End of training
Browse files- all_results.json +16 -0
- eval_results.json +11 -0
- runs/Dec09_03-30-42_8a523c90c79c/events.out.tfevents.1670561042.8a523c90c79c.76.3 +3 -0
- train_results.json +8 -0
- trainer_state.json +1387 -0
all_results.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"eval_loss": 0.20216532051563263,
|
4 |
+
"eval_rouge1": 0.0,
|
5 |
+
"eval_rouge2": 0.0,
|
6 |
+
"eval_rougeL": 0.0,
|
7 |
+
"eval_rougeLsum": 0.0,
|
8 |
+
"eval_runtime": 84.4466,
|
9 |
+
"eval_samples_per_second": 4.82,
|
10 |
+
"eval_steps_per_second": 2.416,
|
11 |
+
"total_flos": 3.2211655354615726e+18,
|
12 |
+
"train_loss": 0.2087684871917057,
|
13 |
+
"train_runtime": 2567.1485,
|
14 |
+
"train_samples_per_second": 6.916,
|
15 |
+
"train_steps_per_second": 0.864
|
16 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"eval_loss": 0.20216532051563263,
|
4 |
+
"eval_rouge1": 0.0,
|
5 |
+
"eval_rouge2": 0.0,
|
6 |
+
"eval_rougeL": 0.0,
|
7 |
+
"eval_rougeLsum": 0.0,
|
8 |
+
"eval_runtime": 84.4466,
|
9 |
+
"eval_samples_per_second": 4.82,
|
10 |
+
"eval_steps_per_second": 2.416
|
11 |
+
}
|
runs/Dec09_03-30-42_8a523c90c79c/events.out.tfevents.1670561042.8a523c90c79c.76.3
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f6ce4a6f65afa407525b3e6f503590942d2a4da51c3fccd7fb005345431a09b0
|
3 |
+
size 514
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"total_flos": 3.2211655354615726e+18,
|
4 |
+
"train_loss": 0.2087684871917057,
|
5 |
+
"train_runtime": 2567.1485,
|
6 |
+
"train_samples_per_second": 6.916,
|
7 |
+
"train_steps_per_second": 0.864
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,1387 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 2.9989861439675565,
|
5 |
+
"global_step": 2217,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.01,
|
12 |
+
"learning_rate": 3.310780333784393e-05,
|
13 |
+
"loss": 0.255,
|
14 |
+
"step": 10
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.03,
|
18 |
+
"learning_rate": 3.288227334235454e-05,
|
19 |
+
"loss": 0.2416,
|
20 |
+
"step": 20
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.04,
|
24 |
+
"learning_rate": 3.2656743346865135e-05,
|
25 |
+
"loss": 0.232,
|
26 |
+
"step": 30
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 0.05,
|
30 |
+
"learning_rate": 3.243121335137573e-05,
|
31 |
+
"loss": 0.2344,
|
32 |
+
"step": 40
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.07,
|
36 |
+
"learning_rate": 3.220568335588634e-05,
|
37 |
+
"loss": 0.2296,
|
38 |
+
"step": 50
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 0.08,
|
42 |
+
"learning_rate": 3.1980153360396934e-05,
|
43 |
+
"loss": 0.2471,
|
44 |
+
"step": 60
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.09,
|
48 |
+
"learning_rate": 3.175462336490753e-05,
|
49 |
+
"loss": 0.2212,
|
50 |
+
"step": 70
|
51 |
+
},
|
52 |
+
{
|
53 |
+
"epoch": 0.11,
|
54 |
+
"learning_rate": 3.1529093369418136e-05,
|
55 |
+
"loss": 0.2219,
|
56 |
+
"step": 80
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 0.12,
|
60 |
+
"learning_rate": 3.1303563373928734e-05,
|
61 |
+
"loss": 0.2305,
|
62 |
+
"step": 90
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 0.14,
|
66 |
+
"learning_rate": 3.107803337843934e-05,
|
67 |
+
"loss": 0.2555,
|
68 |
+
"step": 100
|
69 |
+
},
|
70 |
+
{
|
71 |
+
"epoch": 0.15,
|
72 |
+
"learning_rate": 3.085250338294993e-05,
|
73 |
+
"loss": 0.2398,
|
74 |
+
"step": 110
|
75 |
+
},
|
76 |
+
{
|
77 |
+
"epoch": 0.16,
|
78 |
+
"learning_rate": 3.0626973387460534e-05,
|
79 |
+
"loss": 0.2538,
|
80 |
+
"step": 120
|
81 |
+
},
|
82 |
+
{
|
83 |
+
"epoch": 0.18,
|
84 |
+
"learning_rate": 3.0401443391971135e-05,
|
85 |
+
"loss": 0.2271,
|
86 |
+
"step": 130
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.19,
|
90 |
+
"learning_rate": 3.0175913396481732e-05,
|
91 |
+
"loss": 0.2479,
|
92 |
+
"step": 140
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 0.2,
|
96 |
+
"learning_rate": 2.9950383400992333e-05,
|
97 |
+
"loss": 0.2465,
|
98 |
+
"step": 150
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 0.22,
|
102 |
+
"learning_rate": 2.9724853405502934e-05,
|
103 |
+
"loss": 0.2288,
|
104 |
+
"step": 160
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 0.23,
|
108 |
+
"learning_rate": 2.9499323410013535e-05,
|
109 |
+
"loss": 0.2247,
|
110 |
+
"step": 170
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 0.24,
|
114 |
+
"learning_rate": 2.927379341452413e-05,
|
115 |
+
"loss": 0.2442,
|
116 |
+
"step": 180
