185,
863
+ "step": 2440
864
+ },
865
+ {
866
+ "epoch": 0.29233511586452765,
867
+ "grad_norm": 1.546875,
868
+ "learning_rate": 0.00019349780491342223,
869
+ "loss": 1.3545,
870
+ "step": 2460
871
+ },
872
+ {
873
+ "epoch": 0.2947118241235888,
874
+ "grad_norm": 2.125,
875
+ "learning_rate": 0.00019316267965355528,
876
+ "loss": 1.2694,
877
+ "step": 2480
878
+ },
879
+ {
880
+ "epoch": 0.29708853238265004,
881
+ "grad_norm": 2.03125,
882
+ "learning_rate": 0.00019281943898601645,
883
+ "loss": 1.0126,
884
+ "step": 2500
885
+ },
886
+ {
887
+ "epoch": 0.2994652406417112,
888
+ "grad_norm": 1.6328125,
889
+ "learning_rate": 0.00019246811281052487,
890
+ "loss": 1.2593,
891
+ "step": 2520
892
+ },
893
+ {
894
+ "epoch": 0.3018419489007724,
895
+ "grad_norm": 1.578125,
896
+ "learning_rate": 0.00019210873173112865,
897
+ "loss": 1.3609,
898
+ "step": 2540
899
+ },
900
+ {
901
+ "epoch": 0.30421865715983365,
902
+ "grad_norm": 1.375,
903
+ "learning_rate": 0.0001917413270535393,
904
+ "loss": 0.9368,
905
+ "step": 2560
906
+ },
907
+ {
908
+ "epoch": 0.3065953654188948,
909
+ "grad_norm": 2.25,
910
+ "learning_rate": 0.0001913659307824045,
911
+ "loss": 1.1893,
912
+ "step": 2580
913
+ },
914
+ {
915
+ "epoch": 0.30897207367795604,
916
+ "grad_norm": 2.21875,
917
+ "learning_rate": 0.0001909825756185202,
918
+ "loss": 1.1345,
919
+ "step": 2600
920
+ },
921
+ {
922
+ "epoch": 0.3113487819370172,
923
+ "grad_norm": 1.515625,
924
+ "learning_rate": 0.0001905912949559821,
925
+ "loss": 1.2925,
926
+ "step": 2620
927
+ },
928
+ {
929
+ "epoch": 0.3137254901960784,
930
+ "grad_norm": 1.296875,
931
+ "learning_rate": 0.00019019212287927663,
932
+ "loss": 1.0842,
933
+ "step": 2640
934
+ },
935
+ {
936
+ "epoch": 0.31610219845513965,
937
+ "grad_norm": 2.40625,
938
+ "learning_rate": 0.00018978509416031186,
939
+ "loss": 1.2722,
940
+ "step": 2660
941
+ },
942
+ {
943
+ "epoch": 0.3184789067142008,
944
+ "grad_norm": 1.796875,
945
+ "learning_rate": 0.00018937024425538855,
946
+ "loss": 1.3566,
947
+ "step": 2680
948
+ },
949
+ {
950
+ "epoch": 0.32085561497326204,
951
+ "grad_norm": 2.09375,
952
+ "learning_rate": 0.0001889476093021115,
953
+ "loss": 0.9899,
954
+ "step": 2700
955
+ },
956
+ {
957
+ "epoch": 0.32323232323232326,
958
+ "grad_norm": 1.8046875,
959
+ "learning_rate": 0.00018851722611624164,
960
+ "loss": 1.0713,
961
+ "step": 2720
962
+ },
963
+ {
964
+ "epoch": 0.3256090314913844,
965
+ "grad_norm": 2.25,
966
+ "learning_rate": 0.00018807913218848906,
967
+ "loss": 1.1957,
968
+ "step": 2740
969
+ },
970
+ {
971
+ "epoch": 0.32798573975044565,
972
+ "grad_norm": 1.453125,
973
+ "learning_rate": 0.000187633365681247,
974
+ "loss": 1.3258,
975
+ "step": 2760
976
+ },
977
+ {
978
+ "epoch": 0.3303624480095068,
979
+ "grad_norm": 1.5625,
980
+ "learning_rate": 0.00018717996542526777,
981
+ "loss": 1.1146,
982
+ "step": 2780
983
+ },
984
+ {
985
+ "epoch": 0.33273915626856804,
986
+ "grad_norm": 1.984375,
987
+ "learning_rate": 0.00018671897091627993,
988
+ "loss": 1.0215,
989
+ "step": 2800
990
+ },
991
+ {
992
+ "epoch": 0.33511586452762926,
993
+ "grad_norm": 1.5703125,
994
+ "learning_rate": 0.00018625042231154817,
995
+ "loss": 1.1107,
996
+ "step": 2820
997
+ },
998
+ {
999
+ "epoch": 0.3374925727866904,
1000
+ "grad_norm": 1.8828125,
1001
+ "learning_rate": 0.00018577436042637477,
1002
+ "loss": 1.1655,
1003
+ "step": 2840
1004
+ },
1005
+ {
1006
+ "epoch": 0.33986928104575165,
1007
+ "grad_norm": 1.765625,
1008
+ "learning_rate": 0.00018529082673054457,
1009
+ "loss": 1.1197,
1010
+ "step": 2860
1011
+ },
1012
+ {
1013
+ "epoch": 0.3422459893048128,
1014
+ "grad_norm": 1.8828125,
1015
+ "learning_rate": 0.0001847998633447123,
1016
+ "loss": 1.1792,
1017
+ "step": 2880
1018
+ },
1019
+ {
1020
+ "epoch": 0.34462269756387404,
1021
+ "grad_norm": 1.9765625,
1022
+ "learning_rate": 0.0001843015130367335,
1023
+ "loss": 1.1933,
1024
+ "step": 2900
1025
+ },
1026
+ {
1027
+ "epoch": 0.34699940582293526,
1028
+ "grad_norm": 2.15625,
1029
+ "learning_rate": 0.00018379581921793914,
1030
+ "loss": 1.1603,
1031
+ "step": 2920
1032
+ },
1033
+ {
1034
+ "epoch": 0.3493761140819964,
1035
+ "grad_norm": 1.09375,
1036
+ "learning_rate": 0.00018328282593935377,
1037
+ "loss": 1.1653,
1038
+ "step": 2940
1039
+ },
1040
+ {
1041
+ "epoch": 0.35175282234105765,
1042
+ "grad_norm": 1.7578125,
1043
+ "learning_rate": 0.00018276257788785855,
1044
+ "loss": 1.0807,
1045
+ "step": 2960
1046
+ },
1047
+ {
1048
+ "epoch": 0.3541295306001188,
1049
+ "grad_norm": 2.296875,
1050
+ "learning_rate": 0.00018223512038229833,
1051
+ "loss": 1.0923,
1052
+ "step": 2980
1053
+ },
1054
+ {
1055
+ "epoch": 0.35650623885918004,
1056
+ "grad_norm": 1.7109375,
1057
+ "learning_rate": 0.00018170049936953406,
1058
+ "loss": 1.3022,
1059
+ "step": 3000
1060
+ },
1061
+ {
1062
+ "epoch": 0.35888294711824126,
1063
+ "grad_norm": 1.625,
1064
+ "learning_rate": 0.00018115876142044032,
1065
+ "loss": 1.2415,
1066
+ "step": 3020
1067
+ },
1068
+ {
1069
+ "epoch": 0.3612596553773024,
1070
+ "grad_norm": 1.4453125,
1071
+ "learning_rate": 0.0001806099537258485,
1072
+ "loss": 1.0849,
1073
+ "step": 3040
1074
+ },
1075
+ {
1076
+ "epoch": 0.36363636363636365,
1077
+ "grad_norm": 1.8203125,
1078
+ "learning_rate": 0.00018005412409243606,
1079
+ "loss": 1.1712,
1080
+ "step": 3060
1081
+ },
1082
+ {
1083
+ "epoch": 0.3660130718954248,
1084
+ "grad_norm": 1.4921875,
1085
+ "learning_rate": 0.000179491320938562,
1086
+ "loss": 1.2795,
1087
+ "step": 3080
1088
+ },
1089
+ {
1090
+ "epoch": 0.36838978015448604,
1091
+ "grad_norm": 1.734375,
1092
+ "learning_rate": 0.00017892159329004916,
1093
+ "loss": 1.1304,
1094
+ "step": 3100
1095
+ },
1096
+ {
1097
+ "epoch": 0.37076648841354726,
1098
+ "grad_norm": 2.265625,
1099
+ "learning_rate": 0.00017834499077591374,
1100
+ "loss": 1.277,
1101
+ "step": 3120
1102
+ },
1103
+ {
1104
+ "epoch": 0.3731431966726084,
1105
+ "grad_norm": 1.5546875,
1106
+ "learning_rate": 0.00017776156362404186,
1107
+ "loss": 1.2222,
1108
+ "step": 3140
1109
+ },
1110
+ {
1111
+ "epoch": 0.37551990493166965,
1112
+ "grad_norm": 2.4375,
1113
+ "learning_rate": 0.0001771713626568143,
1114
+ "loss": 1.2639,
1115
+ "step": 3160
1116
+ },
1117
+ {
1118
+ "epoch": 0.3778966131907308,
1119
+ "grad_norm": 1.7109375,
1120
+ "learning_rate": 0.0001765744392866795,
1121
+ "loss": 1.1679,
1122
+ "step": 3180
1123
+ },
1124
+ {
1125
+ "epoch": 0.38027332144979203,
1126
+ "grad_norm": 1.9921875,
1127
+ "learning_rate": 0.00017597084551167476,
1128
+ "loss": 1.0657,
1129
+ "step": 3200
1130
+ },
1131
+ {
1132
+ "epoch": 0.38265002970885326,
