GaetanMichelet commited on
Commit
0315e33
1 Parent(s): 3bf50cd

Model save

Browse files
README.md ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
3
+ library_name: peft
4
+ license: llama3.1
5
+ tags:
6
+ - trl
7
+ - sft
8
+ - generated_from_trainer
9
+ model-index:
10
+ - name: Llama-31-8B_task-1_180-samples_config-3_full
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # Llama-31-8B_task-1_180-samples_config-3_full
18
+
19
+ This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on the None dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.9400
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 1e-05
41
+ - train_batch_size: 1
42
+ - eval_batch_size: 1
43
+ - seed: 42
44
+ - distributed_type: multi-GPU
45
+ - gradient_accumulation_steps: 8
46
+ - total_train_batch_size: 8
47
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
+ - lr_scheduler_type: cosine
49
+ - lr_scheduler_warmup_ratio: 0.1
50
+ - num_epochs: 150
51
+
52
+ ### Training results
53
+
54
+ | Training Loss | Epoch | Step | Validation Loss |
55
+ |:-------------:|:-----:|:----:|:---------------:|
56
+ | 2.4542 | 1.0 | 17 | 2.4259 |
57
+ | 2.4022 | 2.0 | 34 | 2.3882 |
58
+ | 2.3317 | 3.0 | 51 | 2.3140 |
59
+ | 2.2607 | 4.0 | 68 | 2.2050 |
60
+ | 2.1352 | 5.0 | 85 | 2.0643 |
61
+ | 1.9456 | 6.0 | 102 | 1.8885 |
62
+ | 1.7528 | 7.0 | 119 | 1.7025 |
63
+ | 1.4935 | 8.0 | 136 | 1.4674 |
64
+ | 1.2733 | 9.0 | 153 | 1.2421 |
65
+ | 1.1154 | 10.0 | 170 | 1.1134 |
66
+ | 1.1202 | 11.0 | 187 | 1.0689 |
67
+ | 0.9449 | 12.0 | 204 | 1.0450 |
68
+ | 0.9973 | 13.0 | 221 | 1.0253 |
69
+ | 1.0562 | 14.0 | 238 | 1.0091 |
70
+ | 0.9947 | 15.0 | 255 | 0.9928 |
71
+ | 1.0096 | 16.0 | 272 | 0.9804 |
72
+ | 0.9222 | 17.0 | 289 | 0.9692 |
73
+ | 0.8838 | 18.0 | 306 | 0.9603 |
74
+ | 0.8942 | 19.0 | 323 | 0.9511 |
75
+ | 0.9058 | 20.0 | 340 | 0.9432 |
76
+ | 0.8837 | 21.0 | 357 | 0.9354 |
77
+ | 0.795 | 22.0 | 374 | 0.9315 |
78
+ | 0.8395 | 23.0 | 391 | 0.9243 |
79
+ | 0.8308 | 24.0 | 408 | 0.9169 |
80
+ | 0.7863 | 25.0 | 425 | 0.9138 |
81
+ | 0.7468 | 26.0 | 442 | 0.9068 |
82
+ | 0.7658 | 27.0 | 459 | 0.9008 |
83
+ | 0.7128 | 28.0 | 476 | 0.8992 |
84
+ | 0.6474 | 29.0 | 493 | 0.9064 |
85
+ | 0.6387 | 30.0 | 510 | 0.9089 |
86
+ | 0.6846 | 31.0 | 527 | 0.9096 |
87
+ | 0.6424 | 32.0 | 544 | 0.9173 |
88
+ | 0.6598 | 33.0 | 561 | 0.9238 |
89
+ | 0.6634 | 34.0 | 578 | 0.9290 |
90
+ | 0.5893 | 35.0 | 595 | 0.9400 |
91
+
92
+
93
+ ### Framework versions
94
+
95
+ - PEFT 0.12.0
96
+ - Transformers 4.44.0
97
+ - Pytorch 2.1.2+cu121
98
+ - Datasets 2.20.0
99
+ - Tokenizers 0.19.1
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:beeef946c1131a7c9ccdca29cae429c194bc319cce456247a84a615aba242965
3
  size 167832240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01ec7a291e7130b2a2288f91a204e2717a24f2776c6c423148f280acd735d755
3
  size 167832240
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 35.0,
3
+ "total_flos": 7.576813686279373e+16,
4
+ "train_loss": 1.1527870081052058,
5
+ "train_runtime": 5617.142,
6
+ "train_samples": 136,
7
+ "train_samples_per_second": 3.632,
8
+ "train_steps_per_second": 0.454
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 35.0,
3
+ "total_flos": 7.576813686279373e+16,
4
+ "train_loss": 1.1527870081052058,
5
+ "train_runtime": 5617.142,
6
+ "train_samples": 136,
7
+ "train_samples_per_second": 3.632,
8
+ "train_steps_per_second": 0.454
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,2417 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8991827368736267,
3
+ "best_model_checkpoint": "data/Llama-31-8B_task-1_180-samples_config-3_full/checkpoint-476",
4
+ "epoch": 35.0,
5
+ "eval_steps": 500,
6
+ "global_step": 595,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.058823529411764705,
13
+ "grad_norm": 1.858755111694336,
14
+ "learning_rate": 3.9215686274509804e-08,
15
+ "loss": 2.4463,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.11764705882352941,
20
+ "grad_norm": 1.8338021039962769,
21
+ "learning_rate": 7.843137254901961e-08,
22
+ "loss": 2.4612,
23
+ "step": 2
24
+ },
25
+ {
26
+ "epoch": 0.23529411764705882,
27
+ "grad_norm": 1.7695642709732056,
28
+ "learning_rate": 1.5686274509803921e-07,
29
+ "loss": 2.3799,
30
+ "step": 4
31
+ },
32
+ {
33
+ "epoch": 0.35294117647058826,
34
+ "grad_norm": 1.7253705263137817,
35
+ "learning_rate": 2.3529411764705883e-07,
36
+ "loss": 2.4519,
37
+ "step": 6
38
+ },
39
+ {
40
+ "epoch": 0.47058823529411764,
41
+ "grad_norm": 2.068340301513672,
42
+ "learning_rate": 3.1372549019607843e-07,
43
+ "loss": 2.4357,
44
+ "step": 8
45
+ },
46
+ {
47
+ "epoch": 0.5882352941176471,
48
+ "grad_norm": 1.71905517578125,
49
+ "learning_rate": 3.921568627450981e-07,
50
+ "loss": 2.4114,
51
+ "step": 10
52
+ },
53
+ {
54
+ "epoch": 0.7058823529411765,
55
+ "grad_norm": 1.9960722923278809,
56
+ "learning_rate": 4.7058823529411767e-07,
57
+ "loss": 2.5452,
58
+ "step": 12
59
+ },
60
+ {
61
+ "epoch": 0.8235294117647058,
62
+ "grad_norm": 1.846751093864441,
63
+ "learning_rate": 5.490196078431373e-07,
64
+ "loss": 2.4838,
65
+ "step": 14
66
+ },
67
+ {
68
+ "epoch": 0.9411764705882353,
69
+ "grad_norm": 1.7892955541610718,
70
+ "learning_rate": 6.274509803921569e-07,
71
+ "loss": 2.4542,
72
+ "step": 16
73
+ },
74
+ {
75
+ "epoch": 1.0,
76
+ "eval_loss": 2.4258534908294678,
77
+ "eval_runtime": 14.49,
78
+ "eval_samples_per_second": 2.484,
79
+ "eval_steps_per_second": 2.484,
80
+ "step": 17
81
+ },
82
+ {
83
+ "epoch": 1.0588235294117647,
84
+ "grad_norm": 1.7462923526763916,
85
+ "learning_rate": 7.058823529411766e-07,
86
+ "loss": 2.4066,
87
+ "step": 18
88
+ },
89
+ {
90
+ "epoch": 1.1764705882352942,
91
+ "grad_norm": 1.6423271894454956,
92
+ "learning_rate": 7.843137254901962e-07,
93
+ "loss": 2.4084,
94
+ "step": 20
95
+ },
96
+ {
97
+ "epoch": 1.2941176470588236,
98
+ "grad_norm": 1.6562241315841675,
99
+ "learning_rate": 8.627450980392157e-07,
100
+ "loss": 2.4685,
101
+ "step": 22
102
+ },
103
+ {
104
+ "epoch": 1.4117647058823528,
105
+ "grad_norm": 1.5601104497909546,
106
+ "learning_rate": 9.411764705882353e-07,
107
+ "loss": 2.3986,
108
+ "step": 24
109
+ },
110
+ {
111
+ "epoch": 1.5294117647058822,
112
+ "grad_norm": 1.689158320426941,
113
+ "learning_rate": 1.019607843137255e-06,
114
+ "loss": 2.512,
115
+ "step": 26
116
+ },
117
+ {
118
+ "epoch": 1.6470588235294117,
119
+ "grad_norm": 1.7012155055999756,
120
+ "learning_rate": 1.0980392156862745e-06,
121
+ "loss": 2.434,
122
+ "step": 28
123
+ },
124
+ {
125
+ "epoch": 1.7647058823529411,
126
+ "grad_norm": 1.533742070198059,
127
+ "learning_rate": 1.1764705882352942e-06,
128
+ "loss": 2.4145,
129
+ "step": 30
130
+ },
131
+ {
132
+ "epoch": 1.8823529411764706,
133
+ "grad_norm": 1.6920032501220703,
134
+ "learning_rate": 1.2549019607843137e-06,
135
+ "loss": 2.4019,
136
+ "step": 32
137
+ },
138
+ {
139
+ "epoch": 2.0,
140
+ "grad_norm": 1.5552300214767456,
141
+ "learning_rate": 1.3333333333333334e-06,
142
+ "loss": 2.4022,
143
+ "step": 34
144
+ },
145
+ {
146
+ "epoch": 2.0,
147
+ "eval_loss": 2.3881916999816895,
148
+ "eval_runtime": 14.4942,
149
+ "eval_samples_per_second": 2.484,
150
+ "eval_steps_per_second": 2.484,
151
+ "step": 34
152
+ },
153
+ {
154
+ "epoch": 2.1176470588235294,
155
+ "grad_norm": 1.5124330520629883,
156
+ "learning_rate": 1.4117647058823531e-06,
157
+ "loss": 2.3961,
158
+ "step": 36
159
+ },
160
+ {
161
+ "epoch": 2.235294117647059,
162
+ "grad_norm": 1.4209071397781372,
163
+ "learning_rate": 1.4901960784313726e-06,
164
+ "loss": 2.4343,
165
+ "step": 38
166
+ },
167
+ {
168
+ "epoch": 2.3529411764705883,
169
+ "grad_norm": 1.6290644407272339,
170
+ "learning_rate": 1.5686274509803923e-06,
171
+ "loss": 2.3528,
172
+ "step": 40
173
+ },
174
+ {
175
+ "epoch": 2.4705882352941178,
176
+ "grad_norm": 1.5042191743850708,
177
+ "learning_rate": 1.6470588235294118e-06,
178
+ "loss": 2.3555,
179
+ "step": 42
180
+ },
181
+ {
182
+ "epoch": 2.588235294117647,
183
+ "grad_norm": 1.6211644411087036,
184
+ "learning_rate": 1.7254901960784315e-06,
185
+ "loss": 2.3944,
186
+ "step": 44
187
+ },
188
+ {
189
+ "epoch": 2.7058823529411766,
190
+ "grad_norm": 1.5416394472122192,
191
