AlekseyKorshuk commited on
Commit
5d2a9ba
1 Parent(s): 9ef6b0c

End of training

Browse files
Files changed (4) hide show
  1. all_results.json +10 -10
  2. eval_results.json +7 -6
  3. train_results.json +4 -4
  4. trainer_state.json +13 -13
all_results.json CHANGED
@@ -1,15 +1,15 @@
1
  {
2
  "epoch": 0.99,
3
- "eval_accuracy": 0.03542682926829268,
4
- "eval_loss": 3.253744602203369,
5
- "eval_runtime": 7.9335,
6
  "eval_samples": 100,
7
- "eval_samples_per_second": 12.605,
8
- "eval_steps_per_second": 0.504,
9
- "perplexity": 25.887095523441527,
10
- "train_loss": 1.0375519144919612,
11
- "train_runtime": 120.7864,
12
  "train_samples": 1000,
13
- "train_samples_per_second": 8.279,
14
- "train_steps_per_second": 0.257
15
  }
 
1
  {
2
  "epoch": 0.99,
3
+ "eval_accuracy": 0.037317073170731706,
4
+ "eval_loss": 3.0486419200897217,
5
+ "eval_runtime": 2.253,
6
  "eval_samples": 100,
7
+ "eval_samples_per_second": 44.386,
8
+ "eval_steps_per_second": 1.775,
9
+ "perplexity": 21.086687561040215,
10
+ "train_loss": 3.260736426999492,
11
+ "train_runtime": 119.8588,
12
  "train_samples": 1000,
13
+ "train_samples_per_second": 8.343,
14
+ "train_steps_per_second": 0.259
15
  }
eval_results.json CHANGED
@@ -1,9 +1,10 @@
1
  {
2
- "eval_accuracy": 0.03542682926829268,
3
- "eval_loss": 3.253744602203369,
4
- "eval_runtime": 7.9335,
 
5
  "eval_samples": 100,
6
- "eval_samples_per_second": 12.605,
7
- "eval_steps_per_second": 0.504,
8
- "perplexity": 25.887095523441527
9
  }
 
1
  {
2
+ "epoch": 0.99,
3
+ "eval_accuracy": 0.037317073170731706,
4
+ "eval_loss": 3.0486419200897217,
5
+ "eval_runtime": 2.253,
6
  "eval_samples": 100,
7
+ "eval_samples_per_second": 44.386,
8
+ "eval_steps_per_second": 1.775,
9
+ "perplexity": 21.086687561040215
10
  }
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "epoch": 0.99,
3
- "train_loss": 1.0375519144919612,
4
- "train_runtime": 120.7864,
5
  "train_samples": 1000,
6
- "train_samples_per_second": 8.279,
7
- "train_steps_per_second": 0.257
8
  }
 
1
  {
2
  "epoch": 0.99,
3
+ "train_loss": 3.260736426999492,
4
+ "train_runtime": 119.8588,
5
  "train_samples": 1000,
6
+ "train_samples_per_second": 8.343,
7
+ "train_steps_per_second": 0.259
8
  }
trainer_state.json CHANGED
@@ -9,33 +9,33 @@
9
  "log_history": [
10
  {
11
  "epoch": 0.03,
12
- "learning_rate": 5e-05,
13
- "loss": 13.7576,
14
  "step": 1
15
  },
16
  {
17
  "epoch": 0.8,
18
- "learning_rate": 1.129032258064516e-05,
19
- "loss": 0.761,
20
  "step": 25
21
  },
22
  {
23
  "epoch": 0.8,
24
- "eval_accuracy": 0.0851829268292683,
25
- "eval_loss": 0.21037571132183075,
26
- "eval_runtime": 2.0103,
27
- "eval_samples_per_second": 49.743,
28
- "eval_steps_per_second": 1.99,
29
  "step": 25
30
  },
31
  {
32
  "epoch": 0.99,
33
  "step": 31,
34
  "total_flos": 938904995758080.0,
35
- "train_loss": 1.0375519144919612,
36
- "train_runtime": 120.7864,
37
- "train_samples_per_second": 8.279,
38
- "train_steps_per_second": 0.257
39
  }
40
  ],
41
  "max_steps": 31,
 
9
  "log_history": [
10
  {
11
  "epoch": 0.03,
12
+ "learning_rate": 4.8387096774193554e-05,
13
+ "loss": 3.5576,
14
  "step": 1
15
  },
16
  {
17
  "epoch": 0.8,
18
+ "learning_rate": 9.67741935483871e-06,
19
+ "loss": 3.2766,
20
  "step": 25
21
  },
22
  {
23
  "epoch": 0.8,
24
+ "eval_accuracy": 0.03723577235772358,
25
+ "eval_loss": 3.0587494373321533,
26
+ "eval_runtime": 2.0061,
27
+ "eval_samples_per_second": 49.847,
28
+ "eval_steps_per_second": 1.994,
29
  "step": 25
30
  },
31
  {
32
  "epoch": 0.99,
33
  "step": 31,
34
  "total_flos": 938904995758080.0,
35
+ "train_loss": 3.260736426999492,
36
+ "train_runtime": 119.8588,
37
+ "train_samples_per_second": 8.343,
38
+ "train_steps_per_second": 0.259
39
  }
40
  ],
41
  "max_steps": 31,