1133
+ "grad_norm": 1.2109375,
1134
+ "learning_rate": 0.00017536063391089697,
1135
+ "loss": 1.1886,
1136
+ "step": 3220
1137
+ },
1138
+ {
1139
+ "epoch": 0.3850267379679144,
1140
+ "grad_norm": 1.1875,
1141
+ "learning_rate": 0.00017474385763992212,
1142
+ "loss": 1.1526,
1143
+ "step": 3240
1144
+ },
1145
+ {
1146
+ "epoch": 0.38740344622697565,
1147
+ "grad_norm": 1.6171875,
1148
+ "learning_rate": 0.00017412057042617525,
1149
+ "loss": 1.0648,
1150
+ "step": 3260
1151
+ },
1152
+ {
1153
+ "epoch": 0.3897801544860368,
1154
+ "grad_norm": 1.8515625,
1155
+ "learning_rate": 0.00017349082656424995,
1156
+ "loss": 1.2938,
1157
+ "step": 3280
1158
+ },
1159
+ {
1160
+ "epoch": 0.39215686274509803,
1161
+ "grad_norm": 1.34375,
1162
+ "learning_rate": 0.00017285468091117904,
1163
+ "loss": 1.1246,
1164
+ "step": 3300
1165
+ },
1166
+ {
1167
+ "epoch": 0.39453357100415926,
1168
+ "grad_norm": 2.03125,
1169
+ "learning_rate": 0.00017221218888165572,
1170
+ "loss": 1.0179,
1171
+ "step": 3320
1172
+ },
1173
+ {
1174
+ "epoch": 0.3969102792632204,
1175
+ "grad_norm": 1.8671875,
1176
+ "learning_rate": 0.0001715634064432065,
1177
+ "loss": 0.914,
1178
+ "step": 3340
1179
+ },
1180
+ {
1181
+ "epoch": 0.39928698752228164,
1182
+ "grad_norm": 1.4296875,
1183
+ "learning_rate": 0.0001709083901113159,
1184
+ "loss": 1.1192,
1185
+ "step": 3360
1186
+ },
1187
+ {
1188
+ "epoch": 0.40166369578134287,
1189
+ "grad_norm": 3.203125,
1190
+ "learning_rate": 0.00017024719694450337,
1191
+ "loss": 1.1462,
1192
+ "step": 3380
1193
+ },
1194
+ {
1195
+ "epoch": 0.40404040404040403,
1196
+ "grad_norm": 1.3515625,
1197
+ "learning_rate": 0.00016957988453935276,
1198
+ "loss": 0.9452,
1199
+ "step": 3400
1200
+ },
1201
+ {
1202
+ "epoch": 0.40641711229946526,
1203
+ "grad_norm": 2.28125,
1204
+ "learning_rate": 0.00016890651102549538,
1205
+ "loss": 1.1837,
1206
+ "step": 3420
1207
+ },
1208
+ {
1209
+ "epoch": 0.4087938205585264,
1210
+ "grad_norm": 2.515625,
1211
+ "learning_rate": 0.00016822713506054604,
1212
+ "loss": 0.8889,
1213
+ "step": 3440
1214
+ },
1215
+ {
1216
+ "epoch": 0.41117052881758764,
1217
+ "grad_norm": 1.515625,
1218
+ "learning_rate": 0.0001675418158249935,
1219
+ "loss": 1.0873,
1220
+ "step": 3460
1221
+ },
1222
+ {
1223
+ "epoch": 0.41354723707664887,
1224
+ "grad_norm": 1.3515625,
1225
+ "learning_rate": 0.0001668506130170453,
1226
+ "loss": 0.995,
1227
+ "step": 3480
1228
+ },
1229
+ {
1230
+ "epoch": 0.41592394533571003,
1231
+ "grad_norm": 1.7421875,
1232
+ "learning_rate": 0.0001661535868474273,
1233
+ "loss": 1.1595,
1234
+ "step": 3500
1235
+ },
1236
+ {
1237
+ "epoch": 0.41830065359477125,
1238
+ "grad_norm": 2.15625,
1239
+ "learning_rate": 0.00016545079803413892,
1240
+ "loss": 1.1134,
1241
+ "step": 3520
1242
+ },
1243
+ {
1244
+ "epoch": 0.4206773618538324,
1245
+ "grad_norm": 0.92578125,
1246
+ "learning_rate": 0.00016474230779716384,
1247
+ "loss": 0.7802,
1248
+ "step": 3540
1249
+ },
1250
+ {
1251
+ "epoch": 0.42305407011289364,
1252
+ "grad_norm": 1.671875,
1253
+ "learning_rate": 0.00016402817785313712,
1254
+ "loss": 0.9085,
1255
+ "step": 3560
1256
+ },
1257
+ {
1258
+ "epoch": 0.42543077837195487,
1259
+ "grad_norm": 2.390625,
1260
+ "learning_rate": 0.00016330847040996915,
1261
+ "loss": 1.0042,
1262
+ "step": 3580
1263
+ },
1264
+ {
1265
+ "epoch": 0.42780748663101603,
1266
+ "grad_norm": 2.46875,
1267
+ "learning_rate": 0.00016258324816142668,
1268
+ "loss": 1.0318,
1269
+ "step": 3600
1270
+ },
1271
+ {
1272
+ "epoch": 0.43018419489007725,
1273
+ "grad_norm": 3.375,
1274
+ "learning_rate": 0.00016185257428167143,
1275
+ "loss": 1.0833,
1276
+ "step": 3620
1277
+ },
1278
+ {
1279
+ "epoch": 0.4325609031491384,
1280
+ "grad_norm": 2.078125,
1281
+ "learning_rate": 0.00016111651241975734,
1282
+ "loss": 1.0208,
1283
+ "step": 3640
1284
+ },
1285
+ {
1286
+ "epoch": 0.43493761140819964,
1287
+ "grad_norm": 2.125,
1288
+ "learning_rate": 0.00016037512669408565,
1289
+ "loss": 1.0074,
1290
+ "step": 3660
1291
+ },
1292
+ {
1293
+ "epoch": 0.43731431966726086,
1294
+ "grad_norm": 1.5703125,
1295
+ "learning_rate": 0.0001596284816868198,
1296
+ "loss": 1.2403,
1297
+ "step": 3680
1298
+ },
1299
+ {
1300
+ "epoch": 0.43969102792632203,
1301
+ "grad_norm": 2.59375,
1302
+ "learning_rate": 0.00015887664243825967,
1303
+ "loss": 1.0694,
1304
+ "step": 3700
1305
+ },
1306
+ {
1307
+ "epoch": 0.44206773618538325,
1308
+ "grad_norm": 1.3203125,
1309
+ "learning_rate": 0.0001581196744411759,
1310
+ "loss": 1.0033,
1311
+ "step": 3720
1312
+ },
1313
+ {
1314
+ "epoch": 0.4444444444444444,
1315
+ "grad_norm": 2.359375,
1316
+ "learning_rate": 0.0001573576436351046,
1317
+ "loss": 1.0574,
1318
+ "step": 3740
1319
+ },
1320
+ {
1321
+ "epoch": 0.44682115270350564,
1322
+ "grad_norm": 2.109375,
1323
+ "learning_rate": 0.00015659061640060378,
1324
+ "loss": 1.2204,
1325
+ "step": 3760
1326
+ },
1327
+ {
1328
+ "epoch": 0.44919786096256686,
1329
+ "grad_norm": 1.8046875,
1330
+ "learning_rate": 0.0001558186595534705,
1331
+ "loss": 1.157,
1332
+ "step": 3780
1333
+ },
1334
+ {
1335
+ "epoch": 0.45157456922162803,
1336
+ "grad_norm": 1.6171875,
1337
+ "learning_rate": 0.0001550418403389208,
1338
+ "loss": 0.9542,
1339
+ "step": 3800
1340
+ },
1341
+ {
1342
+ "epoch": 0.45395127748068925,
1343
+ "grad_norm": 2.390625,
1344
+ "learning_rate": 0.00015426022642573193,
1345
+ "loss": 0.9281,
1346
+ "step": 3820
1347
+ },
1348
+ {
1349
+ "epoch": 0.4563279857397504,
1350
+ "grad_norm": 1.4453125,
1351
+ "learning_rate": 0.00015347388590034757,
1352
+ "loss": 0.9331,
1353
+ "step": 3840
1354
+ },
1355
+ {
1356
+ "epoch": 0.45870469399881164,
1357
+ "grad_norm": 1.3828125,
1358
+ "learning_rate": 0.00015268288726094705,
1359
+ "loss": 0.7978,
1360
+ "step": 3860
1361
+ },
1362
+ {
1363
+ "epoch": 0.46108140225787286,
1364
+ "grad_norm": 2.421875,
1365
+ "learning_rate": 0.00015188729941147824,
1366
+ "loss": 0.9807,
1367
+ "step": 3880
1368
+ },
1369
+ {
1370
+ "epoch": 0.46345811051693403,
1371
+ "grad_norm": 2.640625,
1372
+ "learning_rate": 0.0001510871916556555,
1373
+ "loss": 1.0974,
1374
+ "step": 3900
1375
+ },
1376
+ {
1377
+ "epoch": 0.46583481877599525,
1378
+ "grad_norm": 1.171875,
1379
+ "learning_rate": 0.00015028263369092253,
1380
+ "loss": 0.9807,
1381
+ "step": 3920
1382
+ },
1383
+ {
1384
+ "epoch": 0.4682115270350565,
1385
+ "grad_norm": 1.6484375,
1386
+ "learning_rate": 0.00014947369560238104,
1387
+ "loss": 1.0588,
1388
+ "step": 3940
1389
+ },
1390
+ {
1391
+ "epoch": 0.47058823529411764,
1392
+ "grad_norm": 2.546875,
1393
+ "learning_rate": 0.00014866044785668563,
1394
+ "loss": 1.0442,