+ "learning_rate": 1.8039215686274512e-06,
192
+ "loss": 2.3917,
193
+ "step": 46
194
+ },
195
+ {
196
+ "epoch": 2.8235294117647056,
197
+ "grad_norm": 1.697242259979248,
198
+ "learning_rate": 1.8823529411764707e-06,
199
+ "loss": 2.3457,
200
+ "step": 48
201
+ },
202
+ {
203
+ "epoch": 2.9411764705882355,
204
+ "grad_norm": 1.8130015134811401,
205
+ "learning_rate": 1.96078431372549e-06,
206
+ "loss": 2.3317,
207
+ "step": 50
208
+ },
209
+ {
210
+ "epoch": 3.0,
211
+ "eval_loss": 2.314044952392578,
212
+ "eval_runtime": 14.4935,
213
+ "eval_samples_per_second": 2.484,
214
+ "eval_steps_per_second": 2.484,
215
+ "step": 51
216
+ },
217
+ {
218
+ "epoch": 3.0588235294117645,
219
+ "grad_norm": 1.830562710762024,
220
+ "learning_rate": 2.03921568627451e-06,
221
+ "loss": 2.2938,
222
+ "step": 52
223
+ },
224
+ {
225
+ "epoch": 3.176470588235294,
226
+ "grad_norm": 1.8372972011566162,
227
+ "learning_rate": 2.1176470588235296e-06,
228
+ "loss": 2.3541,
229
+ "step": 54
230
+ },
231
+ {
232
+ "epoch": 3.2941176470588234,
233
+ "grad_norm": 1.8857609033584595,
234
+ "learning_rate": 2.196078431372549e-06,
235
+ "loss": 2.2888,
236
+ "step": 56
237
+ },
238
+ {
239
+ "epoch": 3.411764705882353,
240
+ "grad_norm": 1.7633429765701294,
241
+ "learning_rate": 2.274509803921569e-06,
242
+ "loss": 2.2616,
243
+ "step": 58
244
+ },
245
+ {
246
+ "epoch": 3.5294117647058822,
247
+ "grad_norm": 1.5656747817993164,
248
+ "learning_rate": 2.3529411764705885e-06,
249
+ "loss": 2.2801,
250
+ "step": 60
251
+ },
252
+ {
253
+ "epoch": 3.6470588235294117,
254
+ "grad_norm": 1.6285021305084229,
255
+ "learning_rate": 2.431372549019608e-06,
256
+ "loss": 2.3144,
257
+ "step": 62
258
+ },
259
+ {
260
+ "epoch": 3.764705882352941,
261
+ "grad_norm": 1.531112790107727,
262
+ "learning_rate": 2.5098039215686274e-06,
263
+ "loss": 2.2294,
264
+ "step": 64
265
+ },
266
+ {
267
+ "epoch": 3.8823529411764706,
268
+ "grad_norm": 1.527350664138794,
269
+ "learning_rate": 2.5882352941176473e-06,
270
+ "loss": 2.1976,
271
+ "step": 66
272
+ },
273
+ {
274
+ "epoch": 4.0,
275
+ "grad_norm": 2.009220838546753,
276
+ "learning_rate": 2.666666666666667e-06,
277
+ "loss": 2.2607,
278
+ "step": 68
279
+ },
280
+ {
281
+ "epoch": 4.0,
282
+ "eval_loss": 2.2050342559814453,
283
+ "eval_runtime": 14.4977,
284
+ "eval_samples_per_second": 2.483,
285
+ "eval_steps_per_second": 2.483,
286
+ "step": 68
287
+ },
288
+ {
289
+ "epoch": 4.117647058823529,
290
+ "grad_norm": 2.025496006011963,
291
+ "learning_rate": 2.7450980392156867e-06,
292
+ "loss": 2.1834,
293
+ "step": 70
294
+ },
295
+ {
296
+ "epoch": 4.235294117647059,
297
+ "grad_norm": 1.714682698249817,
298
+ "learning_rate": 2.8235294117647062e-06,
299
+ "loss": 2.1785,
300
+ "step": 72
301
+ },
302
+ {
303
+ "epoch": 4.352941176470588,
304
+ "grad_norm": 1.1758134365081787,
305
+ "learning_rate": 2.901960784313726e-06,
306
+ "loss": 2.2148,
307
+ "step": 74
308
+ },
309
+ {
310
+ "epoch": 4.470588235294118,
311
+ "grad_norm": 1.2965394258499146,
312
+ "learning_rate": 2.980392156862745e-06,
313
+ "loss": 2.1797,
314
+ "step": 76
315
+ },
316
+ {
317
+ "epoch": 4.588235294117647,
318
+ "grad_norm": 1.1413812637329102,
319
+ "learning_rate": 3.058823529411765e-06,
320
+ "loss": 2.1743,
321
+ "step": 78
322
+ },
323
+ {
324
+ "epoch": 4.705882352941177,
325
+ "grad_norm": 1.0636754035949707,
326
+ "learning_rate": 3.1372549019607846e-06,
327
+ "loss": 2.0559,
328
+ "step": 80
329
+ },
330
+ {
331
+ "epoch": 4.823529411764706,
332
+ "grad_norm": 1.1126306056976318,
333
+ "learning_rate": 3.2156862745098045e-06,
334
+ "loss": 2.1355,
335
+ "step": 82
336
+ },
337
+ {
338
+ "epoch": 4.9411764705882355,
339
+ "grad_norm": 1.1904844045639038,
340
+ "learning_rate": 3.2941176470588236e-06,
341
+ "loss": 2.1352,
342
+ "step": 84
343
+ },
344
+ {
345
+ "epoch": 5.0,
346
+ "eval_loss": 2.0643370151519775,
347
+ "eval_runtime": 14.5066,
348
+ "eval_samples_per_second": 2.482,
349
+ "eval_steps_per_second": 2.482,
350
+ "step": 85
351
+ },
352
+ {
353
+ "epoch": 5.0588235294117645,
354
+ "grad_norm": 1.0779309272766113,
355
+ "learning_rate": 3.3725490196078435e-06,
356
+ "loss": 2.0248,
357
+ "step": 86
358
+ },
359
+ {
360
+ "epoch": 5.176470588235294,
361
+ "grad_norm": 1.107112169265747,
362
+ "learning_rate": 3.450980392156863e-06,
363
+ "loss": 2.1049,
364
+ "step": 88
365
+ },
366
+ {
367
+ "epoch": 5.294117647058823,
368
+ "grad_norm": 0.9876514077186584,
369
+ "learning_rate": 3.529411764705883e-06,
370
+ "loss": 2.0005,
371
+ "step": 90
372
+ },
373
+ {
374
+ "epoch": 5.411764705882353,
375
+ "grad_norm": 1.073117733001709,
376
+ "learning_rate": 3.6078431372549024e-06,
377
+ "loss": 1.9499,
378
+ "step": 92
379
+ },
380
+ {
381
+ "epoch": 5.529411764705882,
382
+ "grad_norm": 1.0594408512115479,
383
+ "learning_rate": 3.6862745098039223e-06,
384
+ "loss": 1.9899,
385
+ "step": 94
386
+ },
387
+ {
388
+ "epoch": 5.647058823529412,
389
+ "grad_norm": 1.0870219469070435,
390
+ "learning_rate": 3.7647058823529414e-06,
391
+ "loss": 1.9852,
392
+ "step": 96
393
+ },
394
+ {
395
+ "epoch": 5.764705882352941,
396
+ "grad_norm": 0.9945081472396851,
397
+ "learning_rate": 3.843137254901962e-06,
398
+ "loss": 1.9981,
399
+ "step": 98
400
+ },
401
+ {
402
+ "epoch": 5.882352941176471,
403
+ "grad_norm": 0.8944886326789856,
404
+ "learning_rate": 3.92156862745098e-06,
405
+ "loss": 1.9126,
406
+ "step": 100
407
+ },
408
+ {
409
+ "epoch": 6.0,
410
+ "grad_norm": 0.8814469575881958,
411
+ "learning_rate": 4.000000000000001e-06,
412
+ "loss": 1.9456,
413
+ "step": 102
414
+ },
415
+ {
416
+ "epoch": 6.0,
417
+ "eval_loss": 1.888541340827942,
418
+ "eval_runtime": 14.5125,
419
+ "eval_samples_per_second": 2.481,
420
+ "eval_steps_per_second": 2.481,
421
+ "step": 102
422
+ },
423
+ {
424
+ "epoch": 6.117647058823529,
425
+ "grad_norm": 0.8919170498847961,
426
+ "learning_rate": 4.07843137254902e-06,
427
+ "loss": 1.8157,
428
+ "step": 104
429
+ },
430
+ {
431
+ "epoch": 6.235294117647059,
432
+ "grad_norm": 0.8234829902648926,
433
+ "learning_rate": 4.15686274509804e-06,
434
+ "loss": 1.9187,
435
+ "step": 106
436
+ },
437
+ {
438
+ "epoch": 6.352941176470588,
439
+ "grad_norm": 0.8216582536697388,
440
+ "learning_rate": 4.235294117647059e-06,
441
+ "loss": 1.8121,
442
+ "step": 108
443
+ },
444
+ {
445
+ "epoch": 6.470588235294118,
446
+ "grad_norm": 0.8760618567466736,
447
+ "learning_rate": 4.313725490196079e-06,
448
+ "loss": 1.8794,
449
+ "step": 110
450
+ },
451
+ {
452
+ "epoch": 6.588235294117647,
453
+ "grad_norm": 0.90522301197052,
454
+ "learning_rate": 4.392156862745098e-06,
455
+ "loss": 1.7899,
456
+ "step": 112
457
+ },
458
+ {
459
+ "epoch": 6.705882352941177,
460
+ "grad_norm": 0.8919849395751953,
461
+ "learning_rate": 4.4705882352941184e-06,
462
+ "loss": 1.7929,
463
+ "step": 114
464
+ },
465
+ {
466
+ "epoch": 6.823529411764706,
467
+ "grad_norm": 1.0193332433700562,
468
+ "learning_rate": 4.549019607843138e-06,
469
+ "loss": 1.7409,
470
+ "step": 116
471
+ },
472
+ {
473
+ "epoch": 6.9411764705882355,
474
+ "grad_norm": 0.9497600793838501,
475
+ "learning_rate": 4.627450980392157e-06,
476
+ "loss": 1.7528,
477
+ "step": 118
478
+ },
479
+ {
480
+ "epoch": 7.0,
481
+ "eval_loss": 1.7024633884429932,
482
+ "eval_runtime": 14.5072,
483
+ "eval_samples_per_second": 2.482,
484
+ "eval_steps_per_second": 2.482,
485
+ "step": 119
486
+ },
487
+ {
488
+ "epoch": 7.0588235294117645,
489
+ "grad_norm": 0.9311454892158508,
490
+ "learning_rate": 4.705882352941177e-06,
491
+ "loss": 1.7333,
492
+ "step": 120
493
+ },
494
+ {
495
+ "epoch": 7.176470588235294,
496
+ "grad_norm": 1.0313152074813843,
497
+ "learning_rate": 4.784313725490196e-06,
498
+ "loss": 1.7217,
499
+ "step": 122
500
+ },
501
+ {
502
+ "epoch": 7.294117647058823,
503
+ "grad_norm": 1.1278079748153687,
504
+ "learning_rate": 4.862745098039216e-06,
505
+ "loss": 1.6414,
506
+ "step": 124
507
+ },
508
+ {
509
+ "epoch": 7.411764705882353,
510
+ "grad_norm": 0.9751306176185608,
511
+ "learning_rate": 4.941176470588236e-06,
512
+ "loss": 1.6047,
513
+ "step": 126
514
+ },
515
+ {
516
+ "epoch": 7.529411764705882,
517
+ "grad_norm": 0.9619643688201904,
518