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 0.26,
|
120 |
+
"learning_rate": 2.9048263419034734e-05,
|
121 |
+
"loss": 0.2321,
|
122 |
+
"step": 190
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 0.27,
|
126 |
+
"learning_rate": 2.8822733423545335e-05,
|
127 |
+
"loss": 0.2342,
|
128 |
+
"step": 200
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.28,
|
132 |
+
"learning_rate": 2.8597203428055936e-05,
|
133 |
+
"loss": 0.2342,
|
134 |
+
"step": 210
|
135 |
+
},
|
136 |
+
{
|
137 |
+
"epoch": 0.3,
|
138 |
+
"learning_rate": 2.837167343256653e-05,
|
139 |
+
"loss": 0.2351,
|
140 |
+
"step": 220
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 0.31,
|
144 |
+
"learning_rate": 2.814614343707713e-05,
|
145 |
+
"loss": 0.2155,
|
146 |
+
"step": 230
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"epoch": 0.32,
|
150 |
+
"learning_rate": 2.7920613441587736e-05,
|
151 |
+
"loss": 0.2517,
|
152 |
+
"step": 240
|
153 |
+
},
|
154 |
+
{
|
155 |
+
"epoch": 0.34,
|
156 |
+
"learning_rate": 2.769508344609833e-05,
|
157 |
+
"loss": 0.2431,
|
158 |
+
"step": 250
|
159 |
+
},
|
160 |
+
{
|
161 |
+
"epoch": 0.35,
|
162 |
+
"learning_rate": 2.746955345060893e-05,
|
163 |
+
"loss": 0.2409,
|
164 |
+
"step": 260
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 0.36,
|
168 |
+
"learning_rate": 2.7244023455119532e-05,
|
169 |
+
"loss": 0.2458,
|
170 |
+
"step": 270
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.38,
|
174 |
+
"learning_rate": 2.7018493459630133e-05,
|
175 |
+
"loss": 0.2268,
|
176 |
+
"step": 280
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 0.39,
|
180 |
+
"learning_rate": 2.679296346414073e-05,
|
181 |
+
"loss": 0.2256,
|
182 |
+
"step": 290
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 0.41,
|
186 |
+
"learning_rate": 2.6567433468651332e-05,
|
187 |
+
"loss": 0.2408,
|
188 |
+
"step": 300
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 0.42,
|
192 |
+
"learning_rate": 2.6341903473161933e-05,
|
193 |
+
"loss": 0.224,
|
194 |
+
"step": 310
|
195 |
+
},
|
196 |
+
{
|
197 |
+
"epoch": 0.43,
|
198 |
+
"learning_rate": 2.6116373477672534e-05,
|
199 |
+
"loss": 0.2328,
|
200 |
+
"step": 320
|
201 |
+
},
|
202 |
+
{
|
203 |
+
"epoch": 0.45,
|
204 |
+
"learning_rate": 2.589084348218313e-05,
|
205 |
+
"loss": 0.2359,
|
206 |
+
"step": 330
|
207 |
+
},
|
208 |
+
{
|
209 |
+
"epoch": 0.46,
|
210 |
+
"learning_rate": 2.5665313486693732e-05,
|
211 |
+
"loss": 0.23,
|
212 |
+
"step": 340
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.47,
|
216 |
+
"learning_rate": 2.5439783491204333e-05,
|
217 |
+
"loss": 0.2371,
|
218 |
+
"step": 350
|
219 |
+
},
|
220 |
+
{
|
221 |
+
"epoch": 0.49,
|
222 |
+
"learning_rate": 2.5214253495714928e-05,
|
223 |
+
"loss": 0.2443,
|
224 |
+
"step": 360
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"epoch": 0.5,
|
228 |
+
"learning_rate": 2.4988723500225532e-05,
|
229 |
+
"loss": 0.2411,
|
230 |
+
"step": 370
|
231 |
+
},
|
232 |
+
{
|
233 |
+
"epoch": 0.51,
|
234 |
+
"learning_rate": 2.4763193504736133e-05,
|
235 |
+
"loss": 0.2294,
|
236 |
+
"step": 380
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 0.53,
|
240 |
+
"learning_rate": 2.453766350924673e-05,
|
241 |
+
"loss": 0.2108,
|
242 |
+
"step": 390
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 0.54,
|
246 |
+
"learning_rate": 2.4312133513757332e-05,
|
247 |
+
"loss": 0.2424,
|
248 |
+
"step": 400
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 0.55,
|
252 |
+
"learning_rate": 2.408660351826793e-05,
|
253 |
+
"loss": 0.2246,
|
254 |
+
"step": 410
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.57,
|
258 |
+
"learning_rate": 2.386107352277853e-05,
|
259 |
+
"loss": 0.2329,
|
260 |
+
"step": 420
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 0.58,
|
264 |
+
"learning_rate": 2.363554352728913e-05,
|
265 |
+
"loss": 0.2497,
|
266 |
+
"step": 430
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 0.59,
|
270 |
+
"learning_rate": 2.341001353179973e-05,
|
271 |
+
"loss": 0.2363,
|
272 |
+
"step": 440
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 0.61,
|
276 |
+
"learning_rate": 2.318448353631033e-05,
|
277 |
+
"loss": 0.2322,
|
278 |
+
"step": 450
|
279 |
+
},
|
280 |
+
{
|
281 |
+
"epoch": 0.62,
|
282 |
+
"learning_rate": 2.2958953540820928e-05,
|
283 |
+
"loss": 0.2233,
|
284 |
+
"step": 460
|
285 |
+
},
|
286 |
+
{
|
287 |
+
"epoch": 0.64,
|
288 |
+
"learning_rate": 2.2733423545331532e-05,
|
289 |
+
"loss": 0.2386,
|
290 |
+
"step": 470