1395
+ "step": 3960
1396
+ },
1397
+ {
1398
+ "epoch": 0.47296494355317886,
1399
+ "grad_norm": 2.9375,
1400
+ "learning_rate": 0.00014784296129590548,
1401
+ "loss": 0.8837,
1402
+ "step": 3980
1403
+ },
1404
+ {
1405
+ "epoch": 0.47534165181224003,
1406
+ "grad_norm": 1.7890625,
1407
+ "learning_rate": 0.00014702130713135317,
1408
+ "loss": 1.129,
1409
+ "step": 4000
1410
+ },
1411
+ {
1412
+ "epoch": 0.47771836007130125,
1413
+ "grad_norm": 2.171875,
1414
+ "learning_rate": 0.00014619555693738166,
1415
+ "loss": 1.0496,
1416
+ "step": 4020
1417
+ },
1418
+ {
1419
+ "epoch": 0.4800950683303625,
1420
+ "grad_norm": 2.328125,
1421
+ "learning_rate": 0.0001453657826451493,
1422
+ "loss": 1.0136,
1423
+ "step": 4040
1424
+ },
1425
+ {
1426
+ "epoch": 0.48247177658942364,
1427
+ "grad_norm": 2.203125,
1428
+ "learning_rate": 0.00014453205653635376,
1429
+ "loss": 0.9794,
1430
+ "step": 4060
1431
+ },
1432
+ {
1433
+ "epoch": 0.48484848484848486,
1434
+ "grad_norm": 2.734375,
1435
+ "learning_rate": 0.00014369445123693596,
1436
+ "loss": 0.7651,
1437
+ "step": 4080
1438
+ },
1439
+ {
1440
+ "epoch": 0.48722519310754603,
1441
+ "grad_norm": 2.15625,
1442
+ "learning_rate": 0.0001428530397107533,
1443
+ "loss": 0.9667,
1444
+ "step": 4100
1445
+ },
1446
+ {
1447
+ "epoch": 0.48960190136660725,
1448
+ "grad_norm": 2.578125,
1449
+ "learning_rate": 0.0001420078952532238,
1450
+ "loss": 0.8897,
1451
+ "step": 4120
1452
+ },
1453
+ {
1454
+ "epoch": 0.4919786096256685,
1455
+ "grad_norm": 4.28125,
1456
+ "learning_rate": 0.0001411590914849415,
1457
+ "loss": 1.0208,
1458
+ "step": 4140
1459
+ },
1460
+ {
1461
+ "epoch": 0.49435531788472964,
1462
+ "grad_norm": 2.34375,
1463
+ "learning_rate": 0.00014030670234526323,
1464
+ "loss": 1.0138,
1465
+ "step": 4160
1466
+ },
1467
+ {
1468
+ "epoch": 0.49673202614379086,
1469
+ "grad_norm": 1.234375,
1470
+ "learning_rate": 0.00013945080208586775,
1471
+ "loss": 0.8435,
1472
+ "step": 4180
1473
+ },
1474
+ {
1475
+ "epoch": 0.49910873440285203,
1476
+ "grad_norm": 2.828125,
1477
+ "learning_rate": 0.0001385914652642877,
1478
+ "loss": 0.7405,
1479
+ "step": 4200
1480
+ },
1481
+ {
1482
+ "epoch": 0.5014854426619133,
1483
+ "grad_norm": 2.59375,
1484
+ "learning_rate": 0.00013772876673741498,
1485
+ "loss": 0.8999,
1486
+ "step": 4220
1487
+ },
1488
+ {
1489
+ "epoch": 0.5038621509209744,
1490
+ "grad_norm": 2.171875,
1491
+ "learning_rate": 0.00013686278165497977,
1492
+ "loss": 0.9063,
1493
+ "step": 4240
1494
+ },
1495
+ {
1496
+ "epoch": 0.5062388591800356,
1497
+ "grad_norm": 3.578125,
1498
+ "learning_rate": 0.00013599358545300438,
1499
+ "loss": 0.8485,
1500
+ "step": 4260
1501
+ },
1502
+ {
1503
+ "epoch": 0.5086155674390969,
1504
+ "grad_norm": 2.5625,
1505
+ "learning_rate": 0.00013512125384723204,
1506
+ "loss": 0.7843,
1507
+ "step": 4280
1508
+ },
1509
+ {
1510
+ "epoch": 0.5109922756981581,
1511
+ "grad_norm": 1.5546875,
1512
+ "learning_rate": 0.00013424586282653116,
1513
+ "loss": 0.7548,
1514
+ "step": 4300
1515
+ },
1516
+ {
1517
+ "epoch": 0.5133689839572193,
1518
+ "grad_norm": 2.8125,
1519
+ "learning_rate": 0.00013336748864627593,
1520
+ "loss": 0.9609,
1521
+ "step": 4320
1522
+ },
1523
+ {
1524
+ "epoch": 0.5157456922162804,
1525
+ "grad_norm": 2.453125,
1526
+ "learning_rate": 0.00013248620782170396,
1527
+ "loss": 0.9358,
1528
+ "step": 4340
1529
+ },
1530
+ {
1531
+ "epoch": 0.5181224004753416,
1532
+ "grad_norm": 2.21875,
1533
+ "learning_rate": 0.00013160209712125074,
1534
+ "loss": 1.0639,
1535
+ "step": 4360
1536
+ },
1537
+ {
1538
+ "epoch": 0.5204991087344029,
1539
+ "grad_norm": 1.515625,
1540
+ "learning_rate": 0.0001307152335598624,
1541
+ "loss": 1.0147,
1542
+ "step": 4380
1543
+ },
1544
+ {
1545
+ "epoch": 0.5228758169934641,
1546
+ "grad_norm": 2.4375,
1547
+ "learning_rate": 0.00012982569439228713,
1548
+ "loss": 0.7749,
1549
+ "step": 4400
1550
+ },
1551
+ {
1552
+ "epoch": 0.5252525252525253,
1553
+ "grad_norm": 2.21875,
1554
+ "learning_rate": 0.0001289335571063453,
1555
+ "loss": 0.642,
1556
+ "step": 4420
1557
+ },
1558
+ {
1559
+ "epoch": 0.5276292335115864,
1560
+ "grad_norm": 2.984375,
1561
+ "learning_rate": 0.00012803889941617944,
1562
+ "loss": 0.863,
1563
+ "step": 4440
1564
+ },
1565
+ {
1566
+ "epoch": 0.5300059417706476,
1567
+ "grad_norm": 2.265625,
1568
+ "learning_rate": 0.0001271417992554849,
1569
+ "loss": 0.8817,
1570
+ "step": 4460
1571
+ },
1572
+ {
1573
+ "epoch": 0.5323826500297089,
1574
+ "grad_norm": 2.3125,
1575
+ "learning_rate": 0.00012624233477072057,
1576
+ "loss": 0.9788,
1577
+ "step": 4480
1578
+ },
1579
+ {
1580
+ "epoch": 0.5347593582887701,
1581
+ "grad_norm": 2.484375,
1582
+ "learning_rate": 0.00012534058431430198,
1583
+ "loss": 0.8273,
1584
+ "step": 4500
1585
+ },
1586
+ {
1587
+ "epoch": 0.5371360665478313,
1588
+ "grad_norm": 1.2265625,
1589
+ "learning_rate": 0.0001244366264377757,
1590
+ "loss": 0.848,
1591
+ "step": 4520
1592
+ },
1593
+ {
1594
+ "epoch": 0.5395127748068924,
1595
+ "grad_norm": 2.625,
1596
+ "learning_rate": 0.00012353053988497684,
1597
+ "loss": 0.8556,
1598
+ "step": 4540
1599
+ },
1600
+ {
1601
+ "epoch": 0.5418894830659536,
1602
+ "grad_norm": 2.375,
1603
+ "learning_rate": 0.00012262240358516967,
1604
+ "loss": 0.8137,
1605
+ "step": 4560
1606
+ },
1607
+ {
1608
+ "epoch": 0.5442661913250149,
1609
+ "grad_norm": 1.1015625,
1610
+ "learning_rate": 0.00012171229664617208,
1611
+ "loss": 0.7362,
1612
+ "step": 4580
1613
+ },
1614
+ {
1615
+ "epoch": 0.5466428995840761,
1616
+ "grad_norm": 2.90625,
1617
+ "learning_rate": 0.00012080029834746447,
1618
+ "loss": 0.7504,
1619
+ "step": 4600
1620
+ },
1621
+ {
1622
+ "epoch": 0.5490196078431373,
1623
+ "grad_norm": 3.109375,
1624
+ "learning_rate": 0.00011988648813328367,
1625
+ "loss": 1.0486,
1626
+ "step": 4620
1627
+ },
1628
+ {
1629
+ "epoch": 0.5513963161021984,
1630
+ "grad_norm": 1.875,
1631
+ "learning_rate": 0.00011897094560570265,
1632
+ "loss": 0.7487,
1633
+ "step": 4640
1634
+ },
1635
+ {
1636
+ "epoch": 0.5537730243612596,
1637
+ "grad_norm": 2.171875,
1638
+ "learning_rate": 0.00011805375051769636,
1639
+ "loss": 0.9325,
1640
+ "step": 4660
1641
+ },
1642
+ {
1643
+ "epoch": 0.5561497326203209,
1644
+ "grad_norm": 2.28125,
1645
+ "learning_rate": 0.00011713498276619432,
1646
+ "loss": 0.906,
1647
+ "step": 4680
1648
+ },
1649
+ {
1650
+ "epoch": 0.5585264408793821,
1651
+ "grad_norm": 1.953125,
1652
+ "learning_rate": 0.0001162147223851209,
1653
+ "loss": 0.7214,
1654
+ "step": 4700
1655
+ },
1656
+ {
1657
+ "epoch": 0.5609031491384433,
1658
+ "grad_norm": 2.34375,
1659
+ "learning_rate": 0.0001152930495384236,