+ "learning_rate": 5.019607843137255e-06,
519
+ "loss": 1.5733,
520
+ "step": 128
521
+ },
522
+ {
523
+ "epoch": 7.647058823529412,
524
+ "grad_norm": 0.9418209195137024,
525
+ "learning_rate": 5.098039215686274e-06,
526
+ "loss": 1.5655,
527
+ "step": 130
528
+ },
529
+ {
530
+ "epoch": 7.764705882352941,
531
+ "grad_norm": 0.986770749092102,
532
+ "learning_rate": 5.176470588235295e-06,
533
+ "loss": 1.5325,
534
+ "step": 132
535
+ },
536
+ {
537
+ "epoch": 7.882352941176471,
538
+ "grad_norm": 0.8657909631729126,
539
+ "learning_rate": 5.254901960784314e-06,
540
+ "loss": 1.5042,
541
+ "step": 134
542
+ },
543
+ {
544
+ "epoch": 8.0,
545
+ "grad_norm": 0.8987972736358643,
546
+ "learning_rate": 5.333333333333334e-06,
547
+ "loss": 1.4935,
548
+ "step": 136
549
+ },
550
+ {
551
+ "epoch": 8.0,
552
+ "eval_loss": 1.467383861541748,
553
+ "eval_runtime": 14.5108,
554
+ "eval_samples_per_second": 2.481,
555
+ "eval_steps_per_second": 2.481,
556
+ "step": 136
557
+ },
558
+ {
559
+ "epoch": 8.117647058823529,
560
+ "grad_norm": 0.8275275826454163,
561
+ "learning_rate": 5.411764705882353e-06,
562
+ "loss": 1.4215,
563
+ "step": 138
564
+ },
565
+ {
566
+ "epoch": 8.235294117647058,
567
+ "grad_norm": 0.9540057182312012,
568
+ "learning_rate": 5.4901960784313735e-06,
569
+ "loss": 1.4698,
570
+ "step": 140
571
+ },
572
+ {
573
+ "epoch": 8.352941176470589,
574
+ "grad_norm": 0.9684072136878967,
575
+ "learning_rate": 5.568627450980393e-06,
576
+ "loss": 1.4359,
577
+ "step": 142
578
+ },
579
+ {
580
+ "epoch": 8.470588235294118,
581
+ "grad_norm": 0.9229031801223755,
582
+ "learning_rate": 5.6470588235294125e-06,
583
+ "loss": 1.3994,
584
+ "step": 144
585
+ },
586
+ {
587
+ "epoch": 8.588235294117647,
588
+ "grad_norm": 0.8458110094070435,
589
+ "learning_rate": 5.725490196078431e-06,
590
+ "loss": 1.3096,
591
+ "step": 146
592
+ },
593
+ {
594
+ "epoch": 8.705882352941176,
595
+ "grad_norm": 0.9069352746009827,
596
+ "learning_rate": 5.803921568627452e-06,
597
+ "loss": 1.2347,
598
+ "step": 148
599
+ },
600
+ {
601
+ "epoch": 8.823529411764707,
602
+ "grad_norm": 0.8469833731651306,
603
+ "learning_rate": 5.882352941176471e-06,
604
+ "loss": 1.332,
605
+ "step": 150
606
+ },
607
+ {
608
+ "epoch": 8.941176470588236,
609
+ "grad_norm": 0.8933460116386414,
610
+ "learning_rate": 5.96078431372549e-06,
611
+ "loss": 1.2733,
612
+ "step": 152
613
+ },
614
+ {
615
+ "epoch": 9.0,
616
+ "eval_loss": 1.2421215772628784,
617
+ "eval_runtime": 14.5471,
618
+ "eval_samples_per_second": 2.475,
619
+ "eval_steps_per_second": 2.475,
620
+ "step": 153
621
+ },
622
+ {
623
+ "epoch": 9.058823529411764,
624
+ "grad_norm": 0.8019786477088928,
625
+ "learning_rate": 6.03921568627451e-06,
626
+ "loss": 1.1929,
627
+ "step": 154
628
+ },
629
+ {
630
+ "epoch": 9.176470588235293,
631
+ "grad_norm": 0.7300643920898438,
632
+ "learning_rate": 6.11764705882353e-06,
633
+ "loss": 1.2392,
634
+ "step": 156
635
+ },
636
+ {
637
+ "epoch": 9.294117647058824,
638
+ "grad_norm": 0.809948742389679,
639
+ "learning_rate": 6.19607843137255e-06,
640
+ "loss": 1.1685,
641
+ "step": 158
642
+ },
643
+ {
644
+ "epoch": 9.411764705882353,
645
+ "grad_norm": 0.6852974891662598,
646
+ "learning_rate": 6.274509803921569e-06,
647
+ "loss": 1.168,
648
+ "step": 160
649
+ },
650
+ {
651
+ "epoch": 9.529411764705882,
652
+ "grad_norm": 0.709697961807251,
653
+ "learning_rate": 6.352941176470589e-06,
654
+ "loss": 1.1333,
655
+ "step": 162
656
+ },
657
+ {
658
+ "epoch": 9.647058823529411,
659
+ "grad_norm": 0.7923583388328552,
660
+ "learning_rate": 6.431372549019609e-06,
661
+ "loss": 1.1475,
662
+ "step": 164
663
+ },
664
+ {
665
+ "epoch": 9.764705882352942,
666
+ "grad_norm": 0.7233794927597046,
667
+ "learning_rate": 6.5098039215686285e-06,
668
+ "loss": 1.1775,
669
+ "step": 166
670
+ },
671
+ {
672
+ "epoch": 9.882352941176471,
673
+ "grad_norm": 0.7074316740036011,
674
+ "learning_rate": 6.588235294117647e-06,
675
+ "loss": 1.1279,
676
+ "step": 168
677
+ },
678
+ {
679
+ "epoch": 10.0,
680
+ "grad_norm": 0.6581458449363708,
681
+ "learning_rate": 6.666666666666667e-06,
682
+ "loss": 1.1154,
683
+ "step": 170
684
+ },
685
+ {
686
+ "epoch": 10.0,
687
+ "eval_loss": 1.1133772134780884,
688
+ "eval_runtime": 14.5122,
689
+ "eval_samples_per_second": 2.481,
690
+ "eval_steps_per_second": 2.481,
691
+ "step": 170
692
+ },
693
+ {
694
+ "epoch": 10.117647058823529,
695
+ "grad_norm": 0.6955820918083191,
696
+ "learning_rate": 6.745098039215687e-06,
697
+ "loss": 1.0662,
698
+ "step": 172
699
+ },
700
+ {
701
+ "epoch": 10.235294117647058,
702
+ "grad_norm": 0.5870165824890137,
703
+ "learning_rate": 6.8235294117647065e-06,
704
+ "loss": 1.0219,
705
+ "step": 174
706
+ },
707
+ {
708
+ "epoch": 10.352941176470589,
709
+ "grad_norm": 0.6177704334259033,
710
+ "learning_rate": 6.901960784313726e-06,
711
+ "loss": 1.0537,
712
+ "step": 176
713
+ },
714
+ {
715
+ "epoch": 10.470588235294118,
716
+ "grad_norm": 0.6390775442123413,
717
+ "learning_rate": 6.9803921568627454e-06,
718
+ "loss": 1.1001,
719
+ "step": 178
720
+ },
721
+ {
722
+ "epoch": 10.588235294117647,
723
+ "grad_norm": 0.4973801374435425,
724
+ "learning_rate": 7.058823529411766e-06,
725
+ "loss": 1.0578,
726
+ "step": 180
727
+ },
728
+ {
729
+ "epoch": 10.705882352941176,
730
+ "grad_norm": 0.518943190574646,
731
+ "learning_rate": 7.137254901960785e-06,
732
+ "loss": 1.1447,
733
+ "step": 182
734
+ },
735
+ {
736
+ "epoch": 10.823529411764707,
737
+ "grad_norm": 0.6414965987205505,
738
+ "learning_rate": 7.215686274509805e-06,
739
+ "loss": 1.0872,
740
+ "step": 184
741
+ },
742
+ {
743
+ "epoch": 10.941176470588236,
744
+ "grad_norm": 0.508786678314209,
745
+ "learning_rate": 7.294117647058823e-06,
746
+ "loss": 1.1202,
747
+ "step": 186
748
+ },
749
+ {
750
+ "epoch": 11.0,
751
+ "eval_loss": 1.0689375400543213,
752
+ "eval_runtime": 14.505,
753
+ "eval_samples_per_second": 2.482,
754
+ "eval_steps_per_second": 2.482,
755
+ "step": 187
756
+ },
757
+ {
758
+ "epoch": 11.058823529411764,
759
+ "grad_norm": 0.48530295491218567,
760
+ "learning_rate": 7.372549019607845e-06,
761
+ "loss": 1.0999,
762
+ "step": 188
763
+ },
764
+ {
765
+ "epoch": 11.176470588235293,
766
+ "grad_norm": 0.5133592486381531,
767
+ "learning_rate": 7.450980392156863e-06,
768
+ "loss": 1.0864,
769
+ "step": 190
770
+ },
771
+ {
772
+ "epoch": 11.294117647058824,
773
+ "grad_norm": 0.49263596534729004,
774
+ "learning_rate": 7.529411764705883e-06,
775
+ "loss": 1.0535,
776
+ "step": 192
777
+ },
778
+ {
779
+ "epoch": 11.411764705882353,
780
+ "grad_norm": 0.4610048532485962,
781
+ "learning_rate": 7.607843137254902e-06,
782
+ "loss": 1.0462,
783
+ "step": 194
784
+ },
785
+ {
786
+ "epoch": 11.529411764705882,
787
+ "grad_norm": 0.5121297836303711,
788
+ "learning_rate": 7.686274509803923e-06,
789
+ "loss": 1.0862,
790
+ "step": 196
791
+ },
792
+ {
793
+ "epoch": 11.647058823529411,
794
+ "grad_norm": 0.5441015958786011,
795
+ "learning_rate": 7.764705882352941e-06,
796
+ "loss": 1.0068,
797
+ "step": 198
798
+ },
799
+ {
800
+ "epoch": 11.764705882352942,
801
+ "grad_norm": 0.5135095119476318,
802
+ "learning_rate": 7.84313725490196e-06,
803
+ "loss": 1.0548,
804
+ "step": 200
805
+ },
806
+ {
807
+ "epoch": 11.882352941176471,
808
+ "grad_norm": 0.4792177081108093,
809
+ "learning_rate": 7.92156862745098e-06,
810
+ "loss": 0.9711,
811
+ "step": 202
812
+ },
813
+ {
814
+ "epoch": 12.0,
815
+ "grad_norm": 0.45314979553222656,
816
+ "learning_rate": 8.000000000000001e-06,
817
+ "loss": 0.9449,
818
+ "step": 204
819
+ },
820
+ {
821
+ "epoch": 12.0,
822
+ "eval_loss": 1.0450434684753418,
823
+ "eval_runtime": 14.5066,
824
+ "eval_samples_per_second": 2.482,
825
+ "eval_steps_per_second": 2.482,
826
+ "step": 204
827
+ },
828
+ {
829
+ "epoch": 12.117647058823529,
830
+ "grad_norm": 0.5007625818252563,
831
+ "learning_rate": 8.07843137254902e-06,
832
+ "loss": 1.0258,
833
+ "step": 206
834
+ },
835
+ {
836
+ "epoch": 12.235294117647058,
837
+ "grad_norm": 0.5184361934661865,
838
+ "learning_rate": 8.15686274509804e-06,
839
+ "loss": 1.0845,
840
+ "step": 208
841
+ },
842
+ {
843
+ "epoch": 12.352941176470589,
844
+ "grad_norm": 0.44266751408576965,