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 0.65,
|
294 |
+
"learning_rate": 2.250789354984213e-05,
|
295 |
+
"loss": 0.2283,
|
296 |
+
"step": 480
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.66,
|
300 |
+
"learning_rate": 2.228236355435273e-05,
|
301 |
+
"loss": 0.2178,
|
302 |
+
"step": 490
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 0.68,
|
306 |
+
"learning_rate": 2.205683355886333e-05,
|
307 |
+
"loss": 0.2386,
|
308 |
+
"step": 500
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"epoch": 0.69,
|
312 |
+
"learning_rate": 2.183130356337393e-05,
|
313 |
+
"loss": 0.2261,
|
314 |
+
"step": 510
|
315 |
+
},
|
316 |
+
{
|
317 |
+
"epoch": 0.7,
|
318 |
+
"learning_rate": 2.160577356788453e-05,
|
319 |
+
"loss": 0.2236,
|
320 |
+
"step": 520
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 0.72,
|
324 |
+
"learning_rate": 2.1380243572395128e-05,
|
325 |
+
"loss": 0.2247,
|
326 |
+
"step": 530
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 0.73,
|
330 |
+
"learning_rate": 2.115471357690573e-05,
|
331 |
+
"loss": 0.2361,
|
332 |
+
"step": 540
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 0.74,
|
336 |
+
"learning_rate": 2.0929183581416327e-05,
|
337 |
+
"loss": 0.2198,
|
338 |
+
"step": 550
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 0.76,
|
342 |
+
"learning_rate": 2.070365358592693e-05,
|
343 |
+
"loss": 0.2425,
|
344 |
+
"step": 560
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 0.77,
|
348 |
+
"learning_rate": 2.047812359043753e-05,
|
349 |
+
"loss": 0.2273,
|
350 |
+
"step": 570
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 0.78,
|
354 |
+
"learning_rate": 2.025259359494813e-05,
|
355 |
+
"loss": 0.2501,
|
356 |
+
"step": 580
|
357 |
+
},
|
358 |
+
{
|
359 |
+
"epoch": 0.8,
|
360 |
+
"learning_rate": 2.0027063599458727e-05,
|
361 |
+
"loss": 0.2364,
|
362 |
+
"step": 590
|
363 |
+
},
|
364 |
+
{
|
365 |
+
"epoch": 0.81,
|
366 |
+
"learning_rate": 1.980153360396933e-05,
|
367 |
+
"loss": 0.2386,
|
368 |
+
"step": 600
|
369 |
+
},
|
370 |
+
{
|
371 |
+
"epoch": 0.82,
|
372 |
+
"learning_rate": 1.957600360847993e-05,
|
373 |
+
"loss": 0.234,
|
374 |
+
"step": 610
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 0.84,
|
378 |
+
"learning_rate": 1.9350473612990527e-05,
|
379 |
+
"loss": 0.2277,
|
380 |
+
"step": 620
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 0.85,
|
384 |
+
"learning_rate": 1.9124943617501128e-05,
|
385 |
+
"loss": 0.2352,
|
386 |
+
"step": 630
|
387 |
+
},
|
388 |
+
{
|
389 |
+
"epoch": 0.87,
|
390 |
+
"learning_rate": 1.8899413622011726e-05,
|
391 |
+
"loss": 0.2087,
|
392 |
+
"step": 640
|
393 |
+
},
|
394 |
+
{
|
395 |
+
"epoch": 0.88,
|
396 |
+
"learning_rate": 1.867388362652233e-05,
|
397 |
+
"loss": 0.226,
|
398 |
+
"step": 650
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 0.89,
|
402 |
+
"learning_rate": 1.8448353631032928e-05,
|
403 |
+
"loss": 0.2267,
|
404 |
+
"step": 660
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 0.91,
|
408 |
+
"learning_rate": 1.822282363554353e-05,
|
409 |
+
"loss": 0.2257,
|
410 |
+
"step": 670
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 0.92,
|
414 |
+
"learning_rate": 1.7997293640054127e-05,
|
415 |
+
"loss": 0.2315,
|
416 |
+
"step": 680
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 0.93,
|
420 |
+
"learning_rate": 1.7771763644564728e-05,
|
421 |
+
"loss": 0.2219,
|
422 |
+
"step": 690
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 0.95,
|
426 |
+
"learning_rate": 1.754623364907533e-05,
|
427 |
+
"loss": 0.2263,
|
428 |
+
"step": 700
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 0.96,
|
432 |
+
"learning_rate": 1.732070365358593e-05,
|
433 |
+
"loss": 0.2072,
|
434 |
+
"step": 710
|
435 |
+
},
|
436 |
+
{
|
437 |
+
"epoch": 0.97,
|
438 |
+
"learning_rate": 1.7095173658096527e-05,
|
439 |
+
"loss": 0.2179,
|
440 |
+
"step": 720
|
441 |
+
},
|
442 |
+
{
|
443 |
+
"epoch": 0.99,
|
444 |
+
"learning_rate": 1.6869643662607125e-05,
|
445 |
+
"loss": 0.2359,
|
446 |
+
"step": 730
|
447 |
+
},
|
448 |
+
{
|
449 |
+
"epoch": 1.0,
|
450 |
+
"eval_loss": 0.19105975329875946,
|
451 |
+
"eval_rouge1": 0.0,
|
452 |
+
"eval_rouge2": 0.0,
|
453 |
+
"eval_rougeL": 0.0,
|
454 |
+
"eval_rougeLsum": 0.0,
|
455 |
+
"eval_runtime": 72.427,
|
456 |
+
"eval_samples_per_second": 5.619,
|
457 |
+
"eval_steps_per_second": 2.817,
|
458 |
+
"step": 739
|
459 |
+
},
|
460 |
+
{
|
461 |
+
"epoch": 1.0,
|
462 |
+
"learning_rate": 1.664411366711773e-05,