1660
+ "loss": 1.01,
1661
+ "step": 4720
1662
+ },
1663
+ {
1664
+ "epoch": 0.5632798573975044,
1665
+ "grad_norm": 2.53125,
1666
+ "learning_rate": 0.00011437004451308983,
1667
+ "loss": 1.0897,
1668
+ "step": 4740
1669
+ },
1670
+ {
1671
+ "epoch": 0.5656565656565656,
1672
+ "grad_norm": 3.296875,
1673
+ "learning_rate": 0.00011344578771215319,
1674
+ "loss": 0.8222,
1675
+ "step": 4760
1676
+ },
1677
+ {
1678
+ "epoch": 0.5680332739156269,
1679
+ "grad_norm": 2.921875,
1680
+ "learning_rate": 0.00011252035964768961,
1681
+ "loss": 0.8411,
1682
+ "step": 4780
1683
+ },
1684
+ {
1685
+ "epoch": 0.5704099821746881,
1686
+ "grad_norm": 1.9375,
1687
+ "learning_rate": 0.00011159384093380377,
1688
+ "loss": 0.747,
1689
+ "step": 4800
1690
+ },
1691
+ {
1692
+ "epoch": 0.5727866904337493,
1693
+ "grad_norm": 2.375,
1694
+ "learning_rate": 0.00011066631227960693,
1695
+ "loss": 0.7352,
1696
+ "step": 4820
1697
+ },
1698
+ {
1699
+ "epoch": 0.5751633986928104,
1700
+ "grad_norm": 2.484375,
1701
+ "learning_rate": 0.00010973785448218639,
1702
+ "loss": 0.901,
1703
+ "step": 4840
1704
+ },
1705
+ {
1706
+ "epoch": 0.5775401069518716,
1707
+ "grad_norm": 1.7265625,
1708
+ "learning_rate": 0.00010880854841956712,
1709
+ "loss": 0.6489,
1710
+ "step": 4860
1711
+ },
1712
+ {
1713
+ "epoch": 0.5799168152109329,
1714
+ "grad_norm": 1.3359375,
1715
+ "learning_rate": 0.00010787847504366649,
1716
+ "loss": 0.7034,
1717
+ "step": 4880
1718
+ },
1719
+ {
1720
+ "epoch": 0.5822935234699941,
1721
+ "grad_norm": 3.046875,
1722
+ "learning_rate": 0.00010694771537324269,
1723
+ "loss": 0.7427,
1724
+ "step": 4900
1725
+ },
1726
+ {
1727
+ "epoch": 0.5846702317290553,
1728
+ "grad_norm": 3.40625,
1729
+ "learning_rate": 0.00010601635048683698,
1730
+ "loss": 0.7655,
1731
+ "step": 4920
1732
+ },
1733
+ {
1734
+ "epoch": 0.5870469399881164,
1735
+ "grad_norm": 2.21875,
1736
+ "learning_rate": 0.00010508446151571109,
1737
+ "loss": 0.7467,
1738
+ "step": 4940
1739
+ },
1740
+ {
1741
+ "epoch": 0.5894236482471776,
1742
+ "grad_norm": 1.9375,
1743
+ "learning_rate": 0.0001041521296367798,
1744
+ "loss": 0.7942,
1745
+ "step": 4960
1746
+ },
1747
+ {
1748
+ "epoch": 0.5918003565062389,
1749
+ "grad_norm": 3.34375,
1750
+ "learning_rate": 0.00010321943606553961,
1751
+ "loss": 1.033,
1752
+ "step": 4980
1753
+ },
1754
+ {
1755
+ "epoch": 0.5941770647653001,
1756
+ "grad_norm": 2.765625,
1757
+ "learning_rate": 0.00010228646204899401,
1758
+ "loss": 0.7049,
1759
+ "step": 5000
1760
+ },
1761
+ {
1762
+ "epoch": 0.5965537730243613,
1763
+ "grad_norm": 1.59375,
1764
+ "learning_rate": 0.0001013532888585762,
1765
+ "loss": 0.7016,
1766
+ "step": 5020
1767
+ },
1768
+ {
1769
+ "epoch": 0.5989304812834224,
1770
+ "grad_norm": 2.875,
1771
+ "learning_rate": 0.00010041999778306936,
1772
+ "loss": 0.9511,
1773
+ "step": 5040
1774
+ },
1775
+ {
1776
+ "epoch": 0.6013071895424836,
1777
+ "grad_norm": 3.640625,
1778
+ "learning_rate": 9.948667012152566e-05,
1779
+ "loss": 0.6895,
1780
+ "step": 5060
1781
+ },
1782
+ {
1783
+ "epoch": 0.6036838978015449,
1784
+ "grad_norm": 3.3125,
1785
+ "learning_rate": 9.855338717618432e-05,
1786
+ "loss": 0.8696,
1787
+ "step": 5080
1788
+ },
1789
+ {
1790
+ "epoch": 0.6060606060606061,
1791
+ "grad_norm": 2.625,
1792
+ "learning_rate": 9.762023024538926e-05,
1793
+ "loss": 0.8621,
1794
+ "step": 5100
1795
+ },
1796
+ {
1797
+ "epoch": 0.6084373143196673,
1798
+ "grad_norm": 1.984375,
1799
+ "learning_rate": 9.668728061650733e-05,
1800
+ "loss": 0.9548,
1801
+ "step": 5120
1802
+ },
1803
+ {
1804
+ "epoch": 0.6108140225787284,
1805
+ "grad_norm": 3.40625,
1806
+ "learning_rate": 9.575461955884726e-05,
1807
+ "loss": 0.8048,
1808
+ "step": 5140
1809
+ },
1810
+ {
1811
+ "epoch": 0.6131907308377896,
1812
+ "grad_norm": 2.265625,
1813
+ "learning_rate": 9.482232831658034e-05,
1814
+ "loss": 0.8866,
1815
+ "step": 5160
1816
+ },
1817
+ {
1818
+ "epoch": 0.6155674390968509,
1819
+ "grad_norm": 1.5390625,
1820
+ "learning_rate": 9.389048810166317e-05,
1821
+ "loss": 0.7404,
1822
+ "step": 5180
1823
+ },
1824
+ {
1825
+ "epoch": 0.6179441473559121,
1826
+ "grad_norm": 1.359375,
1827
+ "learning_rate": 9.29591800867634e-05,
1828
+ "loss": 0.6817,
1829
+ "step": 5200
1830
+ },
1831
+ {
1832
+ "epoch": 0.6203208556149733,
1833
+ "grad_norm": 5.0,
1834
+ "learning_rate": 9.202848539818865e-05,
1835
+ "loss": 0.7504,
1836
+ "step": 5220
1837
+ },
1838
+ {
1839
+ "epoch": 0.6226975638740344,
1840
+ "grad_norm": 2.484375,
1841
+ "learning_rate": 9.109848510881961e-05,
1842
+ "loss": 0.7767,
1843
+ "step": 5240
1844
+ },
1845
+ {
1846
+ "epoch": 0.6250742721330956,
1847
+ "grad_norm": 2.53125,
1848
+ "learning_rate": 9.016926023104789e-05,
1849
+ "loss": 0.7846,
1850
+ "step": 5260
1851
+ },
1852
+ {
1853
+ "epoch": 0.6274509803921569,
1854
+ "grad_norm": 3.421875,
1855
+ "learning_rate": 8.924089170971887e-05,
1856
+ "loss": 0.8327,
1857
+ "step": 5280
1858
+ },
1859
+ {
1860
+ "epoch": 0.6298276886512181,
1861
+ "grad_norm": 3.765625,
1862
+ "learning_rate": 8.831346041508069e-05,
1863
+ "loss": 0.8035,
1864
+ "step": 5300
1865
+ },
1866
+ {
1867
+ "epoch": 0.6322043969102793,
1868
+ "grad_norm": 2.25,
1869
+ "learning_rate": 8.738704713573959e-05,
1870
+ "loss": 0.9569,
1871
+ "step": 5320
1872
+ },
1873
+ {
1874
+ "epoch": 0.6345811051693404,
1875
+ "grad_norm": 2.34375,
1876
+ "learning_rate": 8.646173257162245e-05,
1877
+ "loss": 0.6629,
1878
+ "step": 5340
1879
+ },
1880
+ {
1881
+ "epoch": 0.6369578134284016,
1882
+ "grad_norm": 2.875,
1883
+ "learning_rate": 8.553759732694696e-05,
1884
+ "loss": 0.6597,
1885
+ "step": 5360
1886
+ },
1887
+ {
1888
+ "epoch": 0.6393345216874629,
1889
+ "grad_norm": 2.859375,
1890
+ "learning_rate": 8.461472190320021e-05,
1891
+ "loss": 0.6857,
1892
+ "step": 5380
1893
+ },
1894
+ {
1895
+ "epoch": 0.6417112299465241,
1896
+ "grad_norm": 1.5234375,
1897
+ "learning_rate": 8.369318669212625e-05,
1898
+ "loss": 0.7234,
1899
+ "step": 5400
1900
+ },
1901
+ {
1902
+ "epoch": 0.6440879382055853,
1903
+ "grad_norm": 3.359375,
1904
+ "learning_rate": 8.277307196872303e-05,
1905
+ "loss": 0.7483,
1906
+ "step": 5420
1907
+ },
1908
+ {
1909
+ "epoch": 0.6464646464646465,
1910
+ "grad_norm": 1.21875,
1911
+ "learning_rate": 8.185445788424974e-05,
1912
+ "loss": 0.7538,
1913
+ "step": 5440
1914
+ },
1915
+ {
1916
+ "epoch": 0.6488413547237076,
1917
+ "grad_norm": 3.234375,
1918
+ "learning_rate": 8.093742445924491e-05,
1919
+ "loss": 0.7882,
1920
+ "step": 5460
1921
+ },
1922
+ {
1923
+ "epoch": 0.6512180629827689,
1924
+ "grad_norm": 3.546875,
1925
+ "learning_rate": 8.002205157655554e-05,
