Hugging Face's logo

GItaf
/
JointGPT2-warmup-from-CLS

Text Generation
Transformers
PyTorch
gpt2
generated_from_trainer
Inference Endpoints
text-generation-inference
Model card Files Community
GItaf commited on
Commit
5b0b191
•
1 Parent(s): 279e2dd

End of training

Browse files
Files changed (3) hide show
  1. all_results.json +15 -15
  2. train_results.json +4 -4
  3. with_eval_results.json +11 -11
all_results.json CHANGED
@@ -1,19 +1,19 @@
1
  {
2
  "epoch": 3.0,
3
- "eval_CLS loss": 2.33956241607666,
4
- "eval_CLS_accuracy": 0.622478386167147,
5
- "eval_CLS_f1": 0.615766951227354,
6
- "eval_CLS_precision": 0.619580054950422,
7
- "eval_CLS_recall": 0.622478386167147,
8
- "eval_LM loss": 11.057974815368652,
9
- "eval_Perplexity": "63447.93",
10
- "eval_loss": 2.3409106731414795,
11
- "eval_runtime": 116.8259,
12
- "eval_samples_per_second": 14.851,
13
- "eval_steps_per_second": 7.43,
14
  "total_flos": 1.088098614116352e+16,
15
- "train_loss": 0.9344328573411984,
16
- "train_runtime": 3737.2556,
17
- "train_samples_per_second": 5.571,
18
- "train_steps_per_second": 2.785
19
  }
 
1
  {
2
  "epoch": 3.0,
3
+ "eval_CLS loss": 1.8206377029418945,
4
+ "eval_CLS_accuracy": 0.6121037463976945,
5
+ "eval_CLS_f1": 0.6063092264509424,
6
+ "eval_CLS_precision": 0.6096390716759156,
7
+ "eval_CLS_recall": 0.6121037463976945,
8
+ "eval_LM loss": 4.016696453094482,
9
+ "eval_Perplexity": "55.52",
10
+ "eval_loss": 4.472171783447266,
11
+ "eval_runtime": 103.1433,
12
+ "eval_samples_per_second": 16.821,
13
+ "eval_steps_per_second": 8.415,
14
  "total_flos": 1.088098614116352e+16,
15
+ "train_loss": 4.316991906069885,
16
+ "train_runtime": 5832.0918,
17
+ "train_samples_per_second": 3.57,
18
+ "train_steps_per_second": 1.785
19
  }
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "epoch": 3.0,
3
  "total_flos": 1.088098614116352e+16,
4
- "train_loss": 0.9344328573411984,
5
- "train_runtime": 3737.2556,
6
- "train_samples_per_second": 5.571,
7
- "train_steps_per_second": 2.785
8
  }
 
1
  {
2
  "epoch": 3.0,
3
  "total_flos": 1.088098614116352e+16,
4
+ "train_loss": 4.316991906069885,
5
+ "train_runtime": 5832.0918,
6
+ "train_samples_per_second": 3.57,
7
+ "train_steps_per_second": 1.785
8
  }
with_eval_results.json CHANGED
@@ -1,14 +1,14 @@
1
  {
2
  "epoch": 3.0,
3
- "eval_CLS loss": 2.33956241607666,
4
- "eval_CLS_accuracy": 0.622478386167147,
5
- "eval_CLS_f1": 0.615766951227354,
6
- "eval_CLS_precision": 0.619580054950422,
7
- "eval_CLS_recall": 0.622478386167147,
8
- "eval_LM loss": 11.057974815368652,
9
- "eval_Perplexity": "63447.93",
10
- "eval_loss": 2.3409106731414795,
11
- "eval_runtime": 116.8259,
12
- "eval_samples_per_second": 14.851,
13
- "eval_steps_per_second": 7.43
14
  }
 
1
  {
2
  "epoch": 3.0,
3
+ "eval_CLS loss": 1.8206377029418945,
4
+ "eval_CLS_accuracy": 0.6121037463976945,
5
+ "eval_CLS_f1": 0.6063092264509424,
6
+ "eval_CLS_precision": 0.6096390716759156,
7
+ "eval_CLS_recall": 0.6121037463976945,
8
+ "eval_LM loss": 4.016696453094482,
9
+ "eval_Perplexity": "55.52",
10
+ "eval_loss": 4.472171783447266,
11
+ "eval_runtime": 103.1433,
12
+ "eval_samples_per_second": 16.821,
13
+ "eval_steps_per_second": 8.415
14
  }