|
463 |
+
"loss": 0.2357,
|
464 |
+
"step": 740
|
465 |
+
},
|
466 |
+
{
|
467 |
+
"epoch": 1.01,
|
468 |
+
"learning_rate": 1.6418583671628327e-05,
|
469 |
+
"loss": 0.2042,
|
470 |
+
"step": 750
|
471 |
+
},
|
472 |
+
{
|
473 |
+
"epoch": 1.03,
|
474 |
+
"learning_rate": 1.6193053676138928e-05,
|
475 |
+
"loss": 0.2101,
|
476 |
+
"step": 760
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 1.04,
|
480 |
+
"learning_rate": 1.5967523680649526e-05,
|
481 |
+
"loss": 0.1908,
|
482 |
+
"step": 770
|
483 |
+
},
|
484 |
+
{
|
485 |
+
"epoch": 1.06,
|
486 |
+
"learning_rate": 1.5741993685160127e-05,
|
487 |
+
"loss": 0.2128,
|
488 |
+
"step": 780
|
489 |
+
},
|
490 |
+
{
|
491 |
+
"epoch": 1.07,
|
492 |
+
"learning_rate": 1.5516463689670728e-05,
|
493 |
+
"loss": 0.1934,
|
494 |
+
"step": 790
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 1.08,
|
498 |
+
"learning_rate": 1.529093369418133e-05,
|
499 |
+
"loss": 0.2087,
|
500 |
+
"step": 800
|
501 |
+
},
|
502 |
+
{
|
503 |
+
"epoch": 1.1,
|
504 |
+
"learning_rate": 1.5065403698691926e-05,
|
505 |
+
"loss": 0.194,
|
506 |
+
"step": 810
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 1.11,
|
510 |
+
"learning_rate": 1.4839873703202526e-05,
|
511 |
+
"loss": 0.1993,
|
512 |
+
"step": 820
|
513 |
+
},
|
514 |
+
{
|
515 |
+
"epoch": 1.12,
|
516 |
+
"learning_rate": 1.4614343707713127e-05,
|
517 |
+
"loss": 0.1979,
|
518 |
+
"step": 830
|
519 |
+
},
|
520 |
+
{
|
521 |
+
"epoch": 1.14,
|
522 |
+
"learning_rate": 1.4388813712223726e-05,
|
523 |
+
"loss": 0.2281,
|
524 |
+
"step": 840
|
525 |
+
},
|
526 |
+
{
|
527 |
+
"epoch": 1.15,
|
528 |
+
"learning_rate": 1.4163283716734327e-05,
|
529 |
+
"loss": 0.205,
|
530 |
+
"step": 850
|
531 |
+
},
|
532 |
+
{
|
533 |
+
"epoch": 1.16,
|
534 |
+
"learning_rate": 1.3937753721244926e-05,
|
535 |
+
"loss": 0.1895,
|
536 |
+
"step": 860
|
537 |
+
},
|
538 |
+
{
|
539 |
+
"epoch": 1.18,
|
540 |
+
"learning_rate": 1.3712223725755527e-05,
|
541 |
+
"loss": 0.1999,
|
542 |
+
"step": 870
|
543 |
+
},
|
544 |
+
{
|
545 |
+
"epoch": 1.19,
|
546 |
+
"learning_rate": 1.3486693730266125e-05,
|
547 |
+
"loss": 0.204,
|
548 |
+
"step": 880
|
549 |
+
},
|
550 |
+
{
|
551 |
+
"epoch": 1.2,
|
552 |
+
"learning_rate": 1.3261163734776728e-05,
|
553 |
+
"loss": 0.2026,
|
554 |
+
"step": 890
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 1.22,
|
558 |
+
"learning_rate": 1.3035633739287325e-05,
|
559 |
+
"loss": 0.2291,
|
560 |
+
"step": 900
|
561 |
+
},
|
562 |
+
{
|
563 |
+
"epoch": 1.23,
|
564 |
+
"learning_rate": 1.2810103743797926e-05,
|
565 |
+
"loss": 0.2036,
|
566 |
+
"step": 910
|
567 |
+
},
|
568 |
+
{
|
569 |
+
"epoch": 1.24,
|
570 |
+
"learning_rate": 1.2584573748308526e-05,
|
571 |
+
"loss": 0.2065,
|
572 |
+
"step": 920
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 1.26,
|
576 |
+
"learning_rate": 1.2359043752819125e-05,
|
577 |
+
"loss": 0.2039,
|
578 |
+
"step": 930
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 1.27,
|
582 |
+
"learning_rate": 1.2133513757329726e-05,
|
583 |
+
"loss": 0.208,
|
584 |
+
"step": 940
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 1.29,
|
588 |
+
"learning_rate": 1.1907983761840325e-05,
|
589 |
+
"loss": 0.2177,
|
590 |
+
"step": 950
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 1.3,
|
594 |
+
"learning_rate": 1.1682453766350926e-05,
|
595 |
+
"loss": 0.1943,
|
596 |
+
"step": 960
|
597 |
+
},
|
598 |
+
{
|
599 |
+
"epoch": 1.31,
|
600 |
+
"learning_rate": 1.1456923770861524e-05,
|
601 |
+
"loss": 0.207,
|
602 |
+
"step": 970
|
603 |
+
},
|
604 |
+
{
|
605 |
+
"epoch": 1.33,
|
606 |
+
"learning_rate": 1.1231393775372125e-05,
|
607 |
+
"loss": 0.2017,
|
608 |
+
"step": 980
|
609 |
+
},
|
610 |
+
{
|
611 |
+
"epoch": 1.34,
|
612 |
+
"learning_rate": 1.1005863779882724e-05,
|
613 |
+
"loss": 0.2115,
|
614 |
+
"step": 990
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 1.35,
|
618 |
+
"learning_rate": 1.0780333784393325e-05,
|
619 |
+
"loss": 0.2098,
|
620 |
+
"step": 1000
|
621 |
+
},
|
622 |
+
{
|
623 |
+
"epoch": 1.37,
|
624 |
+
"learning_rate": 1.0554803788903925e-05,
|
625 |
+
"loss": 0.2105,
|
626 |
+
"step": 1010
|
627 |
+
},
|
628 |
+
{
|
629 |
+
"epoch": 1.38,
|
630 |
+
"learning_rate": 1.0329273793414524e-05,
|
631 |
+
"loss": 0.2149,
|
632 |
+
"step": 1020
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 1.39,
|
636 |
+
"learning_rate": 1.0103743797925125e-05,
|
637 |
+
"loss": 0.2027,
|
638 |
+
"step": 1030
|
639 |
+
},
|
640 |
+