1926
+ "loss": 0.6984,
1927
+ "step": 5480
1928
+ },
1929
+ {
1930
+ "epoch": 0.6535947712418301,
1931
+ "grad_norm": 2.28125,
1932
+ "learning_rate": 7.910841897437875e-05,
1933
+ "loss": 0.7044,
1934
+ "step": 5500
1935
+ },
1936
+ {
1937
+ "epoch": 0.6559714795008913,
1938
+ "grad_norm": 3.390625,
1939
+ "learning_rate": 7.819660623931575e-05,
1940
+ "loss": 0.6569,
1941
+ "step": 5520
1942
+ },
1943
+ {
1944
+ "epoch": 0.6583481877599525,
1945
+ "grad_norm": 1.5859375,
1946
+ "learning_rate": 7.728669279943897e-05,
1947
+ "loss": 0.8069,
1948
+ "step": 5540
1949
+ },
1950
+ {
1951
+ "epoch": 0.6607248960190136,
1952
+ "grad_norm": 3.125,
1953
+ "learning_rate": 7.637875791737299e-05,
1954
+ "loss": 0.759,
1955
+ "step": 5560
1956
+ },
1957
+ {
1958
+ "epoch": 0.6631016042780749,
1959
+ "grad_norm": 1.4921875,
1960
+ "learning_rate": 7.547288068339025e-05,
1961
+ "loss": 0.6806,
1962
+ "step": 5580
1963
+ },
1964
+ {
1965
+ "epoch": 0.6654783125371361,
1966
+ "grad_norm": 2.875,
1967
+ "learning_rate": 7.456914000852123e-05,
1968
+ "loss": 0.7777,
1969
+ "step": 5600
1970
+ },
1971
+ {
1972
+ "epoch": 0.6678550207961973,
1973
+ "grad_norm": 1.453125,
1974
+ "learning_rate": 7.366761461768052e-05,
1975
+ "loss": 0.7375,
1976
+ "step": 5620
1977
+ },
1978
+ {
1979
+ "epoch": 0.6702317290552585,
1980
+ "grad_norm": 1.734375,
1981
+ "learning_rate": 7.276838304280935e-05,
1982
+ "loss": 0.6862,
1983
+ "step": 5640
1984
+ },
1985
+ {
1986
+ "epoch": 0.6726084373143196,
1987
+ "grad_norm": 3.171875,
1988
+ "learning_rate": 7.187152361603432e-05,
1989
+ "loss": 0.6195,
1990
+ "step": 5660
1991
+ },
1992
+ {
1993
+ "epoch": 0.6749851455733809,
1994
+ "grad_norm": 1.8046875,
1995
+ "learning_rate": 7.097711446284405e-05,
1996
+ "loss": 0.5486,
1997
+ "step": 5680
1998
+ },
1999
+ {
2000
+ "epoch": 0.6773618538324421,
2001
+ "grad_norm": 2.203125,
2002
+ "learning_rate": 7.008523349528377e-05,
2003
+ "loss": 0.7452,
2004
+ "step": 5700
2005
+ },
2006
+ {
2007
+ "epoch": 0.6797385620915033,
2008
+ "grad_norm": 3.03125,
2009
+ "learning_rate": 6.919595840516815e-05,
2010
+ "loss": 0.7417,
2011
+ "step": 5720
2012
+ },
2013
+ {
2014
+ "epoch": 0.6821152703505645,
2015
+ "grad_norm": 3.1875,
2016
+ "learning_rate": 6.830936665731371e-05,
2017
+ "loss": 0.6798,
2018
+ "step": 5740
2019
+ },
2020
+ {
2021
+ "epoch": 0.6844919786096256,
2022
+ "grad_norm": 2.328125,
2023
+ "learning_rate": 6.742553548279095e-05,
2024
+ "loss": 0.7222,
2025
+ "step": 5760
2026
+ },
2027
+ {
2028
+ "epoch": 0.6868686868686869,
2029
+ "grad_norm": 1.0625,
2030
+ "learning_rate": 6.654454187219649e-05,
2031
+ "loss": 0.6629,
2032
+ "step": 5780
2033
+ },
2034
+ {
2035
+ "epoch": 0.6892453951277481,
2036
+ "grad_norm": 1.6015625,
2037
+ "learning_rate": 6.56664625689466e-05,
2038
+ "loss": 0.7808,
2039
+ "step": 5800
2040
+ },
2041
+ {
2042
+ "epoch": 0.6916221033868093,
2043
+ "grad_norm": 2.421875,
2044
+ "learning_rate": 6.479137406259206e-05,
2045
+ "loss": 0.7898,
2046
+ "step": 5820
2047
+ },
2048
+ {
2049
+ "epoch": 0.6939988116458705,
2050
+ "grad_norm": 1.53125,
2051
+ "learning_rate": 6.39193525821551e-05,
2052
+ "loss": 0.8595,
2053
+ "step": 5840
2054
+ },
2055
+ {
2056
+ "epoch": 0.6963755199049316,
2057
+ "grad_norm": 3.21875,
2058
+ "learning_rate": 6.305047408948908e-05,
2059
+ "loss": 0.8169,
2060
+ "step": 5860
2061
+ },
2062
+ {
2063
+ "epoch": 0.6987522281639929,
2064
+ "grad_norm": 2.796875,
2065
+ "learning_rate": 6.21848142726615e-05,
2066
+ "loss": 0.7892,
2067
+ "step": 5880
2068
+ },
2069
+ {
2070
+ "epoch": 0.7011289364230541,
2071
+ "grad_norm": 2.671875,
2072
+ "learning_rate": 6.13224485393608e-05,
2073
+ "loss": 0.8179,
2074
+ "step": 5900
2075
+ },
2076
+ {
2077
+ "epoch": 0.7035056446821153,
2078
+ "grad_norm": 1.8046875,
2079
+ "learning_rate": 6.046345201032748e-05,
2080
+ "loss": 0.6881,
2081
+ "step": 5920
2082
+ },
2083
+ {
2084
+ "epoch": 0.7058823529411765,
2085
+ "grad_norm": 1.515625,
2086
+ "learning_rate": 5.960789951281052e-05,
2087
+ "loss": 0.9263,
2088
+ "step": 5940
2089
+ },
2090
+ {
2091
+ "epoch": 0.7082590612002376,
2092
+ "grad_norm": 2.21875,
2093
+ "learning_rate": 5.8755865574049016e-05,
2094
+ "loss": 0.7661,
2095
+ "step": 5960
2096
+ },
2097
+ {
2098
+ "epoch": 0.7106357694592988,
2099
+ "grad_norm": 2.5625,
2100
+ "learning_rate": 5.7907424414780135e-05,
2101
+ "loss": 0.7932,
2102
+ "step": 5980
2103
+ },
2104
+ {
2105
+ "epoch": 0.7130124777183601,
2106
+ "grad_norm": 1.8984375,
2107
+ "learning_rate": 5.706264994277386e-05,
2108
+ "loss": 0.8045,
2109
+ "step": 6000
2110
+ },
2111
+ {
2112
+ "epoch": 0.7153891859774213,
2113
+ "grad_norm": 1.09375,
2114
+ "learning_rate": 5.6221615746394644e-05,
2115
+ "loss": 0.6996,
2116
+ "step": 6020
2117
+ },
2118
+ {
2119
+ "epoch": 0.7177658942364825,
2120
+ "grad_norm": 1.796875,
2121
+ "learning_rate": 5.538439508819139e-05,
2122
+ "loss": 0.7662,
2123
+ "step": 6040
2124
+ },
2125
+ {
2126
+ "epoch": 0.7201426024955436,
2127
+ "grad_norm": 1.890625,
2128
+ "learning_rate": 5.4551060898515404e-05,
2129
+ "loss": 0.6049,
2130
+ "step": 6060
2131
+ },
2132
+ {
2133
+ "epoch": 0.7225193107546048,
2134
+ "grad_norm": 2.15625,
2135
+ "learning_rate": 5.372168576916732e-05,
2136
+ "loss": 0.6502,
2137
+ "step": 6080
2138
+ },
2139
+ {
2140
+ "epoch": 0.7248960190136661,
2141
+ "grad_norm": 2.6875,
2142
+ "learning_rate": 5.289634194707387e-05,
2143
+ "loss": 0.6785,
2144
+ "step": 6100
2145
+ },
2146
+ {
2147
+ "epoch": 0.7272727272727273,
2148
+ "grad_norm": 2.890625,
2149
+ "learning_rate": 5.207510132799436e-05,
2150
+ "loss": 0.7902,
2151
+ "step": 6120
2152
+ },
2153
+ {
2154
+ "epoch": 0.7296494355317885,
2155
+ "grad_norm": 2.703125,
2156
+ "learning_rate": 5.125803545025758e-05,
2157
+ "loss": 0.6948,
2158
+ "step": 6140
2159
+ },
2160
+ {
2161
+ "epoch": 0.7320261437908496,
2162
+ "grad_norm": 2.84375,
2163
+ "learning_rate": 5.0445215488530525e-05,
2164
+ "loss": 0.7113,
2165
+ "step": 6160
2166
+ },
2167
+ {
2168
+ "epoch": 0.7344028520499108,
2169
+ "grad_norm": 3.484375,
2170
+ "learning_rate": 4.963671224761808e-05,
2171
+ "loss": 0.6249,
2172
+ "step": 6180
2173
+ },
2174
+ {
2175
+ "epoch": 0.7367795603089721,
2176
+ "grad_norm": 1.59375,
2177
+ "learning_rate": 4.883259615629515e-05,
2178
+ "loss": 0.7295,
2179
+ "step": 6200
2180
+ },
2181
+ {
2182
+ "epoch": 0.7391562685680333,
2183
+ "grad_norm": 1.359375,
2184
+ "learning_rate": 4.8032937261171896e-05,
2185
+ "loss": 0.8041,
2186
+ "step": 6220
2187
+ },
2188
+ {
2189
+ "epoch": 0.7415329768270945,
2190
+ "grad_norm": 1.8046875,
2191
