Hugging Face's logo

BluebrainAI
/
rotating-head-gp-gpt2-medium-wikitext

Feature Extraction
Transformers
Safetensors
rotating-head-gpt2
Generated from Trainer
custom_code
Model card Files
xet
Community
shivanandmn commited on
Commit
c98c6b3
·
verified ·
1 Parent(s): 7b3ec04

Model save

Browse files
Files changed (5) hide show
  1. all_results.json +6 -6
  2. eval_results.json +3 -3
  3. train_results.json +3 -3
  4. trainer_state.json +3 -3
  5. training_args.bin +1 -1
all_results.json CHANGED
@@ -4,14 +4,14 @@
4
  "eval_bleu": 0.13436705184218095,
5
  "eval_loss": 3.19417405128479,
6
  "eval_perplexity": 24.39002046460787,
7
- "eval_runtime": 12.1627,
8
  "eval_samples": 1141,
9
- "eval_samples_per_second": 93.812,
10
- "eval_steps_per_second": 1.48,
11
  "perplexity": 24.39002046460787,
12
  "total_flos": 1.0587061010143642e+18,
13
  "train_loss": 0.0,
14
- "train_runtime": 0.6382,
15
- "train_samples_per_second": 893032.162,
16
- "train_steps_per_second": 13960.237
17
  }
 
4
  "eval_bleu": 0.13436705184218095,
5
  "eval_loss": 3.19417405128479,
6
  "eval_perplexity": 24.39002046460787,
7
+ "eval_runtime": 12.6247,
8
  "eval_samples": 1141,
9
+ "eval_samples_per_second": 90.378,
10
+ "eval_steps_per_second": 1.426,
11
  "perplexity": 24.39002046460787,
12
  "total_flos": 1.0587061010143642e+18,
13
  "train_loss": 0.0,
14
+ "train_runtime": 0.6657,
15
+ "train_samples_per_second": 856247.444,
16
+ "train_steps_per_second": 13385.204
17
  }
eval_results.json CHANGED
@@ -4,9 +4,9 @@
4
  "eval_bleu": 0.13436705184218095,
5
  "eval_loss": 3.19417405128479,
6
  "eval_perplexity": 24.39002046460787,
7
- "eval_runtime": 12.1627,
8
  "eval_samples": 1141,
9
- "eval_samples_per_second": 93.812,
10
- "eval_steps_per_second": 1.48,
11
  "perplexity": 24.39002046460787
12
  }
 
4
  "eval_bleu": 0.13436705184218095,
5
  "eval_loss": 3.19417405128479,
6
  "eval_perplexity": 24.39002046460787,
7
+ "eval_runtime": 12.6247,
8
  "eval_samples": 1141,
9
+ "eval_samples_per_second": 90.378,
10
+ "eval_steps_per_second": 1.426,
11
  "perplexity": 24.39002046460787
12
  }
train_results.json CHANGED
@@ -2,7 +2,7 @@
2
  "epoch": 5.0,
3
  "total_flos": 1.0587061010143642e+18,
4
  "train_loss": 0.0,
5
- "train_runtime": 0.6382,
6
- "train_samples_per_second": 893032.162,
7
- "train_steps_per_second": 13960.237
8
  }
 
2
  "epoch": 5.0,
3
  "total_flos": 1.0587061010143642e+18,
4
  "train_loss": 0.0,
5
+ "train_runtime": 0.6657,
6
+ "train_samples_per_second": 856247.444,
7
+ "train_steps_per_second": 13385.204
8
  }
trainer_state.json CHANGED
@@ -823,9 +823,9 @@
823
  "step": 8910,
824
  "total_flos": 1.0587061010143642e+18,
825
  "train_loss": 0.0,
826
- "train_runtime": 0.6382,
827
- "train_samples_per_second": 893032.162,
828
- "train_steps_per_second": 13960.237
829
  }
830
  ],
831
  "logging_steps": 100,
 
823
  "step": 8910,
824
  "total_flos": 1.0587061010143642e+18,
825
  "train_loss": 0.0,
826
+ "train_runtime": 0.6657,
827
+ "train_samples_per_second": 856247.444,
828
+ "train_steps_per_second": 13385.204
829
  }
830
  ],
831
  "logging_steps": 100,
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d204bc0f0d80a596cb44ec14c19b6ac00cda421783ca4ff18679c8fdfe325c30
3
  size 5560
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f310d1f2acbf9b789d07f19240d1ba587465af941995b095af8542bac8d1d00e
3
  size 5560