{
|
641 |
+
"epoch": 1.41,
|
642 |
+
"learning_rate": 9.878213802435724e-06,
|
643 |
+
"loss": 0.1887,
|
644 |
+
"step": 1040
|
645 |
+
},
|
646 |
+
{
|
647 |
+
"epoch": 1.42,
|
648 |
+
"learning_rate": 9.652683806946325e-06,
|
649 |
+
"loss": 0.1955,
|
650 |
+
"step": 1050
|
651 |
+
},
|
652 |
+
{
|
653 |
+
"epoch": 1.43,
|
654 |
+
"learning_rate": 9.427153811456923e-06,
|
655 |
+
"loss": 0.2065,
|
656 |
+
"step": 1060
|
657 |
+
},
|
658 |
+
{
|
659 |
+
"epoch": 1.45,
|
660 |
+
"learning_rate": 9.201623815967524e-06,
|
661 |
+
"loss": 0.2088,
|
662 |
+
"step": 1070
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 1.46,
|
666 |
+
"learning_rate": 8.976093820478123e-06,
|
667 |
+
"loss": 0.2119,
|
668 |
+
"step": 1080
|
669 |
+
},
|
670 |
+
{
|
671 |
+
"epoch": 1.47,
|
672 |
+
"learning_rate": 8.750563824988724e-06,
|
673 |
+
"loss": 0.2101,
|
674 |
+
"step": 1090
|
675 |
+
},
|
676 |
+
{
|
677 |
+
"epoch": 1.49,
|
678 |
+
"learning_rate": 8.525033829499324e-06,
|
679 |
+
"loss": 0.1985,
|
680 |
+
"step": 1100
|
681 |
+
},
|
682 |
+
{
|
683 |
+
"epoch": 1.5,
|
684 |
+
"learning_rate": 8.299503834009923e-06,
|
685 |
+
"loss": 0.2105,
|
686 |
+
"step": 1110
|
687 |
+
},
|
688 |
+
{
|
689 |
+
"epoch": 1.52,
|
690 |
+
"learning_rate": 8.073973838520524e-06,
|
691 |
+
"loss": 0.2121,
|
692 |
+
"step": 1120
|
693 |
+
},
|
694 |
+
{
|
695 |
+
"epoch": 1.53,
|
696 |
+
"learning_rate": 7.848443843031123e-06,
|
697 |
+
"loss": 0.1942,
|
698 |
+
"step": 1130
|
699 |
+
},
|
700 |
+
{
|
701 |
+
"epoch": 1.54,
|
702 |
+
"learning_rate": 7.6229138475417236e-06,
|
703 |
+
"loss": 0.1874,
|
704 |
+
"step": 1140
|
705 |
+
},
|
706 |
+
{
|
707 |
+
"epoch": 1.56,
|
708 |
+
"learning_rate": 7.397383852052324e-06,
|
709 |
+
"loss": 0.204,
|
710 |
+
"step": 1150
|
711 |
+
},
|
712 |
+
{
|
713 |
+
"epoch": 1.57,
|
714 |
+
"learning_rate": 7.171853856562922e-06,
|
715 |
+
"loss": 0.1883,
|
716 |
+
"step": 1160
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 1.58,
|
720 |
+
"learning_rate": 6.946323861073522e-06,
|
721 |
+
"loss": 0.2032,
|
722 |
+
"step": 1170
|
723 |
+
},
|
724 |
+
{
|
725 |
+
"epoch": 1.6,
|
726 |
+
"learning_rate": 6.7207938655841226e-06,
|
727 |
+
"loss": 0.1878,
|
728 |
+
"step": 1180
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 1.61,
|
732 |
+
"learning_rate": 6.495263870094723e-06,
|
733 |
+
"loss": 0.2057,
|
734 |
+
"step": 1190
|
735 |
+
},
|
736 |
+
{
|
737 |
+
"epoch": 1.62,
|
738 |
+
"learning_rate": 6.269733874605323e-06,
|
739 |
+
"loss": 0.2096,
|
740 |
+
"step": 1200
|
741 |
+
},
|
742 |
+
{
|
743 |
+
"epoch": 1.64,
|
744 |
+
"learning_rate": 6.044203879115923e-06,
|
745 |
+
"loss": 0.1872,
|
746 |
+
"step": 1210
|
747 |
+
},
|
748 |
+
{
|
749 |
+
"epoch": 1.65,
|
750 |
+
"learning_rate": 5.8186738836265224e-06,
|
751 |
+
"loss": 0.2072,
|
752 |
+
"step": 1220
|
753 |
+
},
|
754 |
+
{
|
755 |
+
"epoch": 1.66,
|
756 |
+
"learning_rate": 5.593143888137122e-06,
|
757 |
+
"loss": 0.2098,
|
758 |
+
"step": 1230
|
759 |
+
},
|
760 |
+
{
|
761 |
+
"epoch": 1.68,
|
762 |
+
"learning_rate": 5.367613892647722e-06,
|
763 |
+
"loss": 0.2255,
|
764 |
+
"step": 1240
|
765 |
+
},
|
766 |
+
{
|
767 |
+
"epoch": 1.69,
|
768 |
+
"learning_rate": 5.142083897158322e-06,
|
769 |
+
"loss": 0.1859,
|
770 |
+
"step": 1250
|
771 |
+
},
|
772 |
+
{
|
773 |
+
"epoch": 1.7,
|
774 |
+
"learning_rate": 4.916553901668922e-06,
|
775 |
+
"loss": 0.1849,
|
776 |
+
"step": 1260
|
777 |
+
},
|
778 |
+
{
|
779 |
+
"epoch": 1.72,
|
780 |
+
"learning_rate": 4.6910239061795225e-06,
|
781 |
+
"loss": 0.2022,
|
782 |
+
"step": 1270
|
783 |
+
},
|
784 |
+
{
|
785 |
+
"epoch": 1.73,
|
786 |
+
"learning_rate": 4.465493910690122e-06,
|
787 |
+
"loss": 0.1976,
|
788 |
+
"step": 1280
|
789 |
+
},
|
790 |
+
{
|
791 |
+
"epoch": 1.74,
|
792 |
+
"learning_rate": 4.239963915200722e-06,
|
793 |
+
"loss": 0.1988,
|
794 |
+
"step": 1290
|
795 |
+
},
|
796 |
+
{
|
797 |
+
"epoch": 1.76,
|
798 |
+
"learning_rate": 4.014433919711322e-06,
|
799 |
+
"loss": 0.1983,
|
800 |
+
"step": 1300
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 1.77,
|
804 |
+
"learning_rate": 3.788903924221922e-06,
|
805 |
+
"loss": 0.198,
|
806 |
+
"step": 1310
|
807 |
+
},
|
808 |
+
{
|
809 |
+
"epoch": 1.79,
|
810 |
+
"learning_rate": 3.563373928732522e-06,
|
811 |
+
"loss": 0.1967,
|
812 |
+
"step": 1320
|
813 |
+
},
|
814 |
+
{
|
815 |
+
"epoch": 1.8,
|
816 |
+
"learning_rate": 3.3378439332431214e-06,
|
817 |
+
"loss": 0.202,
|
818 |
+
"step": 1330
|