+ "learning_rate": 4.7237805220591744e-05,
2192
+ "loss": 0.6296,
2193
+ "step": 6240
2194
+ },
2195
+ {
2196
+ "epoch": 0.7439096850861556,
2197
+ "grad_norm": 1.609375,
2198
+ "learning_rate": 4.644726929856342e-05,
2199
+ "loss": 0.8887,
2200
+ "step": 6260
2201
+ },
2202
+ {
2203
+ "epoch": 0.7462863933452168,
2204
+ "grad_norm": 2.421875,
2205
+ "learning_rate": 4.5661398358727524e-05,
2206
+ "loss": 0.6501,
2207
+ "step": 6280
2208
+ },
2209
+ {
2210
+ "epoch": 0.7486631016042781,
2211
+ "grad_norm": 1.7109375,
2212
+ "learning_rate": 4.4880260858357746e-05,
2213
+ "loss": 0.5981,
2214
+ "step": 6300
2215
+ },
2216
+ {
2217
+ "epoch": 0.7510398098633393,
2218
+ "grad_norm": 1.765625,
2219
+ "learning_rate": 4.4103924842397395e-05,
2220
+ "loss": 0.5993,
2221
+ "step": 6320
2222
+ },
2223
+ {
2224
+ "epoch": 0.7534165181224005,
2225
+ "grad_norm": 3.484375,
2226
+ "learning_rate": 4.3332457937532246e-05,
2227
+ "loss": 0.8187,
2228
+ "step": 6340
2229
+ },
2230
+ {
2231
+ "epoch": 0.7557932263814616,
2232
+ "grad_norm": 7.09375,
2233
+ "learning_rate": 4.256592734629947e-05,
2234
+ "loss": 0.826,
2235
+ "step": 6360
2236
+ },
2237
+ {
2238
+ "epoch": 0.7581699346405228,
2239
+ "grad_norm": 1.578125,
2240
+ "learning_rate": 4.18043998412335e-05,
2241
+ "loss": 0.5091,
2242
+ "step": 6380
2243
+ },
2244
+ {
2245
+ "epoch": 0.7605466428995841,
2246
+ "grad_norm": 0.78125,
2247
+ "learning_rate": 4.104794175904966e-05,
2248
+ "loss": 0.609,
2249
+ "step": 6400
2250
+ },
2251
+ {
2252
+ "epoch": 0.7629233511586453,
2253
+ "grad_norm": 3.171875,
2254
+ "learning_rate": 4.02966189948655e-05,
2255
+ "loss": 0.7031,
2256
+ "step": 6420
2257
+ },
2258
+ {
2259
+ "epoch": 0.7653000594177065,
2260
+ "grad_norm": 1.6015625,
2261
+ "learning_rate": 3.955049699646054e-05,
2262
+ "loss": 0.5454,
2263
+ "step": 6440
2264
+ },
2265
+ {
2266
+ "epoch": 0.7676767676767676,
2267
+ "grad_norm": 3.578125,
2268
+ "learning_rate": 3.880964075857535e-05,
2269
+ "loss": 0.629,
2270
+ "step": 6460
2271
+ },
2272
+ {
2273
+ "epoch": 0.7700534759358288,
2274
+ "grad_norm": 2.609375,
2275
+ "learning_rate": 3.80741148172497e-05,
2276
+ "loss": 0.6376,
2277
+ "step": 6480
2278
+ },
2279
+ {
2280
+ "epoch": 0.7724301841948901,
2281
+ "grad_norm": 2.015625,
2282
+ "learning_rate": 3.734398324420073e-05,
2283
+ "loss": 0.6744,
2284
+ "step": 6500
2285
+ },
2286
+ {
2287
+ "epoch": 0.7748068924539513,
2288
+ "grad_norm": 3.140625,
2289
+ "learning_rate": 3.661930964124193e-05,
2290
+ "loss": 0.6133,
2291
+ "step": 6520
2292
+ },
2293
+ {
2294
+ "epoch": 0.7771836007130125,
2295
+ "grad_norm": 1.734375,
2296
+ "learning_rate": 3.5900157134742574e-05,
2297
+ "loss": 0.5723,
2298
+ "step": 6540
2299
+ },
2300
+ {
2301
+ "epoch": 0.7795603089720736,
2302
+ "grad_norm": 3.921875,
2303
+ "learning_rate": 3.5186588370128746e-05,
2304
+ "loss": 0.6248,
2305
+ "step": 6560
2306
+ },
2307
+ {
2308
+ "epoch": 0.7819370172311348,
2309
+ "grad_norm": 3.203125,
2310
+ "learning_rate": 3.447866550642649e-05,
2311
+ "loss": 0.6323,
2312
+ "step": 6580
2313
+ },
2314
+ {
2315
+ "epoch": 0.7843137254901961,
2316
+ "grad_norm": 3.1875,
2317
+ "learning_rate": 3.377645021084701e-05,
2318
+ "loss": 0.7705,
2319
+ "step": 6600
2320
+ },
2321
+ {
2322
+ "epoch": 0.7866904337492573,
2323
+ "grad_norm": 2.546875,
2324
+ "learning_rate": 3.3080003653414724e-05,
2325
+ "loss": 0.5787,
2326
+ "step": 6620
2327
+ },
2328
+ {
2329
+ "epoch": 0.7890671420083185,
2330
+ "grad_norm": 1.2109375,
2331
+ "learning_rate": 3.238938650163899e-05,
2332
+ "loss": 0.7117,
2333
+ "step": 6640
2334
+ },
2335
+ {
2336
+ "epoch": 0.7914438502673797,
2337
+ "grad_norm": 1.640625,
2338
+ "learning_rate": 3.17046589152292e-05,
2339
+ "loss": 0.5979,
2340
+ "step": 6660
2341
+ },
2342
+ {
2343
+ "epoch": 0.7938205585264408,
2344
+ "grad_norm": 4.53125,
2345
+ "learning_rate": 3.10258805408542e-05,
2346
+ "loss": 0.6202,
2347
+ "step": 6680
2348
+ },
2349
+ {
2350
+ "epoch": 0.7961972667855021,
2351
+ "grad_norm": 3.484375,
2352
+ "learning_rate": 3.0353110506946647e-05,
2353
+ "loss": 0.6462,
2354
+ "step": 6700
2355
+ },
2356
+ {
2357
+ "epoch": 0.7985739750445633,
2358
+ "grad_norm": 2.640625,
2359
+ "learning_rate": 2.968640741855223e-05,
2360
+ "loss": 0.576,
2361
+ "step": 6720
2362
+ },
2363
+ {
2364
+ "epoch": 0.8009506833036245,
2365
+ "grad_norm": 1.875,
2366
+ "learning_rate": 2.9025829352224477e-05,
2367
+ "loss": 0.7645,
2368
+ "step": 6740
2369
+ },
2370
+ {
2371
+ "epoch": 0.8033273915626857,
2372
+ "grad_norm": 2.453125,
2373
+ "learning_rate": 2.8371433850965922e-05,
2374
+ "loss": 0.6348,
2375
+ "step": 6760
2376
+ },
2377
+ {
2378
+ "epoch": 0.8057040998217468,
2379
+ "grad_norm": 2.875,
2380
+ "learning_rate": 2.7723277919215397e-05,
2381
+ "loss": 0.7165,
2382
+ "step": 6780
2383
+ },
2384
+ {
2385
+ "epoch": 0.8080808080808081,
2386
+ "grad_norm": 1.7890625,
2387
+ "learning_rate": 2.70814180178823e-05,
2388
+ "loss": 0.6772,
2389
+ "step": 6800
2390
+ },
2391
+ {
2392
+ "epoch": 0.8104575163398693,
2393
+ "grad_norm": 3.25,
2394
+ "learning_rate": 2.644591005942846e-05,
2395
+ "loss": 0.6995,
2396
+ "step": 6820
2397
+ },
2398
+ {
2399
+ "epoch": 0.8128342245989305,
2400
+ "grad_norm": 3.546875,
2401
+ "learning_rate": 2.5816809402997522e-05,
2402
+ "loss": 0.6519,
2403
+ "step": 6840
2404
+ },
2405
+ {
2406
+ "epoch": 0.8152109328579917,
2407
+ "grad_norm": 4.375,
2408
+ "learning_rate": 2.5194170849592492e-05,
2409
+ "loss": 0.612,
2410
+ "step": 6860
2411
+ },
2412
+ {
2413
+ "epoch": 0.8175876411170528,
2414
+ "grad_norm": 3.40625,
2415
+ "learning_rate": 2.4578048637302208e-05,
2416
+ "loss": 0.8347,
2417
+ "step": 6880
2418
+ },
2419
+ {
2420
+ "epoch": 0.8199643493761141,
2421
+ "grad_norm": 3.953125,
2422
+ "learning_rate": 2.396849643657657e-05,
2423
+ "loss": 0.7846,
2424
+ "step": 6900
2425
+ },
2426
+ {
2427
+ "epoch": 0.8223410576351753,
2428
+ "grad_norm": 3.0,
2429
+ "learning_rate": 2.3365567345551233e-05,
2430
+ "loss": 0.6799,
2431
+ "step": 6920
2432
+ },
2433
+ {
2434
+ "epoch": 0.8247177658942365,
2435
+ "grad_norm": 2.15625,
2436
+ "learning_rate": 2.276931388542235e-05,
2437
+ "loss": 0.704,
2438
+ "step": 6940
2439
+ },
2440
+ {
2441
+ "epoch": 0.8270944741532977,
2442
+ "grad_norm": 3.3125,
2443
+ "learning_rate": 2.2179787995871403e-05,
2444
+ "loss": 0.5832,
2445
+ "step": 6960
2446
+ },
2447
+ {
2448
+ "epoch": 0.8294711824123588,
2449
+ "grad_norm": 1.8046875,
2450
+ "learning_rate": 2.1597041030540643e-05,
2451
+ "loss": 0.6464,
2452
+ "step": 6980
2453
+ },
2454
+ {
2455
+ "epoch": 0.8318478906714201,
2456
+ "grad_norm": 1.0859375,
2457