845
+ "learning_rate": 8.23529411764706e-06,
846
+ "loss": 1.0005,
847
+ "step": 210
848
+ },
849
+ {
850
+ "epoch": 12.470588235294118,
851
+ "grad_norm": 0.5165805220603943,
852
+ "learning_rate": 8.31372549019608e-06,
853
+ "loss": 1.0242,
854
+ "step": 212
855
+ },
856
+ {
857
+ "epoch": 12.588235294117647,
858
+ "grad_norm": 0.5037981867790222,
859
+ "learning_rate": 8.392156862745099e-06,
860
+ "loss": 0.9857,
861
+ "step": 214
862
+ },
863
+ {
864
+ "epoch": 12.705882352941176,
865
+ "grad_norm": 0.5604737997055054,
866
+ "learning_rate": 8.470588235294118e-06,
867
+ "loss": 1.0086,
868
+ "step": 216
869
+ },
870
+ {
871
+ "epoch": 12.823529411764707,
872
+ "grad_norm": 0.6752682328224182,
873
+ "learning_rate": 8.549019607843138e-06,
874
+ "loss": 1.0277,
875
+ "step": 218
876
+ },
877
+ {
878
+ "epoch": 12.941176470588236,
879
+ "grad_norm": 0.5517321228981018,
880
+ "learning_rate": 8.627450980392157e-06,
881
+ "loss": 0.9973,
882
+ "step": 220
883
+ },
884
+ {
885
+ "epoch": 13.0,
886
+ "eval_loss": 1.0252662897109985,
887
+ "eval_runtime": 14.5028,
888
+ "eval_samples_per_second": 2.482,
889
+ "eval_steps_per_second": 2.482,
890
+ "step": 221
891
+ },
892
+ {
893
+ "epoch": 13.058823529411764,
894
+ "grad_norm": 0.6104453802108765,
895
+ "learning_rate": 8.705882352941177e-06,
896
+ "loss": 0.9835,
897
+ "step": 222
898
+ },
899
+ {
900
+ "epoch": 13.176470588235293,
901
+ "grad_norm": 0.47119539976119995,
902
+ "learning_rate": 8.784313725490196e-06,
903
+ "loss": 1.0076,
904
+ "step": 224
905
+ },
906
+ {
907
+ "epoch": 13.294117647058824,
908
+ "grad_norm": 0.4882214367389679,
909
+ "learning_rate": 8.862745098039216e-06,
910
+ "loss": 0.9808,
911
+ "step": 226
912
+ },
913
+ {
914
+ "epoch": 13.411764705882353,
915
+ "grad_norm": 0.7123433947563171,
916
+ "learning_rate": 8.941176470588237e-06,
917
+ "loss": 0.9676,
918
+ "step": 228
919
+ },
920
+ {
921
+ "epoch": 13.529411764705882,
922
+ "grad_norm": 0.5918748378753662,
923
+ "learning_rate": 9.019607843137256e-06,
924
+ "loss": 1.0068,
925
+ "step": 230
926
+ },
927
+ {
928
+ "epoch": 13.647058823529411,
929
+ "grad_norm": 0.5302197337150574,
930
+ "learning_rate": 9.098039215686276e-06,
931
+ "loss": 0.9573,
932
+ "step": 232
933
+ },
934
+ {
935
+ "epoch": 13.764705882352942,
936
+ "grad_norm": 0.5693833827972412,
937
+ "learning_rate": 9.176470588235294e-06,
938
+ "loss": 0.9914,
939
+ "step": 234
940
+ },
941
+ {
942
+ "epoch": 13.882352941176471,
943
+ "grad_norm": 0.490904837846756,
944
+ "learning_rate": 9.254901960784315e-06,
945
+ "loss": 1.032,
946
+ "step": 236
947
+ },
948
+ {
949
+ "epoch": 14.0,
950
+ "grad_norm": 0.5507678389549255,
951
+ "learning_rate": 9.333333333333334e-06,
952
+ "loss": 1.0562,
953
+ "step": 238
954
+ },
955
+ {
956
+ "epoch": 14.0,
957
+ "eval_loss": 1.0090599060058594,
958
+ "eval_runtime": 14.506,
959
+ "eval_samples_per_second": 2.482,
960
+ "eval_steps_per_second": 2.482,
961
+ "step": 238
962
+ },
963
+ {
964
+ "epoch": 14.117647058823529,
965
+ "grad_norm": 0.6389086246490479,
966
+ "learning_rate": 9.411764705882354e-06,
967
+ "loss": 0.9853,
968
+ "step": 240
969
+ },
970
+ {
971
+ "epoch": 14.235294117647058,
972
+ "grad_norm": 0.5049781203269958,
973
+ "learning_rate": 9.490196078431373e-06,
974
+ "loss": 1.0067,
975
+ "step": 242
976
+ },
977
+ {
978
+ "epoch": 14.352941176470589,
979
+ "grad_norm": 0.7086266279220581,
980
+ "learning_rate": 9.568627450980393e-06,
981
+ "loss": 0.9387,
982
+ "step": 244
983
+ },
984
+ {
985
+ "epoch": 14.470588235294118,
986
+ "grad_norm": 0.5628448128700256,
987
+ "learning_rate": 9.647058823529412e-06,
988
+ "loss": 1.0068,
989
+ "step": 246
990
+ },
991
+ {
992
+ "epoch": 14.588235294117647,
993
+ "grad_norm": 0.6910731196403503,
994
+ "learning_rate": 9.725490196078432e-06,
995
+ "loss": 1.0007,
996
+ "step": 248
997
+ },
998
+ {
999
+ "epoch": 14.705882352941176,
1000
+ "grad_norm": 0.6134346127510071,
1001
+ "learning_rate": 9.803921568627451e-06,
1002
+ "loss": 0.9456,
1003
+ "step": 250
1004
+ },
1005
+ {
1006
+ "epoch": 14.823529411764707,
1007
+ "grad_norm": 0.6747128963470459,
1008
+ "learning_rate": 9.882352941176472e-06,
1009
+ "loss": 0.9506,
1010
+ "step": 252
1011
+ },
1012
+ {
1013
+ "epoch": 14.941176470588236,
1014
+ "grad_norm": 0.5889897346496582,
1015
+ "learning_rate": 9.960784313725492e-06,
1016
+ "loss": 0.9947,
1017
+ "step": 254
1018
+ },
1019
+ {
1020
+ "epoch": 15.0,
1021
+ "eval_loss": 0.9928128719329834,
1022
+ "eval_runtime": 14.4936,
1023
+ "eval_samples_per_second": 2.484,
1024
+ "eval_steps_per_second": 2.484,
1025
+ "step": 255
1026
+ },
1027
+ {
1028
+ "epoch": 15.058823529411764,
1029
+ "grad_norm": 0.5487807989120483,
1030
+ "learning_rate": 9.999995315380667e-06,
1031
+ "loss": 0.9354,
1032
+ "step": 256
1033
+ },
1034
+ {
1035
+ "epoch": 15.176470588235293,
1036
+ "grad_norm": 0.6178866624832153,
1037
+ "learning_rate": 9.99995783847866e-06,
1038
+ "loss": 0.9655,
1039
+ "step": 258
1040
+ },
1041
+ {
1042
+ "epoch": 15.294117647058824,
1043
+ "grad_norm": 0.5696916580200195,
1044
+ "learning_rate": 9.999882884955554e-06,
1045
+ "loss": 0.9468,
1046
+ "step": 260
1047
+ },
1048
+ {
1049
+ "epoch": 15.411764705882353,
1050
+ "grad_norm": 0.6009863615036011,
1051
+ "learning_rate": 9.99977045537315e-06,
1052
+ "loss": 0.9852,
1053
+ "step": 262
1054
+ },
1055
+ {
1056
+ "epoch": 15.529411764705882,
1057
+ "grad_norm": 0.6040264368057251,
1058
+ "learning_rate": 9.999620550574155e-06,
1059
+ "loss": 0.9553,
1060
+ "step": 264
1061
+ },
1062
+ {
1063
+ "epoch": 15.647058823529411,
1064
+ "grad_norm": 0.6321269869804382,
1065
+ "learning_rate": 9.999433171682158e-06,
1066
+ "loss": 0.9419,
1067
+ "step": 266
1068
+ },
1069
+ {
1070
+ "epoch": 15.764705882352942,
1071
+ "grad_norm": 0.6273146867752075,
1072
+ "learning_rate": 9.999208320101643e-06,
1073
+ "loss": 0.9715,
1074
+ "step": 268
1075
+ },
1076
+ {
1077
+ "epoch": 15.882352941176471,
1078
+ "grad_norm": 0.6734570860862732,
1079
+ "learning_rate": 9.998945997517957e-06,
1080
+ "loss": 0.918,
1081
+ "step": 270
1082
+ },
1083
+ {
1084
+ "epoch": 16.0,
1085
+ "grad_norm": 0.7102432250976562,
1086
+ "learning_rate": 9.99864620589731e-06,
1087
+ "loss": 1.0096,
1088
+ "step": 272
1089
+ },
1090
+ {
1091
+ "epoch": 16.0,
1092
+ "eval_loss": 0.9803969264030457,
1093
+ "eval_runtime": 14.5083,
1094
+ "eval_samples_per_second": 2.481,
1095
+ "eval_steps_per_second": 2.481,
1096
+ "step": 272
1097
+ },
1098
+ {
1099
+ "epoch": 16.11764705882353,
1100
+ "grad_norm": 0.6154859066009521,
1101
+ "learning_rate": 9.998308947486753e-06,
1102
+ "loss": 0.8898,
1103
+ "step": 274
1104
+ },
1105
+ {
1106
+ "epoch": 16.235294117647058,
1107
+ "grad_norm": 0.6435267329216003,
1108
+ "learning_rate": 9.997934224814173e-06,
1109
+ "loss": 0.9271,
1110
+ "step": 276
1111
+ },
1112
+ {
1113
+ "epoch": 16.352941176470587,
1114
+ "grad_norm": 0.7057787775993347,
1115
+ "learning_rate": 9.997522040688258e-06,
1116
+ "loss": 1.0,
1117
+ "step": 278
1118
+ },
1119
+ {
1120
+ "epoch": 16.470588235294116,
1121
+ "grad_norm": 0.6257563233375549,
1122
+ "learning_rate": 9.997072398198492e-06,
1123
+ "loss": 0.973,
1124
+ "step": 280
1125
+ },
1126
+ {
1127
+ "epoch": 16.58823529411765,
1128
+ "grad_norm": 0.6798095703125,
1129
+ "learning_rate": 9.996585300715117e-06,
1130
+ "loss": 0.9625,
1131
+ "step": 282
1132
+ },
1133
+ {
1134
+ "epoch": 16.705882352941178,
1135
+ "grad_norm": 0.7027468681335449,
1136
+ "learning_rate": 9.996060751889114e-06,
1137
+ "loss": 0.9529,
1138
+ "step": 284
1139
+ },
1140
+ {
1141
+ "epoch": 16.823529411764707,
1142
+ "grad_norm": 0.6210634708404541,
1143
+ "learning_rate": 9.995498755652186e-06,
1144
+ "loss": 0.8968,
1145
+ "step": 286
1146
+ },
1147
+ {
1148
+ "epoch": 16.941176470588236,
1149
+ "grad_norm": 0.6995490789413452,
1150
+ "learning_rate": 9.994899316216709e-06,
1151
+ "loss": 0.9222,
1152
+ "step": 288
1153
+ },
1154
+ {
1155
+ "epoch": 17.0,
1156
+ "eval_loss": 0.9691942930221558,
1157
+ "eval_runtime": 14.5044,
1158
+ "eval_samples_per_second": 2.482,
1159
+ "eval_steps_per_second": 2.482,
1160
+ "step": 289
1161
+ },
1162
+ {
1163
+ "epoch": 17.058823529411764,