819 |
+
},
|
820 |
+
{
|
821 |
+
"epoch": 1.81,
|
822 |
+
"learning_rate": 3.112313937753721e-06,
|
823 |
+
"loss": 0.1974,
|
824 |
+
"step": 1340
|
825 |
+
},
|
826 |
+
{
|
827 |
+
"epoch": 1.83,
|
828 |
+
"learning_rate": 2.8867839422643213e-06,
|
829 |
+
"loss": 0.2045,
|
830 |
+
"step": 1350
|
831 |
+
},
|
832 |
+
{
|
833 |
+
"epoch": 1.84,
|
834 |
+
"learning_rate": 2.6612539467749215e-06,
|
835 |
+
"loss": 0.1997,
|
836 |
+
"step": 1360
|
837 |
+
},
|
838 |
+
{
|
839 |
+
"epoch": 1.85,
|
840 |
+
"learning_rate": 2.435723951285521e-06,
|
841 |
+
"loss": 0.1879,
|
842 |
+
"step": 1370
|
843 |
+
},
|
844 |
+
{
|
845 |
+
"epoch": 1.87,
|
846 |
+
"learning_rate": 2.210193955796121e-06,
|
847 |
+
"loss": 0.1983,
|
848 |
+
"step": 1380
|
849 |
+
},
|
850 |
+
{
|
851 |
+
"epoch": 1.88,
|
852 |
+
"learning_rate": 1.9846639603067207e-06,
|
853 |
+
"loss": 0.1943,
|
854 |
+
"step": 1390
|
855 |
+
},
|
856 |
+
{
|
857 |
+
"epoch": 1.89,
|
858 |
+
"learning_rate": 1.7591339648173207e-06,
|
859 |
+
"loss": 0.2045,
|
860 |
+
"step": 1400
|
861 |
+
},
|
862 |
+
{
|
863 |
+
"epoch": 1.91,
|
864 |
+
"learning_rate": 1.5336039693279207e-06,
|
865 |
+
"loss": 0.2134,
|
866 |
+
"step": 1410
|
867 |
+
},
|
868 |
+
{
|
869 |
+
"epoch": 1.92,
|
870 |
+
"learning_rate": 1.3080739738385204e-06,
|
871 |
+
"loss": 0.2181,
|
872 |
+
"step": 1420
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 1.93,
|
876 |
+
"learning_rate": 1.0825439783491206e-06,
|
877 |
+
"loss": 0.1993,
|
878 |
+
"step": 1430
|
879 |
+
},
|
880 |
+
{
|
881 |
+
"epoch": 1.95,
|
882 |
+
"learning_rate": 8.570139828597205e-07,
|
883 |
+
"loss": 0.1982,
|
884 |
+
"step": 1440
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 1.96,
|
888 |
+
"learning_rate": 6.314839873703203e-07,
|
889 |
+
"loss": 0.1947,
|
890 |
+
"step": 1450
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 1.97,
|
894 |
+
"learning_rate": 4.0595399188092017e-07,
|
895 |
+
"loss": 0.199,
|
896 |
+
"step": 1460
|
897 |
+
},
|
898 |
+
{
|
899 |
+
"epoch": 1.99,
|
900 |
+
"learning_rate": 1.8042399639152008e-07,
|
901 |
+
"loss": 0.2169,
|
902 |
+
"step": 1470
|
903 |
+
},
|
904 |
+
{
|
905 |
+
"epoch": 2.0,
|
906 |
+
"eval_loss": 0.20216532051563263,
|
907 |
+
"eval_rouge1": 0.0,
|
908 |
+
"eval_rouge2": 0.0,
|
909 |
+
"eval_rougeL": 0.0,
|
910 |
+
"eval_rougeLsum": 0.0,
|
911 |
+
"eval_runtime": 80.504,
|
912 |
+
"eval_samples_per_second": 5.056,
|
913 |
+
"eval_steps_per_second": 2.534,
|
914 |
+
"step": 1478
|
915 |
+
},
|
916 |
+
{
|
917 |
+
"epoch": 2.0,
|
918 |
+
"learning_rate": 0.0,
|
919 |
+
"loss": 0.2106,
|
920 |
+
"step": 1480
|
921 |
+
},
|
922 |
+
{
|
923 |
+
"epoch": 2.02,
|
924 |
+
"learning_rate": 0.0,
|
925 |
+
"loss": 0.2036,
|
926 |
+
"step": 1490
|
927 |
+
},
|
928 |
+
{
|
929 |
+
"epoch": 2.03,
|
930 |
+
"learning_rate": 0.0,
|
931 |
+
"loss": 0.1853,
|
932 |
+
"step": 1500
|
933 |
+
},
|
934 |
+
{
|
935 |
+
"epoch": 2.04,
|
936 |
+
"learning_rate": 0.0,
|
937 |
+
"loss": 0.1873,
|
938 |
+
"step": 1510
|
939 |
+
},
|
940 |
+
{
|
941 |
+
"epoch": 2.06,
|
942 |
+
"learning_rate": 0.0,
|
943 |
+
"loss": 0.1814,
|
944 |
+
"step": 1520
|
945 |
+
},
|
946 |
+
{
|
947 |
+
"epoch": 2.07,
|
948 |
+
"learning_rate": 0.0,
|
949 |
+
"loss": 0.1864,
|
950 |
+
"step": 1530
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 2.08,
|
954 |
+
"learning_rate": 0.0,
|
955 |
+
"loss": 0.2106,
|
956 |
+
"step": 1540
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 2.1,
|
960 |
+
"learning_rate": 0.0,
|
961 |
+
"loss": 0.1789,
|
962 |
+
"step": 1550
|
963 |
+
},
|
964 |
+
{
|
965 |
+
"epoch": 2.11,
|
966 |
+
"learning_rate": 0.0,
|
967 |
+
"loss": 0.1834,
|
968 |
+
"step": 1560
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 2.12,
|
972 |
+
"learning_rate": 0.0,
|
973 |
+
"loss": 0.2044,
|
974 |
+
"step": 1570
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 2.14,
|
978 |
+
"learning_rate": 0.0,
|
979 |
+
"loss": 0.1963,
|
980 |
+
"step": 1580
|
981 |
+
},
|
982 |
+
{
|
983 |
+
"epoch": 2.15,
|
984 |
+
"learning_rate": 0.0,
|
985 |
+
"loss": 0.1981,
|
986 |
+
"step": 1590
|
987 |
+
},
|
988 |
+
{
|
989 |
+
"epoch": 2.16,
|
990 |
+
"learning_rate": 0.0,
|
991 |
+
"loss": 0.1989,
|
992 |
+
"step": 1600
|
993 |
+
},
|
994 |
+
{
|
995 |
+
"epoch": 2.18,
|
996 |
+
"learning_rate": 0.0,
|
997 |
+
"loss": 0.1933,
|
998 |
+
"step": 1610
|
999 |
+
},
|
1000 |
+
{
|
1001 |
+
"epoch": 2.19,
|
1002 |
+
"learning_rate": 0.0,
|
1003 |
+
"loss": 0.172,
|
1004 |
+
"step": 1620
|