+ "learning_rate": 2.1021123752559836e-05,
2458
+ "loss": 0.6596,
2459
+ "step": 7000
2460
+ },
2461
+ {
2462
+ "epoch": 0.8342245989304813,
2463
+ "grad_norm": 2.4375,
2464
+ "learning_rate": 2.0452086330124164e-05,
2465
+ "loss": 0.6124,
2466
+ "step": 7020
2467
+ },
2468
+ {
2469
+ "epoch": 0.8366013071895425,
2470
+ "grad_norm": 1.953125,
2471
+ "learning_rate": 1.988997833212406e-05,
2472
+ "loss": 0.8721,
2473
+ "step": 7040
2474
+ },
2475
+ {
2476
+ "epoch": 0.8389780154486037,
2477
+ "grad_norm": 1.0234375,
2478
+ "learning_rate": 1.933484872382737e-05,
2479
+ "loss": 0.6729,
2480
+ "step": 7060
2481
+ },
2482
+ {
2483
+ "epoch": 0.8413547237076648,
2484
+ "grad_norm": 3.859375,
2485
+ "learning_rate": 1.8786745862613885e-05,
2486
+ "loss": 0.7747,
2487
+ "step": 7080
2488
+ },
2489
+ {
2490
+ "epoch": 0.8437314319667261,
2491
+ "grad_norm": 2.703125,
2492
+ "learning_rate": 1.8245717493762925e-05,
2493
+ "loss": 0.6807,
2494
+ "step": 7100
2495
+ },
2496
+ {
2497
+ "epoch": 0.8461081402257873,
2498
+ "grad_norm": 2.4375,
2499
+ "learning_rate": 1.7711810746294312e-05,
2500
+ "loss": 0.851,
2501
+ "step": 7120
2502
+ },
2503
+ {
2504
+ "epoch": 0.8484848484848485,
2505
+ "grad_norm": 1.4765625,
2506
+ "learning_rate": 1.7185072128862933e-05,
2507
+ "loss": 0.6449,
2508
+ "step": 7140
2509
+ },
2510
+ {
2511
+ "epoch": 0.8508615567439097,
2512
+ "grad_norm": 2.140625,
2513
+ "learning_rate": 1.6665547525707316e-05,
2514
+ "loss": 0.6683,
2515
+ "step": 7160
2516
+ },
2517
+ {
2518
+ "epoch": 0.8532382650029708,
2519
+ "grad_norm": 2.25,
2520
+ "learning_rate": 1.6153282192652698e-05,
2521
+ "loss": 0.5818,
2522
+ "step": 7180
2523
+ },
2524
+ {
2525
+ "epoch": 0.8556149732620321,
2526
+ "grad_norm": 2.1875,
2527
+ "learning_rate": 1.5648320753168844e-05,
2528
+ "loss": 0.6614,
2529
+ "step": 7200
2530
+ },
2531
+ {
2532
+ "epoch": 0.8579916815210933,
2533
+ "grad_norm": 2.5625,
2534
+ "learning_rate": 1.5150707194482695e-05,
2535
+ "loss": 0.7086,
2536
+ "step": 7220
2537
+ },
2538
+ {
2539
+ "epoch": 0.8603683897801545,
2540
+ "grad_norm": 2.734375,
2541
+ "learning_rate": 1.4660484863746938e-05,
2542
+ "loss": 0.778,
2543
+ "step": 7240
2544
+ },
2545
+ {
2546
+ "epoch": 0.8627450980392157,
2547
+ "grad_norm": 2.109375,
2548
+ "learning_rate": 1.4177696464263723e-05,
2549
+ "loss": 0.7629,
2550
+ "step": 7260
2551
+ },
2552
+ {
2553
+ "epoch": 0.8651218062982768,
2554
+ "grad_norm": 3.1875,
2555
+ "learning_rate": 1.3702384051765005e-05,
2556
+ "loss": 0.8923,
2557
+ "step": 7280
2558
+ },
2559
+ {
2560
+ "epoch": 0.8674985145573381,
2561
+ "grad_norm": 4.125,
2562
+ "learning_rate": 1.3234589030748956e-05,
2563
+ "loss": 0.6005,
2564
+ "step": 7300
2565
+ },
2566
+ {
2567
+ "epoch": 0.8698752228163993,
2568
+ "grad_norm": 2.9375,
2569
+ "learning_rate": 1.2774352150873203e-05,
2570
+ "loss": 0.741,
2571
+ "step": 7320
2572
+ },
2573
+ {
2574
+ "epoch": 0.8722519310754605,
2575
+ "grad_norm": 2.78125,
2576
+ "learning_rate": 1.2321713503405208e-05,
2577
+ "loss": 0.5276,
2578
+ "step": 7340
2579
+ },
2580
+ {
2581
+ "epoch": 0.8746286393345217,
2582
+ "grad_norm": 3.28125,
2583
+ "learning_rate": 1.187671251772987e-05,
2584
+ "loss": 0.5997,
2585
+ "step": 7360
2586
+ },
2587
+ {
2588
+ "epoch": 0.8770053475935828,
2589
+ "grad_norm": 1.8125,
2590
+ "learning_rate": 1.143938795791476e-05,
2591
+ "loss": 0.6653,
2592
+ "step": 7380
2593
+ },
2594
+ {
2595
+ "epoch": 0.8793820558526441,
2596
+ "grad_norm": 2.46875,
2597
+ "learning_rate": 1.1009777919333507e-05,
2598
+ "loss": 0.5658,
2599
+ "step": 7400
2600
+ },
2601
+ {
2602
+ "epoch": 0.8817587641117053,
2603
+ "grad_norm": 2.609375,
2604
+ "learning_rate": 1.0587919825347236e-05,
2605
+ "loss": 0.6695,
2606
+ "step": 7420
2607
+ },
2608
+ {
2609
+ "epoch": 0.8841354723707665,
2610
+ "grad_norm": 1.7734375,
2611
+ "learning_rate": 1.0173850424044596e-05,
2612
+ "loss": 0.68,
2613
+ "step": 7440
2614
+ },
2615
+ {
2616
+ "epoch": 0.8865121806298277,
2617
+ "grad_norm": 3.640625,
2618
+ "learning_rate": 9.76760578504068e-06,
2619
+ "loss": 0.8137,
2620
+ "step": 7460
2621
+ },
2622
+ {
2623
+ "epoch": 0.8888888888888888,
2624
+ "grad_norm": 1.8125,
2625
+ "learning_rate": 9.369221296335006e-06,
2626
+ "loss": 0.7129,
2627
+ "step": 7480
2628
+ },
2629
+ {
2630
+ "epoch": 0.8912655971479501,
2631
+ "grad_norm": 4.625,
2632
+ "learning_rate": 8.978731661228768e-06,
2633
+ "loss": 0.6628,
2634
+ "step": 7500
2635
+ },
2636
+ {
2637
+ "epoch": 0.8936423054070113,
2638
+ "grad_norm": 4.125,
2639
+ "learning_rate": 8.596170895301959e-06,
2640
+ "loss": 0.6621,
2641
+ "step": 7520
2642
+ },
2643
+ {
2644
+ "epoch": 0.8960190136660725,
2645
+ "grad_norm": 2.953125,
2646
+ "learning_rate": 8.221572323450222e-06,
2647
+ "loss": 0.6171,
2648
+ "step": 7540
2649
+ },
2650
+ {
2651
+ "epoch": 0.8983957219251337,
2652
+ "grad_norm": 3.078125,
2653
+ "learning_rate": 7.854968576981824e-06,
2654
+ "loss": 0.5361,
2655
+ "step": 7560
2656
+ },
2657
+ {
2658
+ "epoch": 0.9007724301841948,
2659
+ "grad_norm": 3.8125,
2660
+ "learning_rate": 7.49639159077532e-06,
2661
+ "loss": 0.6402,
2662
+ "step": 7580
2663
+ },
2664
+ {
2665
+ "epoch": 0.9031491384432561,
2666
+ "grad_norm": 3.25,
2667
+ "learning_rate": 7.145872600497561e-06,
2668
+ "loss": 0.5515,
2669
+ "step": 7600
2670
+ },
2671
+ {
2672
+ "epoch": 0.9055258467023173,
2673
+ "grad_norm": 2.078125,
2674
+ "learning_rate": 6.8034421398827765e-06,
2675
+ "loss": 0.6165,
2676
+ "step": 7620
2677
+ },
2678
+ {
2679
+ "epoch": 0.9079025549613785,
2680
+ "grad_norm": 3.78125,
2681
+ "learning_rate": 6.469130038072835e-06,
2682
+ "loss": 0.8606,
2683
+ "step": 7640
2684
+ },
2685
+ {
2686
+ "epoch": 0.9102792632204397,
2687
+ "grad_norm": 2.875,
2688
+ "learning_rate": 6.142965417018798e-06,
2689
+ "loss": 0.6364,
2690
+ "step": 7660
2691
+ },
2692
+ {
2693
+ "epoch": 0.9126559714795008,
2694
+ "grad_norm": 2.53125,
2695
+ "learning_rate": 5.824976688944051e-06,
2696
+ "loss": 0.578,
2697
+ "step": 7680
2698
+ },
2699
+ {
2700
+ "epoch": 0.9150326797385621,
2701
+ "grad_norm": 2.15625,
2702
+ "learning_rate": 5.515191553869381e-06,
2703
+ "loss": 0.5698,
2704
+ "step": 7700
2705
+ },
2706
+ {
2707
+ "epoch": 0.9174093879976233,
2708
+ "grad_norm": 2.96875,
2709
+ "learning_rate": 5.213636997200044e-06,
2710
+ "loss": 0.6748,
2711
+ "step": 7720
2712
+ },
2713
+ {
2714
+ "epoch": 0.9197860962566845,
2715
+ "grad_norm": 4.0625,
2716
+ "learning_rate": 4.920339287374942e-06,
2717
+ "loss": 0.7089,
2718
+ "step": 7740
2719
+ },
2720
+ {
2721
+ "epoch": 0.9221628045157457,
2722
+ "grad_norm": 3.578125,
2723