1164
+ "grad_norm": 0.6503624320030212,
1165
+ "learning_rate": 9.994262438075713e-06,
1166
+ "loss": 0.9487,
1167
+ "step": 290
1168
+ },
1169
+ {
1170
+ "epoch": 17.176470588235293,
1171
+ "grad_norm": 0.6647483706474304,
1172
+ "learning_rate": 9.993588126002848e-06,
1173
+ "loss": 0.9163,
1174
+ "step": 292
1175
+ },
1176
+ {
1177
+ "epoch": 17.294117647058822,
1178
+ "grad_norm": 0.7215944528579712,
1179
+ "learning_rate": 9.992876385052346e-06,
1180
+ "loss": 0.8638,
1181
+ "step": 294
1182
+ },
1183
+ {
1184
+ "epoch": 17.41176470588235,
1185
+ "grad_norm": 0.7234969139099121,
1186
+ "learning_rate": 9.992127220558976e-06,
1187
+ "loss": 0.9037,
1188
+ "step": 296
1189
+ },
1190
+ {
1191
+ "epoch": 17.529411764705884,
1192
+ "grad_norm": 0.7656229138374329,
1193
+ "learning_rate": 9.991340638138022e-06,
1194
+ "loss": 0.9633,
1195
+ "step": 298
1196
+ },
1197
+ {
1198
+ "epoch": 17.647058823529413,
1199
+ "grad_norm": 0.6850258111953735,
1200
+ "learning_rate": 9.990516643685222e-06,
1201
+ "loss": 0.9458,
1202
+ "step": 300
1203
+ },
1204
+ {
1205
+ "epoch": 17.764705882352942,
1206
+ "grad_norm": 0.7975447773933411,
1207
+ "learning_rate": 9.98965524337673e-06,
1208
+ "loss": 0.9801,
1209
+ "step": 302
1210
+ },
1211
+ {
1212
+ "epoch": 17.88235294117647,
1213
+ "grad_norm": 0.7075424790382385,
1214
+ "learning_rate": 9.988756443669081e-06,
1215
+ "loss": 0.888,
1216
+ "step": 304
1217
+ },
1218
+ {
1219
+ "epoch": 18.0,
1220
+ "grad_norm": 0.85096675157547,
1221
+ "learning_rate": 9.987820251299121e-06,
1222
+ "loss": 0.8838,
1223
+ "step": 306
1224
+ },
1225
+ {
1226
+ "epoch": 18.0,
1227
+ "eval_loss": 0.9602956771850586,
1228
+ "eval_runtime": 14.5129,
1229
+ "eval_samples_per_second": 2.481,
1230
+ "eval_steps_per_second": 2.481,
1231
+ "step": 306
1232
+ },
1233
+ {
1234
+ "epoch": 18.11764705882353,
1235
+ "grad_norm": 0.698685884475708,
1236
+ "learning_rate": 9.98684667328398e-06,
1237
+ "loss": 0.8838,
1238
+ "step": 308
1239
+ },
1240
+ {
1241
+ "epoch": 18.235294117647058,
1242
+ "grad_norm": 0.7671274542808533,
1243
+ "learning_rate": 9.985835716921e-06,
1244
+ "loss": 0.9012,
1245
+ "step": 310
1246
+ },
1247
+ {
1248
+ "epoch": 18.352941176470587,
1249
+ "grad_norm": 0.8342521786689758,
1250
+ "learning_rate": 9.984787389787689e-06,
1251
+ "loss": 0.9412,
1252
+ "step": 312
1253
+ },
1254
+ {
1255
+ "epoch": 18.470588235294116,
1256
+ "grad_norm": 0.6886960864067078,
1257
+ "learning_rate": 9.983701699741668e-06,
1258
+ "loss": 0.8946,
1259
+ "step": 314
1260
+ },
1261
+ {
1262
+ "epoch": 18.58823529411765,
1263
+ "grad_norm": 0.7856888175010681,
1264
+ "learning_rate": 9.982578654920601e-06,
1265
+ "loss": 0.9169,
1266
+ "step": 316
1267
+ },
1268
+ {
1269
+ "epoch": 18.705882352941178,
1270
+ "grad_norm": 0.7338317036628723,
1271
+ "learning_rate": 9.981418263742148e-06,
1272
+ "loss": 0.8584,
1273
+ "step": 318
1274
+ },
1275
+ {
1276
+ "epoch": 18.823529411764707,
1277
+ "grad_norm": 0.727165699005127,
1278
+ "learning_rate": 9.980220534903889e-06,
1279
+ "loss": 0.9385,
1280
+ "step": 320
1281
+ },
1282
+ {
1283
+ "epoch": 18.941176470588236,
1284
+ "grad_norm": 0.777866542339325,
1285
+ "learning_rate": 9.978985477383264e-06,
1286
+ "loss": 0.8942,
1287
+ "step": 322
1288
+ },
1289
+ {
1290
+ "epoch": 19.0,
1291
+ "eval_loss": 0.951096773147583,
1292
+ "eval_runtime": 14.4924,
1293
+ "eval_samples_per_second": 2.484,
1294
+ "eval_steps_per_second": 2.484,
1295
+ "step": 323
1296
+ },
1297
+ {
1298
+ "epoch": 19.058823529411764,
1299
+ "grad_norm": 0.6845978498458862,
1300
+ "learning_rate": 9.97771310043751e-06,
1301
+ "loss": 0.8752,
1302
+ "step": 324
1303
+ },
1304
+ {
1305
+ "epoch": 19.176470588235293,
1306
+ "grad_norm": 0.7632399201393127,
1307
+ "learning_rate": 9.97640341360358e-06,
1308
+ "loss": 0.9616,
1309
+ "step": 326
1310
+ },
1311
+ {
1312
+ "epoch": 19.294117647058822,
1313
+ "grad_norm": 0.7852567434310913,
1314
+ "learning_rate": 9.975056426698094e-06,
1315
+ "loss": 0.8884,
1316
+ "step": 328
1317
+ },
1318
+ {
1319
+ "epoch": 19.41176470588235,
1320
+ "grad_norm": 0.7355157136917114,
1321
+ "learning_rate": 9.973672149817232e-06,
1322
+ "loss": 0.8175,
1323
+ "step": 330
1324
+ },
1325
+ {
1326
+ "epoch": 19.529411764705884,
1327
+ "grad_norm": 0.7707788348197937,
1328
+ "learning_rate": 9.972250593336689e-06,
1329
+ "loss": 0.8878,
1330
+ "step": 332
1331
+ },
1332
+ {
1333
+ "epoch": 19.647058823529413,
1334
+ "grad_norm": 1.0082019567489624,
1335
+ "learning_rate": 9.970791767911581e-06,
1336
+ "loss": 0.9118,
1337
+ "step": 334
1338
+ },
1339
+ {
1340
+ "epoch": 19.764705882352942,
1341
+ "grad_norm": 0.8013073205947876,
1342
+ "learning_rate": 9.96929568447637e-06,
1343
+ "loss": 0.8724,
1344
+ "step": 336
1345
+ },
1346
+ {
1347
+ "epoch": 19.88235294117647,
1348
+ "grad_norm": 0.6911207437515259,
1349
+ "learning_rate": 9.967762354244778e-06,
1350
+ "loss": 0.8832,
1351
+ "step": 338
1352
+ },
1353
+ {
1354
+ "epoch": 20.0,
1355
+ "grad_norm": 0.8336138725280762,
1356
+ "learning_rate": 9.966191788709716e-06,
1357
+ "loss": 0.9058,
1358
+ "step": 340
1359
+ },
1360
+ {
1361
+ "epoch": 20.0,
1362
+ "eval_loss": 0.9431850910186768,
1363
+ "eval_runtime": 14.5083,
1364
+ "eval_samples_per_second": 2.481,
1365
+ "eval_steps_per_second": 2.481,
1366
+ "step": 340
1367
+ },
1368
+ {
1369
+ "epoch": 20.11764705882353,
1370
+ "grad_norm": 0.7745249271392822,
1371
+ "learning_rate": 9.964583999643174e-06,
1372
+ "loss": 0.878,
1373
+ "step": 342
1374
+ },
1375
+ {
1376
+ "epoch": 20.235294117647058,
1377
+ "grad_norm": 0.7922182083129883,
1378
+ "learning_rate": 9.962938999096159e-06,
1379
+ "loss": 0.8275,
1380
+ "step": 344
1381
+ },
1382
+ {
1383
+ "epoch": 20.352941176470587,
1384
+ "grad_norm": 0.8610040545463562,
1385
+ "learning_rate": 9.961256799398584e-06,
1386
+ "loss": 0.94,
1387
+ "step": 346
1388
+ },
1389
+ {
1390
+ "epoch": 20.470588235294116,
1391
+ "grad_norm": 0.9406768083572388,
1392
+ "learning_rate": 9.95953741315919e-06,
1393
+ "loss": 0.8779,
1394
+ "step": 348
1395
+ },
1396
+ {
1397
+ "epoch": 20.58823529411765,
1398
+ "grad_norm": 0.8344603180885315,
1399
+ "learning_rate": 9.957780853265441e-06,
1400
+ "loss": 0.8318,
1401
+ "step": 350
1402
+ },
1403
+ {
1404
+ "epoch": 20.705882352941178,
1405
+ "grad_norm": 0.8624390363693237,
1406
+ "learning_rate": 9.955987132883435e-06,
1407
+ "loss": 0.8644,
1408
+ "step": 352
1409
+ },
1410
+ {
1411
+ "epoch": 20.823529411764707,
1412
+ "grad_norm": 0.7996507287025452,
1413
+ "learning_rate": 9.954156265457801e-06,
1414
+ "loss": 0.8656,
1415
+ "step": 354
1416
+ },
1417
+ {
1418
+ "epoch": 20.941176470588236,
1419
+ "grad_norm": 0.9234054684638977,
1420
+ "learning_rate": 9.952288264711601e-06,
1421
+ "loss": 0.8837,
1422
+ "step": 356
1423
+ },
1424
+ {
1425
+ "epoch": 21.0,
1426
+ "eval_loss": 0.9354000091552734,
1427
+ "eval_runtime": 14.5044,
1428
+ "eval_samples_per_second": 2.482,
1429
+ "eval_steps_per_second": 2.482,
1430
+ "step": 357
1431
+ },
1432
+ {
1433
+ "epoch": 21.058823529411764,
1434
+ "grad_norm": 0.793875515460968,
1435
+ "learning_rate": 9.950383144646221e-06,
1436
+ "loss": 0.8662,
1437
+ "step": 358
1438
+ },
1439
+ {
1440
+ "epoch": 21.176470588235293,
1441
+ "grad_norm": 0.8161793947219849,
1442
+ "learning_rate": 9.948440919541277e-06,
1443
+ "loss": 0.8713,
1444
+ "step": 360
1445
+ },
1446
+ {
1447
+ "epoch": 21.294117647058822,
1448
+ "grad_norm": 0.9452466368675232,
1449
+ "learning_rate": 9.946461603954499e-06,
1450
+ "loss": 0.9299,
1451
+ "step": 362
1452
+ },
1453
+ {
1454
+ "epoch": 21.41176470588235,
1455
+ "grad_norm": 0.8712689876556396,
1456
+ "learning_rate": 9.944445212721619e-06,
1457
+ "loss": 0.84,
1458
+ "step": 364
1459
+ },
1460
+ {
1461
+ "epoch": 21.529411764705884,
1462
+ "grad_norm": 0.8613099455833435,
1463
+ "learning_rate": 9.942391760956277e-06,
1464
+ "loss": 0.8523,
1465
+ "step": 366
1466
+ },
1467
+ {
1468
+ "epoch": 21.647058823529413,
1469
+ "grad_norm": 1.0285900831222534,
1470
+ "learning_rate": 9.940301264049885e-06,
1471
+ "loss": 0.8411,
1472
+ "step": 368
1473
+ },
1474
+ {
1475
+ "epoch": 21.764705882352942,
1476
+ "grad_norm": 0.9434134364128113,