1005 |
+
},
|
1006 |
+
{
|
1007 |
+
"epoch": 2.21,
|
1008 |
+
"learning_rate": 0.0,
|
1009 |
+
"loss": 0.1742,
|
1010 |
+
"step": 1630
|
1011 |
+
},
|
1012 |
+
{
|
1013 |
+
"epoch": 2.22,
|
1014 |
+
"learning_rate": 0.0,
|
1015 |
+
"loss": 0.2018,
|
1016 |
+
"step": 1640
|
1017 |
+
},
|
1018 |
+
{
|
1019 |
+
"epoch": 2.23,
|
1020 |
+
"learning_rate": 0.0,
|
1021 |
+
"loss": 0.198,
|
1022 |
+
"step": 1650
|
1023 |
+
},
|
1024 |
+
{
|
1025 |
+
"epoch": 2.25,
|
1026 |
+
"learning_rate": 0.0,
|
1027 |
+
"loss": 0.2013,
|
1028 |
+
"step": 1660
|
1029 |
+
},
|
1030 |
+
{
|
1031 |
+
"epoch": 2.26,
|
1032 |
+
"learning_rate": 0.0,
|
1033 |
+
"loss": 0.1997,
|
1034 |
+
"step": 1670
|
1035 |
+
},
|
1036 |
+
{
|
1037 |
+
"epoch": 2.27,
|
1038 |
+
"learning_rate": 0.0,
|
1039 |
+
"loss": 0.1926,
|
1040 |
+
"step": 1680
|
1041 |
+
},
|
1042 |
+
{
|
1043 |
+
"epoch": 2.29,
|
1044 |
+
"learning_rate": 0.0,
|
1045 |
+
"loss": 0.1905,
|
1046 |
+
"step": 1690
|
1047 |
+
},
|
1048 |
+
{
|
1049 |
+
"epoch": 2.3,
|
1050 |
+
"learning_rate": 0.0,
|
1051 |
+
"loss": 0.1649,
|
1052 |
+
"step": 1700
|
1053 |
+
},
|
1054 |
+
{
|
1055 |
+
"epoch": 2.31,
|
1056 |
+
"learning_rate": 0.0,
|
1057 |
+
"loss": 0.188,
|
1058 |
+
"step": 1710
|
1059 |
+
},
|
1060 |
+
{
|
1061 |
+
"epoch": 2.33,
|
1062 |
+
"learning_rate": 0.0,
|
1063 |
+
"loss": 0.1974,
|
1064 |
+
"step": 1720
|
1065 |
+
},
|
1066 |
+
{
|
1067 |
+
"epoch": 2.34,
|
1068 |
+
"learning_rate": 0.0,
|
1069 |
+
"loss": 0.179,
|
1070 |
+
"step": 1730
|
1071 |
+
},
|
1072 |
+
{
|
1073 |
+
"epoch": 2.35,
|
1074 |
+
"learning_rate": 0.0,
|
1075 |
+
"loss": 0.1846,
|
1076 |
+
"step": 1740
|
1077 |
+
},
|
1078 |
+
{
|
1079 |
+
"epoch": 2.37,
|
1080 |
+
"learning_rate": 0.0,
|
1081 |
+
"loss": 0.1856,
|
1082 |
+
"step": 1750
|
1083 |
+
},
|
1084 |
+
{
|
1085 |
+
"epoch": 2.38,
|
1086 |
+
"learning_rate": 0.0,
|
1087 |
+
"loss": 0.1869,
|
1088 |
+
"step": 1760
|
1089 |
+
},
|
1090 |
+
{
|
1091 |
+
"epoch": 2.39,
|
1092 |
+
"learning_rate": 0.0,
|
1093 |
+
"loss": 0.1838,
|
1094 |
+
"step": 1770
|
1095 |
+
},
|
1096 |
+
{
|
1097 |
+
"epoch": 2.41,
|
1098 |
+
"learning_rate": 0.0,
|
1099 |
+
"loss": 0.1786,
|
1100 |
+
"step": 1780
|
1101 |
+
},
|
1102 |
+
{
|
1103 |
+
"epoch": 2.42,
|
1104 |
+
"learning_rate": 0.0,
|
1105 |
+
"loss": 0.1848,
|
1106 |
+
"step": 1790
|
1107 |
+
},
|
1108 |
+
{
|
1109 |
+
"epoch": 2.44,
|
1110 |
+
"learning_rate": 0.0,
|
1111 |
+
"loss": 0.2055,
|
1112 |
+
"step": 1800
|
1113 |
+
},
|
1114 |
+
{
|
1115 |
+
"epoch": 2.45,
|
1116 |
+
"learning_rate": 0.0,
|
1117 |
+
"loss": 0.1924,
|
1118 |
+
"step": 1810
|
1119 |
+
},
|
1120 |
+
{
|
1121 |
+
"epoch": 2.46,
|
1122 |
+
"learning_rate": 0.0,
|
1123 |
+
"loss": 0.1935,
|
1124 |
+
"step": 1820
|
1125 |
+
},
|
1126 |
+
{
|
1127 |
+
"epoch": 2.48,
|
1128 |
+
"learning_rate": 0.0,
|
1129 |
+
"loss": 0.1819,
|
1130 |
+
"step": 1830
|
1131 |
+
},
|
1132 |
+
{
|
1133 |
+
"epoch": 2.49,
|
1134 |
+
"learning_rate": 0.0,
|
1135 |
+
"loss": 0.1853,
|
1136 |
+
"step": 1840
|
1137 |
+
},
|
1138 |
+
{
|
1139 |
+
"epoch": 2.5,
|
1140 |
+
"learning_rate": 0.0,
|
1141 |
+
"loss": 0.194,
|
1142 |
+
"step": 1850
|
1143 |
+
},
|
1144 |
+
{
|
1145 |
+
"epoch": 2.52,
|
1146 |
+
"learning_rate": 0.0,
|
1147 |
+
"loss": 0.1757,
|
1148 |
+
"step": 1860
|
1149 |
+
},
|
1150 |
+
{
|
1151 |
+
"epoch": 2.53,
|
1152 |
+
"learning_rate": 0.0,
|
1153 |
+
"loss": 0.2099,
|
1154 |
+
"step": 1870
|
1155 |
+
},
|
1156 |
+
{
|
1157 |
+
"epoch": 2.54,
|
1158 |
+
"learning_rate": 0.0,
|
1159 |
+
"loss": 0.177,
|
1160 |
+
"step": 1880
|
1161 |
+
},
|
1162 |
+
{
|
1163 |
+
"epoch": 2.56,
|
1164 |
+
"learning_rate": 0.0,
|
1165 |
+
"loss": 0.1951,
|
1166 |
+
"step": 1890
|
1167 |
+
},
|
1168 |
+
{
|
1169 |
+
"epoch": 2.57,
|
1170 |
+
"learning_rate": 0.0,
|
1171 |
+
"loss": 0.2008,
|
1172 |
+
"step": 1900
|
1173 |
+
},
|
1174 |
+
{
|
1175 |
+
"epoch": 2.58,
|
1176 |
+
"learning_rate": 0.0,
|
1177 |
+
"loss": 0.1756,
|
1178 |
+
"step": 1910
|
1179 |
+
},
|
1180 |
+
{
|
1181 |
+
"epoch": 2.6,
|
1182 |
+
"learning_rate": 0.0,
|
1183 |
+
"loss": 0.201,
|
1184 |
+
"step": 1920
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 2.61,
|
1188 |
+
"learning_rate": 0.0,
|
1189 |
+
"loss": 0.1919,
|
1190 |
+
"step": 1930
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 2.62,
|
1194 |
+
"learning_rate": 0.0,
|
1195 |
+
"loss": 0.1882,
|
1196 |
+
"step": 1940
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 2.64,
|
1200 |
+
"learning_rate": 0.0,
|
1201 |