+ "learning_rate": 4.635323973578543e-06,
2724
+ "loss": 0.7648,
2725
+ "step": 7760
2726
+ },
2727
+ {
2728
+ "epoch": 0.9245395127748068,
2729
+ "grad_norm": 1.0390625,
2730
+ "learning_rate": 4.3586158835151495e-06,
2731
+ "loss": 0.6778,
2732
+ "step": 7780
2733
+ },
2734
+ {
2735
+ "epoch": 0.9269162210338681,
2736
+ "grad_norm": 3.578125,
2737
+ "learning_rate": 4.090239121246231e-06,
2738
+ "loss": 0.7377,
2739
+ "step": 7800
2740
+ },
2741
+ {
2742
+ "epoch": 0.9292929292929293,
2743
+ "grad_norm": 3.5,
2744
+ "learning_rate": 3.830217065090702e-06,
2745
+ "loss": 0.6488,
2746
+ "step": 7820
2747
+ },
2748
+ {
2749
+ "epoch": 0.9316696375519905,
2750
+ "grad_norm": 1.40625,
2751
+ "learning_rate": 3.5785723655884287e-06,
2752
+ "loss": 0.5778,
2753
+ "step": 7840
2754
+ },
2755
+ {
2756
+ "epoch": 0.9340463458110517,
2757
+ "grad_norm": 3.3125,
2758
+ "learning_rate": 3.335326943527117e-06,
2759
+ "loss": 0.6642,
2760
+ "step": 7860
2761
+ },
2762
+ {
2763
+ "epoch": 0.936423054070113,
2764
+ "grad_norm": 3.421875,
2765
+ "learning_rate": 3.100501988032878e-06,
2766
+ "loss": 0.5973,
2767
+ "step": 7880
2768
+ },
2769
+ {
2770
+ "epoch": 0.9387997623291741,
2771
+ "grad_norm": 4.71875,
2772
+ "learning_rate": 2.874117954724309e-06,
2773
+ "loss": 0.618,
2774
+ "step": 7900
2775
+ },
2776
+ {
2777
+ "epoch": 0.9411764705882353,
2778
+ "grad_norm": 3.921875,
2779
+ "learning_rate": 2.656194563930714e-06,
2780
+ "loss": 0.7053,
2781
+ "step": 7920
2782
+ },
2783
+ {
2784
+ "epoch": 0.9435531788472965,
2785
+ "grad_norm": 2.953125,
2786
+ "learning_rate": 2.446750798974229e-06,
2787
+ "loss": 0.6074,
2788
+ "step": 7940
2789
+ },
2790
+ {
2791
+ "epoch": 0.9459298871063577,
2792
+ "grad_norm": 2.46875,
2793
+ "learning_rate": 2.2458049045161244e-06,
2794
+ "loss": 0.5246,
2795
+ "step": 7960
2796
+ },
2797
+ {
2798
+ "epoch": 0.948306595365419,
2799
+ "grad_norm": 3.484375,
2800
+ "learning_rate": 2.0533743849676436e-06,
2801
+ "loss": 0.648,
2802
+ "step": 7980
2803
+ },
2804
+ {
2805
+ "epoch": 0.9506833036244801,
2806
+ "grad_norm": 2.953125,
2807
+ "learning_rate": 1.869476002965065e-06,
2808
+ "loss": 0.6625,
2809
+ "step": 8000
2810
+ },
2811
+ {
2812
+ "epoch": 0.9530600118835413,
2813
+ "grad_norm": 1.5859375,
2814
+ "learning_rate": 1.69412577790955e-06,
2815
+ "loss": 0.7165,
2816
+ "step": 8020
2817
+ },
2818
+ {
2819
+ "epoch": 0.9554367201426025,
2820
+ "grad_norm": 1.3046875,
2821
+ "learning_rate": 1.5273389845717245e-06,
2822
+ "loss": 0.6024,
2823
+ "step": 8040
2824
+ },
2825
+ {
2826
+ "epoch": 0.9578134284016637,
2827
+ "grad_norm": 3.6875,
2828
+ "learning_rate": 1.3691301517610554e-06,
2829
+ "loss": 0.6466,
2830
+ "step": 8060
2831
+ },
2832
+ {
2833
+ "epoch": 0.960190136660725,
2834
+ "grad_norm": 1.171875,
2835
+ "learning_rate": 1.2195130610602623e-06,
2836
+ "loss": 0.6671,
2837
+ "step": 8080
2838
+ },
2839
+ {
2840
+ "epoch": 0.9625668449197861,
2841
+ "grad_norm": 1.046875,
2842
+ "learning_rate": 1.0785007456247886e-06,
2843
+ "loss": 0.628,
2844
+ "step": 8100
2845
+ },
2846
+ {
2847
+ "epoch": 0.9649435531788473,
2848
+ "grad_norm": 2.9375,
2849
+ "learning_rate": 9.461054890474996e-07,
2850
+ "loss": 0.776,
2851
+ "step": 8120
2852
+ },
2853
+ {
2854
+ "epoch": 0.9673202614379085,
2855
+ "grad_norm": 1.671875,
2856
+ "learning_rate": 8.223388242886265e-07,
2857
+ "loss": 0.6368,
2858
+ "step": 8140
2859
+ },
2860
+ {
2861
+ "epoch": 0.9696969696969697,
2862
+ "grad_norm": 2.625,
2863
+ "learning_rate": 7.072115326711704e-07,
2864
+ "loss": 0.5938,
2865
+ "step": 8160
2866
+ },
2867
+ {
2868
+ "epoch": 0.972073677956031,
2869
+ "grad_norm": 1.171875,
2870
+ "learning_rate": 6.00733642941742e-07,
2871
+ "loss": 0.6331,
2872
+ "step": 8180
2873
+ },
2874
+ {
2875
+ "epoch": 0.9744503862150921,
2876
+ "grad_norm": 2.234375,
2877
+ "learning_rate": 5.029144303968724e-07,
2878
+ "loss": 0.566,
2879
+ "step": 8200
2880
+ },
2881
+ {
2882
+ "epoch": 0.9768270944741533,
2883
+ "grad_norm": 3.5,
2884
+ "learning_rate": 4.1376241607518074e-07,
2885
+ "loss": 0.6529,
2886
+ "step": 8220
2887
+ },
2888
+ {
2889
+ "epoch": 0.9792038027332145,
2890
+ "grad_norm": 0.91015625,
2891
+ "learning_rate": 3.332853660149904e-07,
2892
+ "loss": 0.7165,
2893
+ "step": 8240
2894
+ },
2895
+ {
2896
+ "epoch": 0.9815805109922757,
2897
+ "grad_norm": 2.5625,
2898
+ "learning_rate": 2.6149029057785936e-07,
2899
+ "loss": 0.6302,
2900
+ "step": 8260
2901
+ },
2902
+ {
2903
+ "epoch": 0.983957219251337,
2904
+ "grad_norm": 0.97265625,
2905
+ "learning_rate": 1.9838344383793505e-07,
2906
+ "loss": 0.6525,
2907
+ "step": 8280
2908
+ },
2909
+ {
2910
+ "epoch": 0.9863339275103981,
2911
+ "grad_norm": 2.90625,
2912
+ "learning_rate": 1.4397032303715697e-07,
2913
+ "loss": 0.6024,
2914
+ "step": 8300
2915
+ },
2916
+ {
2917
+ "epoch": 0.9887106357694593,
2918
+ "grad_norm": 2.421875,
2919
+ "learning_rate": 9.825566810633958e-08,
2920
+ "loss": 0.4994,
2921
+ "step": 8320
2922
+ },
2923
+ {
2924
+ "epoch": 0.9910873440285205,
2925
+ "grad_norm": 2.71875,
2926
+ "learning_rate": 6.124346125233604e-08,
2927
+ "loss": 0.7707,
2928
+ "step": 8340
2929
+ },
2930
+ {
2931
+ "epoch": 0.9934640522875817,
2932
+ "grad_norm": 1.796875,
2933
+ "learning_rate": 3.2936926611149e-08,
2934
+ "loss": 0.6724,
2935
+ "step": 8360
2936
+ },
2937
+ {
2938
+ "epoch": 0.995840760546643,
2939
+ "grad_norm": 1.5390625,
2940
+ "learning_rate": 1.3338529967010793e-08,
2941
+ "loss": 0.6268,
2942
+ "step": 8380
2943
+ },
2944
+ {
2945
+ "epoch": 0.9982174688057041,
2946
+ "grad_norm": 3.828125,
2947
+ "learning_rate": 2.449978537655273e-09,
2948
+ "loss": 0.8173,
2949
+ "step": 8400
2950
+ },
2951
+ {
2952
+ "epoch": 1.0,
2953
+ "step": 8415,
2954
+ "total_flos": 9.688250026819584e+16,
2955
+ "train_loss": 1.1022938394574842,
2956
+ "train_runtime": 7319.3616,
2957
+ "train_samples_per_second": 1.15,
2958
+ "train_steps_per_second": 1.15
2959
+ }
2960
+ ],
2961
+ "logging_steps": 20,
2962
+ "max_steps": 8415,
2963
+ "num_input_tokens_seen": 0,
2964
+ "num_train_epochs": 1,
2965
+ "save_steps": 100,
2966
+ "stateful_callbacks": {
2967
+ "TrainerControl": {
2968
+ "args": {
2969
+ "should_epoch_stop": false,
2970
+ "should_evaluate": false,
2971
+ "should_log": false,
2972
+ "should_save": true,
2973
+ "should_training_stop": false
2974
+ },
2975
+ "attributes": {}
2976
+ }
2977
+ },
2978
+ "total_flos": 9.688250026819584e+16,
2979
+ "train_batch_size": 1,
2980
+ "trial_name": null,
2981
+ "trial_params": null
2982
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5117070c8b2e3ab5a5ac01c548a28d6e941cd369ac7221133952f0d014487846
3
+ size 5112