1477
+ "learning_rate": 9.938173737671531e-06,
1478
+ "loss": 0.819,
1479
+ "step": 370
1480
+ },
1481
+ {
1482
+ "epoch": 21.88235294117647,
1483
+ "grad_norm": 0.9282283782958984,
1484
+ "learning_rate": 9.936009197767847e-06,
1485
+ "loss": 0.8783,
1486
+ "step": 372
1487
+ },
1488
+ {
1489
+ "epoch": 22.0,
1490
+ "grad_norm": 0.9603204131126404,
1491
+ "learning_rate": 9.933807660562898e-06,
1492
+ "loss": 0.795,
1493
+ "step": 374
1494
+ },
1495
+ {
1496
+ "epoch": 22.0,
1497
+ "eval_loss": 0.9314696788787842,
1498
+ "eval_runtime": 14.5014,
1499
+ "eval_samples_per_second": 2.483,
1500
+ "eval_steps_per_second": 2.483,
1501
+ "step": 374
1502
+ },
1503
+ {
1504
+ "epoch": 22.11764705882353,
1505
+ "grad_norm": 0.9283419847488403,
1506
+ "learning_rate": 9.931569142558057e-06,
1507
+ "loss": 0.8911,
1508
+ "step": 376
1509
+ },
1510
+ {
1511
+ "epoch": 22.235294117647058,
1512
+ "grad_norm": 0.985173761844635,
1513
+ "learning_rate": 9.929293660531889e-06,
1514
+ "loss": 0.8351,
1515
+ "step": 378
1516
+ },
1517
+ {
1518
+ "epoch": 22.352941176470587,
1519
+ "grad_norm": 0.9488443732261658,
1520
+ "learning_rate": 9.926981231540007e-06,
1521
+ "loss": 0.8245,
1522
+ "step": 380
1523
+ },
1524
+ {
1525
+ "epoch": 22.470588235294116,
1526
+ "grad_norm": 1.0252861976623535,
1527
+ "learning_rate": 9.924631872914967e-06,
1528
+ "loss": 0.8096,
1529
+ "step": 382
1530
+ },
1531
+ {
1532
+ "epoch": 22.58823529411765,
1533
+ "grad_norm": 0.8986847996711731,
1534
+ "learning_rate": 9.922245602266119e-06,
1535
+ "loss": 0.8311,
1536
+ "step": 384
1537
+ },
1538
+ {
1539
+ "epoch": 22.705882352941178,
1540
+ "grad_norm": 0.9069613218307495,
1541
+ "learning_rate": 9.919822437479488e-06,
1542
+ "loss": 0.7961,
1543
+ "step": 386
1544
+ },
1545
+ {
1546
+ "epoch": 22.823529411764707,
1547
+ "grad_norm": 0.8006130456924438,
1548
+ "learning_rate": 9.91736239671763e-06,
1549
+ "loss": 0.866,
1550
+ "step": 388
1551
+ },
1552
+ {
1553
+ "epoch": 22.941176470588236,
1554
+ "grad_norm": 0.8258039355278015,
1555
+ "learning_rate": 9.91486549841951e-06,
1556
+ "loss": 0.8395,
1557
+ "step": 390
1558
+ },
1559
+ {
1560
+ "epoch": 23.0,
1561
+ "eval_loss": 0.9243198037147522,
1562
+ "eval_runtime": 14.5142,
1563
+ "eval_samples_per_second": 2.48,
1564
+ "eval_steps_per_second": 2.48,
1565
+ "step": 391
1566
+ },
1567
+ {
1568
+ "epoch": 23.058823529411764,
1569
+ "grad_norm": 1.0394818782806396,
1570
+ "learning_rate": 9.912331761300341e-06,
1571
+ "loss": 0.787,
1572
+ "step": 392
1573
+ },
1574
+ {
1575
+ "epoch": 23.176470588235293,
1576
+ "grad_norm": 0.9367055892944336,
1577
+ "learning_rate": 9.909761204351469e-06,
1578
+ "loss": 0.8501,
1579
+ "step": 394
1580
+ },
1581
+ {
1582
+ "epoch": 23.294117647058822,
1583
+ "grad_norm": 1.0531871318817139,
1584
+ "learning_rate": 9.90715384684021e-06,
1585
+ "loss": 0.8855,
1586
+ "step": 396
1587
+ },
1588
+ {
1589
+ "epoch": 23.41176470588235,
1590
+ "grad_norm": 0.9447432160377502,
1591
+ "learning_rate": 9.904509708309723e-06,
1592
+ "loss": 0.7717,
1593
+ "step": 398
1594
+ },
1595
+ {
1596
+ "epoch": 23.529411764705884,
1597
+ "grad_norm": 0.9809987545013428,
1598
+ "learning_rate": 9.901828808578846e-06,
1599
+ "loss": 0.7949,
1600
+ "step": 400
1601
+ },
1602
+ {
1603
+ "epoch": 23.647058823529413,
1604
+ "grad_norm": 0.9656151533126831,
1605
+ "learning_rate": 9.899111167741966e-06,
1606
+ "loss": 0.8286,
1607
+ "step": 402
1608
+ },
1609
+ {
1610
+ "epoch": 23.764705882352942,
1611
+ "grad_norm": 1.0195831060409546,
1612
+ "learning_rate": 9.896356806168851e-06,
1613
+ "loss": 0.8478,
1614
+ "step": 404
1615
+ },
1616
+ {
1617
+ "epoch": 23.88235294117647,
1618
+ "grad_norm": 1.144056797027588,
1619
+ "learning_rate": 9.89356574450451e-06,
1620
+ "loss": 0.7723,
1621
+ "step": 406
1622
+ },
1623
+ {
1624
+ "epoch": 24.0,
1625
+ "grad_norm": 1.0349133014678955,
1626
+ "learning_rate": 9.890738003669029e-06,
1627
+ "loss": 0.8308,
1628
+ "step": 408
1629
+ },
1630
+ {
1631
+ "epoch": 24.0,
1632
+ "eval_loss": 0.9168965816497803,
1633
+ "eval_runtime": 14.494,
1634
+ "eval_samples_per_second": 2.484,
1635
+ "eval_steps_per_second": 2.484,
1636
+ "step": 408
1637
+ },
1638
+ {
1639
+ "epoch": 24.11764705882353,
1640
+ "grad_norm": 1.003952980041504,
1641
+ "learning_rate": 9.887873604857424e-06,
1642
+ "loss": 0.8492,
1643
+ "step": 410
1644
+ },
1645
+ {
1646
+ "epoch": 24.235294117647058,
1647
+ "grad_norm": 1.1212753057479858,
1648
+ "learning_rate": 9.884972569539471e-06,
1649
+ "loss": 0.8037,
1650
+ "step": 412
1651
+ },
1652
+ {
1653
+ "epoch": 24.352941176470587,
1654
+ "grad_norm": 0.995343029499054,
1655
+ "learning_rate": 9.882034919459556e-06,
1656
+ "loss": 0.765,
1657
+ "step": 414
1658
+ },
1659
+ {
1660
+ "epoch": 24.470588235294116,
1661
+ "grad_norm": 1.0651168823242188,
1662
+ "learning_rate": 9.879060676636502e-06,
1663
+ "loss": 0.8008,
1664
+ "step": 416
1665
+ },
1666
+ {
1667
+ "epoch": 24.58823529411765,
1668
+ "grad_norm": 1.1323087215423584,
1669
+ "learning_rate": 9.876049863363415e-06,
1670
+ "loss": 0.8154,
1671
+ "step": 418
1672
+ },
1673
+ {
1674
+ "epoch": 24.705882352941178,
1675
+ "grad_norm": 1.118166446685791,
1676
+ "learning_rate": 9.873002502207502e-06,
1677
+ "loss": 0.7665,
1678
+ "step": 420
1679
+ },
1680
+ {
1681
+ "epoch": 24.823529411764707,
1682
+ "grad_norm": 1.1308856010437012,
1683
+ "learning_rate": 9.86991861600992e-06,
1684
+ "loss": 0.8056,
1685
+ "step": 422
1686
+ },
1687
+ {
1688
+ "epoch": 24.941176470588236,
1689
+ "grad_norm": 1.0739870071411133,
1690
+ "learning_rate": 9.866798227885588e-06,
1691
+ "loss": 0.7863,
1692
+ "step": 424
1693
+ },
1694
+ {
1695
+ "epoch": 25.0,
1696
+ "eval_loss": 0.9137818217277527,
1697
+ "eval_runtime": 14.4961,
1698
+ "eval_samples_per_second": 2.483,
1699
+ "eval_steps_per_second": 2.483,
1700
+ "step": 425
1701
+ },
1702
+ {
1703
+ "epoch": 25.058823529411764,
1704
+ "grad_norm": 0.947708785533905,
1705
+ "learning_rate": 9.863641361223025e-06,
1706
+ "loss": 0.746,
1707
+ "step": 426
1708
+ },
1709
+ {
1710
+ "epoch": 25.176470588235293,
1711
+ "grad_norm": 1.226585030555725,
1712
+ "learning_rate": 9.860448039684169e-06,
1713
+ "loss": 0.7622,
1714
+ "step": 428
1715
+ },
1716
+ {
1717
+ "epoch": 25.294117647058822,
1718
+ "grad_norm": 1.2553542852401733,
1719
+ "learning_rate": 9.857218287204204e-06,
1720
+ "loss": 0.784,
1721
+ "step": 430
1722
+ },
1723
+ {
1724
+ "epoch": 25.41176470588235,
1725
+ "grad_norm": 1.130286455154419,
1726
+ "learning_rate": 9.853952127991374e-06,
1727
+ "loss": 0.78,
1728
+ "step": 432
1729
+ },
1730
+ {
1731
+ "epoch": 25.529411764705884,
1732
+ "grad_norm": 1.2538301944732666,
1733
+ "learning_rate": 9.850649586526808e-06,
1734
+ "loss": 0.7608,
1735
+ "step": 434
1736
+ },
1737
+ {
1738
+ "epoch": 25.647058823529413,
1739
+ "grad_norm": 1.170310378074646,
1740
+ "learning_rate": 9.847310687564335e-06,
1741
+ "loss": 0.8389,
1742
+ "step": 436
1743
+ },
1744
+ {
1745
+ "epoch": 25.764705882352942,
1746
+ "grad_norm": 0.9732166528701782,
1747
+ "learning_rate": 9.843935456130295e-06,
1748
+ "loss": 0.8158,
1749
+ "step": 438
1750
+ },
1751
+ {
1752
+ "epoch": 25.88235294117647,
1753
+ "grad_norm": 1.2474738359451294,
1754
+ "learning_rate": 9.840523917523354e-06,
1755
+ "loss": 0.7528,
1756
+ "step": 440
1757
+ },
1758
+ {
1759
+ "epoch": 26.0,
1760
+ "grad_norm": 1.130893349647522,
1761
+ "learning_rate": 9.83707609731432e-06,
1762
+ "loss": 0.7468,
1763
+ "step": 442
1764
+ },
1765
+ {
1766
+ "epoch": 26.0,
1767
+ "eval_loss": 0.9068209528923035,
1768
+ "eval_runtime": 14.5043,
1769
+ "eval_samples_per_second": 2.482,
1770
+ "eval_steps_per_second": 2.482,
1771
+ "step": 442
1772
+ },
1773
+ {
1774
+ "epoch": 26.11764705882353,
1775
+ "grad_norm": 1.0452311038970947,
1776
+ "learning_rate": 9.833592021345938e-06,
1777
+ "loss": 0.7589,
1778
+ "step": 444
1779
+ },
1780
+ {
1781
+ "epoch": 26.235294117647058,
1782
+ "grad_norm": 0.9809611439704895,
1783
+ "learning_rate": 9.830071715732708e-06,
1784
+ "loss": 0.8016,
1785
+ "step": 446
1786
+ },
1787
+ {
1788
+ "epoch": 26.352941176470587,
1789
+ "grad_norm": 1.0656489133834839,
1790