+
"loss": 0.202,
|
1202 |
+
"step": 1950
|
1203 |
+
},
|
1204 |
+
{
|
1205 |
+
"epoch": 2.65,
|
1206 |
+
"learning_rate": 0.0,
|
1207 |
+
"loss": 0.191,
|
1208 |
+
"step": 1960
|
1209 |
+
},
|
1210 |
+
{
|
1211 |
+
"epoch": 2.67,
|
1212 |
+
"learning_rate": 0.0,
|
1213 |
+
"loss": 0.1776,
|
1214 |
+
"step": 1970
|
1215 |
+
},
|
1216 |
+
{
|
1217 |
+
"epoch": 2.68,
|
1218 |
+
"learning_rate": 0.0,
|
1219 |
+
"loss": 0.1813,
|
1220 |
+
"step": 1980
|
1221 |
+
},
|
1222 |
+
{
|
1223 |
+
"epoch": 2.69,
|
1224 |
+
"learning_rate": 0.0,
|
1225 |
+
"loss": 0.1911,
|
1226 |
+
"step": 1990
|
1227 |
+
},
|
1228 |
+
{
|
1229 |
+
"epoch": 2.71,
|
1230 |
+
"learning_rate": 0.0,
|
1231 |
+
"loss": 0.1921,
|
1232 |
+
"step": 2000
|
1233 |
+
},
|
1234 |
+
{
|
1235 |
+
"epoch": 2.72,
|
1236 |
+
"learning_rate": 0.0,
|
1237 |
+
"loss": 0.1927,
|
1238 |
+
"step": 2010
|
1239 |
+
},
|
1240 |
+
{
|
1241 |
+
"epoch": 2.73,
|
1242 |
+
"learning_rate": 0.0,
|
1243 |
+
"loss": 0.202,
|
1244 |
+
"step": 2020
|
1245 |
+
},
|
1246 |
+
{
|
1247 |
+
"epoch": 2.75,
|
1248 |
+
"learning_rate": 0.0,
|
1249 |
+
"loss": 0.1814,
|
1250 |
+
"step": 2030
|
1251 |
+
},
|
1252 |
+
{
|
1253 |
+
"epoch": 2.76,
|
1254 |
+
"learning_rate": 0.0,
|
1255 |
+
"loss": 0.1868,
|
1256 |
+
"step": 2040
|
1257 |
+
},
|
1258 |
+
{
|
1259 |
+
"epoch": 2.77,
|
1260 |
+
"learning_rate": 0.0,
|
1261 |
+
"loss": 0.1784,
|
1262 |
+
"step": 2050
|
1263 |
+
},
|
1264 |
+
{
|
1265 |
+
"epoch": 2.79,
|
1266 |
+
"learning_rate": 0.0,
|
1267 |
+
"loss": 0.1932,
|
1268 |
+
"step": 2060
|
1269 |
+
},
|
1270 |
+
{
|
1271 |
+
"epoch": 2.8,
|
1272 |
+
"learning_rate": 0.0,
|
1273 |
+
"loss": 0.1857,
|
1274 |
+
"step": 2070
|
1275 |
+
},
|
1276 |
+
{
|
1277 |
+
"epoch": 2.81,
|
1278 |
+
"learning_rate": 0.0,
|
1279 |
+
"loss": 0.2104,
|
1280 |
+
"step": 2080
|
1281 |
+
},
|
1282 |
+
{
|
1283 |
+
"epoch": 2.83,
|
1284 |
+
"learning_rate": 0.0,
|
1285 |
+
"loss": 0.1869,
|
1286 |
+
"step": 2090
|
1287 |
+
},
|
1288 |
+
{
|
1289 |
+
"epoch": 2.84,
|
1290 |
+
"learning_rate": 0.0,
|
1291 |
+
"loss": 0.1905,
|
1292 |
+
"step": 2100
|
1293 |
+
},
|
1294 |
+
{
|
1295 |
+
"epoch": 2.85,
|
1296 |
+
"learning_rate": 0.0,
|
1297 |
+
"loss": 0.1778,
|
1298 |
+
"step": 2110
|
1299 |
+
},
|
1300 |
+
{
|
1301 |
+
"epoch": 2.87,
|
1302 |
+
"learning_rate": 0.0,
|
1303 |
+
"loss": 0.1961,
|
1304 |
+
"step": 2120
|
1305 |
+
},
|
1306 |
+
{
|
1307 |
+
"epoch": 2.88,
|
1308 |
+
"learning_rate": 0.0,
|
1309 |
+
"loss": 0.1979,
|
1310 |
+
"step": 2130
|
1311 |
+
},
|
1312 |
+
{
|
1313 |
+
"epoch": 2.89,
|
1314 |
+
"learning_rate": 0.0,
|
1315 |
+
"loss": 0.2026,
|
1316 |
+
"step": 2140
|
1317 |
+
},
|
1318 |
+
{
|
1319 |
+
"epoch": 2.91,
|
1320 |
+
"learning_rate": 0.0,
|
1321 |
+
"loss": 0.1929,
|
1322 |
+
"step": 2150
|
1323 |
+
},
|
1324 |
+
{
|
1325 |
+
"epoch": 2.92,
|
1326 |
+
"learning_rate": 0.0,
|
1327 |
+
"loss": 0.1891,
|
1328 |
+
"step": 2160
|
1329 |
+
},
|
1330 |
+
{
|
1331 |
+
"epoch": 2.94,
|
1332 |
+
"learning_rate": 0.0,
|
1333 |
+
"loss": 0.1844,
|
1334 |
+
"step": 2170
|
1335 |
+
},
|
1336 |
+
{
|
1337 |
+
"epoch": 2.95,
|
1338 |
+
"learning_rate": 0.0,
|
1339 |
+
"loss": 0.1908,
|
1340 |
+
"step": 2180
|
1341 |
+
},
|
1342 |
+
{
|
1343 |
+
"epoch": 2.96,
|
1344 |
+
"learning_rate": 0.0,
|
1345 |
+
"loss": 0.2041,
|
1346 |
+
"step": 2190
|
1347 |
+
},
|
1348 |
+
{
|
1349 |
+
"epoch": 2.98,
|
1350 |
+
"learning_rate": 0.0,
|
1351 |
+
"loss": 0.2078,
|
1352 |
+
"step": 2200
|
1353 |
+
},
|
1354 |
+
{
|
1355 |
+
"epoch": 2.99,
|
1356 |
+
"learning_rate": 0.0,
|
1357 |
+
"loss": 0.1739,
|
1358 |
+
"step": 2210
|
1359 |
+
},
|
1360 |
+
{
|
1361 |
+
"epoch": 3.0,
|
1362 |
+
"eval_loss": 0.20216532051563263,
|
1363 |
+
"eval_rouge1": 0.0,
|
1364 |
+
"eval_rouge2": 0.0,
|
1365 |
+
"eval_rougeL": 0.0,
|
1366 |
+
"eval_rougeLsum": 0.0,
|
1367 |
+
"eval_runtime": 75.4566,
|
1368 |
+
"eval_samples_per_second": 5.394,
|
1369 |
+
"eval_steps_per_second": 2.704,
|
1370 |
+
"step": 2217
|
1371 |
+
},
|
1372 |
+
{
|
1373 |
+
"epoch": 3.0,
|
1374 |
+
"step": 2217,
|
1375 |
+
"total_flos": 3.2211655354615726e+18,
|
1376 |
+
"train_loss": 0.2087684871917057,
|
1377 |
+
"train_runtime": 2567.1485,
|
1378 |
+
"train_samples_per_second": 6.916,
|
1379 |
+
"train_steps_per_second": 0.864
|
1380 |
+
}
|
1381 |
+
],
|
1382 |
+
"max_steps": 2217,
|
1383 |
+
"num_train_epochs": 3,
|
1384 |
+
"total_flos": 3.2211655354615726e+18,
|
1385 |
+
"trial_name": null,
|
1386 |
+
"trial_params": null
|
1387 |
+
}
|