+ "learning_rate": 9.826515206860683e-06,
1791
+ "loss": 0.7417,
1792
+ "step": 448
1793
+ },
1794
+ {
1795
+ "epoch": 26.470588235294116,
1796
+ "grad_norm": 1.1188890933990479,
1797
+ "learning_rate": 9.822922521387277e-06,
1798
+ "loss": 0.7569,
1799
+ "step": 450
1800
+ },
1801
+ {
1802
+ "epoch": 26.58823529411765,
1803
+ "grad_norm": 1.087983250617981,
1804
+ "learning_rate": 9.819293686241057e-06,
1805
+ "loss": 0.7596,
1806
+ "step": 452
1807
+ },
1808
+ {
1809
+ "epoch": 26.705882352941178,
1810
+ "grad_norm": 1.0073840618133545,
1811
+ "learning_rate": 9.81562872862155e-06,
1812
+ "loss": 0.7423,
1813
+ "step": 454
1814
+ },
1815
+ {
1816
+ "epoch": 26.823529411764707,
1817
+ "grad_norm": 1.0083576440811157,
1818
+ "learning_rate": 9.811927675999035e-06,
1819
+ "loss": 0.7533,
1820
+ "step": 456
1821
+ },
1822
+ {
1823
+ "epoch": 26.941176470588236,
1824
+ "grad_norm": 1.0545302629470825,
1825
+ "learning_rate": 9.808190556114333e-06,
1826
+ "loss": 0.7658,
1827
+ "step": 458
1828
+ },
1829
+ {
1830
+ "epoch": 27.0,
1831
+ "eval_loss": 0.9007807970046997,
1832
+ "eval_runtime": 14.5307,
1833
+ "eval_samples_per_second": 2.478,
1834
+ "eval_steps_per_second": 2.478,
1835
+ "step": 459
1836
+ },
1837
+ {
1838
+ "epoch": 27.058823529411764,
1839
+ "grad_norm": 0.9539656043052673,
1840
+ "learning_rate": 9.804417396978605e-06,
1841
+ "loss": 0.7658,
1842
+ "step": 460
1843
+ },
1844
+ {
1845
+ "epoch": 27.176470588235293,
1846
+ "grad_norm": 1.044712781906128,
1847
+ "learning_rate": 9.800608226873143e-06,
1848
+ "loss": 0.6566,
1849
+ "step": 462
1850
+ },
1851
+ {
1852
+ "epoch": 27.294117647058822,
1853
+ "grad_norm": 1.3112603425979614,
1854
+ "learning_rate": 9.796763074349147e-06,
1855
+ "loss": 0.8283,
1856
+ "step": 464
1857
+ },
1858
+ {
1859
+ "epoch": 27.41176470588235,
1860
+ "grad_norm": 1.1589727401733398,
1861
+ "learning_rate": 9.792881968227533e-06,
1862
+ "loss": 0.6633,
1863
+ "step": 466
1864
+ },
1865
+ {
1866
+ "epoch": 27.529411764705884,
1867
+ "grad_norm": 0.9757166504859924,
1868
+ "learning_rate": 9.788964937598688e-06,
1869
+ "loss": 0.7725,
1870
+ "step": 468
1871
+ },
1872
+ {
1873
+ "epoch": 27.647058823529413,
1874
+ "grad_norm": 1.1313936710357666,
1875
+ "learning_rate": 9.78501201182228e-06,
1876
+ "loss": 0.7581,
1877
+ "step": 470
1878
+ },
1879
+ {
1880
+ "epoch": 27.764705882352942,
1881
+ "grad_norm": 1.1437342166900635,
1882
+ "learning_rate": 9.781023220527013e-06,
1883
+ "loss": 0.7226,
1884
+ "step": 472
1885
+ },
1886
+ {
1887
+ "epoch": 27.88235294117647,
1888
+ "grad_norm": 1.1630206108093262,
1889
+ "learning_rate": 9.776998593610428e-06,
1890
+ "loss": 0.7693,
1891
+ "step": 474
1892
+ },
1893
+ {
1894
+ "epoch": 28.0,
1895
+ "grad_norm": 1.0083279609680176,
1896
+ "learning_rate": 9.77293816123866e-06,
1897
+ "loss": 0.7128,
1898
+ "step": 476
1899
+ },
1900
+ {
1901
+ "epoch": 28.0,
1902
+ "eval_loss": 0.8991827368736267,
1903
+ "eval_runtime": 14.5006,
1904
+ "eval_samples_per_second": 2.483,
1905
+ "eval_steps_per_second": 2.483,
1906
+ "step": 476
1907
+ },
1908
+ {
1909
+ "epoch": 28.11764705882353,
1910
+ "grad_norm": 1.1530383825302124,
1911
+ "learning_rate": 9.768841953846225e-06,
1912
+ "loss": 0.6908,
1913
+ "step": 478
1914
+ },
1915
+ {
1916
+ "epoch": 28.235294117647058,
1917
+ "grad_norm": 1.0489223003387451,
1918
+ "learning_rate": 9.764710002135784e-06,
1919
+ "loss": 0.675,
1920
+ "step": 480
1921
+ },
1922
+ {
1923
+ "epoch": 28.352941176470587,
1924
+ "grad_norm": 1.2449612617492676,
1925
+ "learning_rate": 9.760542337077914e-06,
1926
+ "loss": 0.7516,
1927
+ "step": 482
1928
+ },
1929
+ {
1930
+ "epoch": 28.470588235294116,
1931
+ "grad_norm": 1.1940374374389648,
1932
+ "learning_rate": 9.75633898991088e-06,
1933
+ "loss": 0.7681,
1934
+ "step": 484
1935
+ },
1936
+ {
1937
+ "epoch": 28.58823529411765,
1938
+ "grad_norm": 1.1063061952590942,
1939
+ "learning_rate": 9.752099992140401e-06,
1940
+ "loss": 0.7693,
1941
+ "step": 486
1942
+ },
1943
+ {
1944
+ "epoch": 28.705882352941178,
1945
+ "grad_norm": 1.1479785442352295,
1946
+ "learning_rate": 9.747825375539401e-06,
1947
+ "loss": 0.7108,
1948
+ "step": 488
1949
+ },
1950
+ {
1951
+ "epoch": 28.823529411764707,
1952
+ "grad_norm": 1.2331879138946533,
1953
+ "learning_rate": 9.743515172147793e-06,
1954
+ "loss": 0.7786,
1955
+ "step": 490
1956
+ },
1957
+ {
1958
+ "epoch": 28.941176470588236,
1959
+ "grad_norm": 1.1679853200912476,
1960
+ "learning_rate": 9.739169414272219e-06,
1961
+ "loss": 0.6474,
1962
+ "step": 492
1963
+ },
1964
+ {
1965
+ "epoch": 29.0,
1966
+ "eval_loss": 0.906444787979126,
1967
+ "eval_runtime": 14.4932,
1968
+ "eval_samples_per_second": 2.484,
1969
+ "eval_steps_per_second": 2.484,
1970
+ "step": 493
1971
+ },
1972
+ {
1973
+ "epoch": 29.058823529411764,
1974
+ "grad_norm": 1.0444296598434448,
1975
+ "learning_rate": 9.734788134485817e-06,
1976
+ "loss": 0.6756,
1977
+ "step": 494
1978
+ },
1979
+ {
1980
+ "epoch": 29.176470588235293,
1981
+ "grad_norm": 1.5380338430404663,
1982
+ "learning_rate": 9.73037136562798e-06,
1983
+ "loss": 0.7099,
1984
+ "step": 496
1985
+ },
1986
+ {
1987
+ "epoch": 29.294117647058822,
1988
+ "grad_norm": 1.166580080986023,
1989
+ "learning_rate": 9.7259191408041e-06,
1990
+ "loss": 0.7595,
1991
+ "step": 498
1992
+ },
1993
+ {
1994
+ "epoch": 29.41176470588235,
1995
+ "grad_norm": 1.2345106601715088,
1996
+ "learning_rate": 9.721431493385322e-06,
1997
+ "loss": 0.7026,
1998
+ "step": 500
1999
+ },
2000
+ {
2001
+ "epoch": 29.529411764705884,
2002
+ "grad_norm": 1.0901451110839844,
2003
+ "learning_rate": 9.71690845700831e-06,
2004
+ "loss": 0.6719,
2005
+ "step": 502
2006
+ },
2007
+ {
2008
+ "epoch": 29.647058823529413,
2009
+ "grad_norm": 1.1619518995285034,
2010
+ "learning_rate": 9.71235006557497e-06,
2011
+ "loss": 0.7517,
2012
+ "step": 504
2013
+ },
2014
+ {
2015
+ "epoch": 29.764705882352942,
2016
+ "grad_norm": 1.1259740591049194,
2017
+ "learning_rate": 9.707756353252213e-06,
2018
+ "loss": 0.7052,
2019
+ "step": 506
2020
+ },
2021
+ {
2022
+ "epoch": 29.88235294117647,
2023
+ "grad_norm": 1.1172682046890259,
2024
+ "learning_rate": 9.70312735447169e-06,
2025
+ "loss": 0.655,
2026
+ "step": 508
2027
+ },
2028
+ {
2029
+ "epoch": 30.0,
2030
+ "grad_norm": 1.306216835975647,
2031
+ "learning_rate": 9.698463103929542e-06,
2032
+ "loss": 0.6387,
2033
+ "step": 510
2034
+ },
2035
+ {
2036
+ "epoch": 30.0,
2037
+ "eval_loss": 0.9089268445968628,
2038
+ "eval_runtime": 14.4939,
2039
+ "eval_samples_per_second": 2.484,
2040
+ "eval_steps_per_second": 2.484,
2041
+ "step": 510
2042
+ },
2043
+ {
2044
+ "epoch": 30.11764705882353,
2045
+ "grad_norm": 1.258402705192566,
2046
+ "learning_rate": 9.693763636586135e-06,
2047
+ "loss": 0.716,
2048
+ "step": 512
2049
+ },
2050
+ {
2051
+ "epoch": 30.235294117647058,
2052
+ "grad_norm": 1.143336296081543,
2053
+ "learning_rate": 9.689028987665797e-06,
2054
+ "loss": 0.6283,
2055
+ "step": 514
2056
+ },
2057
+ {
2058
+ "epoch": 30.352941176470587,
2059
+ "grad_norm": 1.1861103773117065,
2060
+ "learning_rate": 9.684259192656554e-06,
2061
+ "loss": 0.6445,
2062
+ "step": 516
2063
+ },
2064
+ {
2065
+ "epoch": 30.470588235294116,
2066
+ "grad_norm": 1.2192977666854858,
2067
+ "learning_rate": 9.679454287309868e-06,
2068
+ "loss": 0.6928,
2069
+ "step": 518
2070
+ },
2071
+ {
2072
+ "epoch": 30.58823529411765,
2073
+ "grad_norm": 1.3194884061813354,
2074
+ "learning_rate": 9.674614307640368e-06,
2075
+ "loss": 0.733,
2076
+ "step": 520
2077
+ },
2078
+ {
2079
+ "epoch": 30.705882352941178,
2080
+ "grad_norm": 1.3853224515914917,
2081
+ "learning_rate": 9.669739289925578e-06,
2082
+ "loss": 0.6438,
2083
+ "step": 522
2084
+ },
2085
+ {
2086
+ "epoch": 30.823529411764707,
2087
+ "grad_norm": 1.4584524631500244,
2088
+ "learning_rate": 9.664829270705638e-06,
2089
+ "loss": 0.7003,
2090
+ "step": 524
2091
+ },
2092
+ {
2093
+ "epoch": 30.941176470588236,
2094
+ "grad_norm": 1.637763500213623,
2095
+ "learning_rate": 9.659884286783052e-06,
2096
+ "loss": 0.6846,
2097
+ "step": 526
2098
+ },
2099
+ {
2100
+ "epoch": 31.0,
2101
+ "eval_loss": 0.9096066355705261,
2102
+ "eval_runtime": 14.5201,
2103
+ "eval_samples_per_second": 2.479,
2104
+ "eval_steps_per_second": 2.479,
2105
+ "step": 527
2106
+ },
2107
+ {
2108
+ "epoch": 31.058823529411764,
2109
+ "grad_norm": 1.399101972579956,
2110
+ "learning_rate": 9.654904375222384e-06,
2111
+ "loss": 0.617,
2112
+ "step": 528
2113
+ },
2114
+ {
2115
+ "epoch": 31.176470588235293,
2116
+ "grad_norm": 1.3545421361923218,
2117
+ "learning_rate": 9.649889573350006e-06,
2118
+ "loss": 0.6534,
2119
+ "step": 530
2120
+ },
2121
+ {
2122
+ "epoch": 31.294117647058822,
2123
+ "grad_norm": 1.4606151580810547,
2124
+ "learning_rate": 9.644839918753796e-06,
2125
+ "loss": 0.6815,
2126
+ "step": 532
2127
+ },
2128
+ {
2129
+ "epoch": 31.41176470588235,
2130
+ "grad_norm": 1.435264229774475,
2131
+ "learning_rate": 9.639755449282874e-06,
2132
+ "loss": 0.6696,
2133
+ "step": 534
2134
+ },
2135
+ {
2136
+ "epoch": 31.529411764705884,
2137
+ "grad_norm": 1.2791359424591064,
2138
+ "learning_rate": 9.634636203047309e-06,
2139
+ "loss": 0.642,
2140
+ "step": 536
2141
+ },
2142
+ {
2143
+ "epoch": 31.647058823529413,
2144
+ "grad_norm": 1.2923133373260498,
2145
+ "learning_rate": 9.629482218417834e-06,
2146
+ "loss": 0.712,
2147
+ "step": 538
2148
+ },
2149
+ {
2150
+ "epoch": 31.764705882352942,
2151
+ "grad_norm": 1.2450653314590454,
2152
+ "learning_rate": 9.62429353402556e-06,
2153
+ "loss": 0.6357,
2154
+ "step": 540
2155
+ },
2156
+ {
2157
+ "epoch": 31.88235294117647,
2158
+ "grad_norm": 1.31989586353302,
2159
+ "learning_rate": 9.619070188761687e-06,
2160
+ "loss": 0.6692,
2161
+ "step": 542
2162
+ },
2163
+ {
2164
+ "epoch": 32.0,
2165
+ "grad_norm": 1.3321213722229004,
2166
+ "learning_rate": 9.613812221777212e-06,
2167
+ "loss": 0.6424,
2168
+ "step": 544
2169
+ },
2170
+ {
2171
+ "epoch": 32.0,
2172
+ "eval_loss": 0.9172940850257874,
2173
+ "eval_runtime": 14.4984,
2174
+ "eval_samples_per_second": 2.483,
2175
+ "eval_steps_per_second": 2.483,
2176
+ "step": 544
2177
+ },
2178
+ {
2179
+ "epoch": 32.11764705882353,
2180
+ "grad_norm": 1.2186630964279175,
2181
+ "learning_rate": 9.608519672482635e-06,
2182
+ "loss": 0.5872,
2183
+ "step": 546
2184
+ },
2185
+ {
2186
+ "epoch": 32.23529411764706,
2187
+ "grad_norm": 1.5495742559432983,
2188
+ "learning_rate": 9.603192580547664e-06,
2189
+ "loss": 0.6069,
2190
+ "step": 548
2191
+ },
2192
+ {
2193
+ "epoch": 32.35294117647059,
2194
+ "grad_norm": 1.551956295967102,
2195
+ "learning_rate": 9.597830985900913e-06,
2196
+ "loss": 0.6971,
2197
+ "step": 550
2198
+ },
2199
+ {
2200
+ "epoch": 32.470588235294116,
2201
+ "grad_norm": 1.5809985399246216,
2202
+ "learning_rate": 9.592434928729617e-06,
2203
+ "loss": 0.6887,
2204
+ "step": 552
2205
+ },
2206
+ {
2207
+ "epoch": 32.588235294117645,
2208
+ "grad_norm": 1.5837764739990234,
2209
+ "learning_rate": 9.58700444947931e-06,
2210
+ "loss": 0.6228,
2211
+ "step": 554
2212
+ },
2213
+ {
2214
+ "epoch": 32.705882352941174,
2215
+ "grad_norm": 1.4612311124801636,
2216
+ "learning_rate": 9.581539588853539e-06,
2217
+ "loss": 0.6002,
2218
+ "step": 556
2219
+ },
2220
+ {
2221
+ "epoch": 32.8235294117647,
2222
+ "grad_norm": 1.4830561876296997,
2223
+ "learning_rate": 9.576040387813553e-06,
2224
+ "loss": 0.6673,
2225
+ "step": 558
2226
+ },
2227
+ {
2228
+ "epoch": 32.94117647058823,
2229
+ "grad_norm": 1.5311380624771118,
2230
+ "learning_rate": 9.570506887577994e-06,
2231
+ "loss": 0.6598,
2232
+ "step": 560
2233
+ },
2234
+ {
2235
+ "epoch": 33.0,
2236
+ "eval_loss": 0.9237830638885498,
2237
+ "eval_runtime": 14.5075,
2238
+ "eval_samples_per_second": 2.481,
2239
+ "eval_steps_per_second": 2.481,
2240
+ "step": 561
2241
+ },
2242
+ {
2243
+ "epoch": 33.05882352941177,
2244
+ "grad_norm": 1.3405797481536865,
2245
+ "learning_rate": 9.564939129622591e-06,
2246
+ "loss": 0.6105,
2247
+ "step": 562
2248
+ },
2249
+ {
2250
+ "epoch": 33.1764705882353,
2251
+ "grad_norm": 1.4336148500442505,
2252
+ "learning_rate": 9.559337155679843e-06,
2253
+ "loss": 0.572,
2254
+ "step": 564
2255
+ },
2256
+ {
2257
+ "epoch": 33.294117647058826,
2258
+ "grad_norm": 1.4750621318817139,
2259
+ "learning_rate": 9.553701007738717e-06,
2260
+ "loss": 0.5598,
2261
+ "step": 566
2262
+ },
2263
+ {
2264
+ "epoch": 33.411764705882355,
2265
+ "grad_norm": 1.4853854179382324,
2266
+ "learning_rate": 9.54803072804433e-06,
2267
+ "loss": 0.6175,
2268
+ "step": 568
2269
+ },
2270
+ {
2271
+ "epoch": 33.529411764705884,
2272
+ "grad_norm": 1.5611326694488525,
2273
+ "learning_rate": 9.542326359097619e-06,
2274
+ "loss": 0.5898,
2275
+ "step": 570
2276
+ },
2277
+ {
2278
+ "epoch": 33.64705882352941,
2279
+ "grad_norm": 1.4341068267822266,
2280
+ "learning_rate": 9.536587943655043e-06,
2281
+ "loss": 0.6158,
2282
+ "step": 572
2283
+ },
2284
+ {
2285
+ "epoch": 33.76470588235294,
2286
+ "grad_norm": 1.3872367143630981,
2287
+ "learning_rate": 9.530815524728245e-06,
2288
+ "loss": 0.6776,
2289
+ "step": 574
2290
+ },
2291
+ {
2292
+ "epoch": 33.88235294117647,
2293
+ "grad_norm": 1.3841159343719482,
2294
+ "learning_rate": 9.525009145583746e-06,
2295
+ "loss": 0.6208,
2296
+ "step": 576
2297
+ },
2298
+ {
2299
+ "epoch": 34.0,
2300
+ "grad_norm": 1.5026782751083374,
2301
+ "learning_rate": 9.519168849742603e-06,
2302
+ "loss": 0.6634,
2303
+ "step": 578
2304
+ },
2305
+ {
2306
+ "epoch": 34.0,
2307
+ "eval_loss": 0.9289535880088806,
2308
+ "eval_runtime": 14.5116,
2309
+ "eval_samples_per_second": 2.481,
2310
+ "eval_steps_per_second": 2.481,
2311
+ "step": 578
2312
+ },
2313
+ {
2314
+ "epoch": 34.11764705882353,
2315
+ "grad_norm": 1.5542646646499634,
2316
+ "learning_rate": 9.5132946809801e-06,
2317
+ "loss": 0.6259,
2318
+ "step": 580
2319
+ },
2320
+ {
2321
+ "epoch": 34.23529411764706,
2322
+ "grad_norm": 1.337219476699829,
2323
+ "learning_rate": 9.507386683325404e-06,
2324
+ "loss": 0.5992,
2325
+ "step": 582
2326
+ },
2327
+ {
2328
+ "epoch": 34.35294117647059,
2329
+ "grad_norm": 1.744362235069275,
2330
+ "learning_rate": 9.501444901061248e-06,
2331
+ "loss": 0.5903,
2332
+ "step": 584
2333
+ },
2334
+ {
2335
+ "epoch": 34.470588235294116,
2336
+ "grad_norm": 1.5578619241714478,
2337
+ "learning_rate": 9.495469378723592e-06,
2338
+ "loss": 0.5371,
2339
+ "step": 586
2340
+ },
2341
+ {
2342
+ "epoch": 34.588235294117645,
2343
+ "grad_norm": 1.679646611213684,
2344
+ "learning_rate": 9.489460161101291e-06,
2345
+ "loss": 0.617,
2346
+ "step": 588
2347
+ },
2348
+ {
2349
+ "epoch": 34.705882352941174,
2350
+ "grad_norm": 1.5505824089050293,
2351
+ "learning_rate": 9.483417293235759e-06,
2352
+ "loss": 0.6008,
2353
+ "step": 590
2354
+ },
2355
+ {
2356
+ "epoch": 34.8235294117647,
2357
+ "grad_norm": 1.9452924728393555,
2358
+ "learning_rate": 9.477340820420633e-06,
2359
+ "loss": 0.5852,
2360
+ "step": 592
2361
+ },
2362
+ {
2363
+ "epoch": 34.94117647058823,
2364
+ "grad_norm": 1.5196162462234497,
2365
+ "learning_rate": 9.471230788201429e-06,
2366
+ "loss": 0.5893,
2367
+ "step": 594
2368
+ },
2369
+ {
2370
+ "epoch": 35.0,
2371
+ "eval_loss": 0.9399586915969849,
2372
+ "eval_runtime": 14.5128,
2373
+ "eval_samples_per_second": 2.481,
2374
+ "eval_steps_per_second": 2.481,
2375
+ "step": 595
2376
+ },
2377
+ {
2378
+ "epoch": 35.0,
2379
+ "step": 595,
2380
+ "total_flos": 7.576813686279373e+16,
2381
+ "train_loss": 1.1527870081052058,
2382
+ "train_runtime": 5617.142,
2383
+ "train_samples_per_second": 3.632,
2384
+ "train_steps_per_second": 0.454
2385
+ }
2386
+ ],
2387
+ "logging_steps": 2,
2388
+ "max_steps": 2550,
2389
+ "num_input_tokens_seen": 0,
2390
+ "num_train_epochs": 150,
2391
+ "save_steps": 25,
2392
+ "stateful_callbacks": {
2393
+ "EarlyStoppingCallback": {
2394
+ "args": {
2395
+ "early_stopping_patience": 7,
2396
+ "early_stopping_threshold": 0.0
2397
+ },
2398
+ "attributes": {
2399
+ "early_stopping_patience_counter": 0
2400
+ }
2401
+ },
2402
+ "TrainerControl": {
2403
+ "args": {
2404
+ "should_epoch_stop": false,
2405
+ "should_evaluate": false,
2406
+ "should_log": false,
2407
+ "should_save": true,
2408
+ "should_training_stop": true
2409
+ },
2410
+ "attributes": {}
2411
+ }
2412
+ },
2413
+ "total_flos": 7.576813686279373e+16,
2414
+ "train_batch_size": 1,
2415
+ "trial_name": null,
2416
+ "trial_params": null
2417
+ }