Hugging Face's logo

Shahradmz
/
OLMo-1B-hf-DPO-constitution-1

Transformers
Safetensors
Generated from Trainer
trl
dpo
Model card Files Community
Shahradmz commited on
Commit
97df398
·
verified ·
1 Parent(s): 836d5b9

Model save

Browse files
Files changed (2) hide show
  1. all_results.json +15 -0
  2. eval_results.json +15 -0
all_results.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.9998604326587579,
3
+ "eval_logits/chosen": -1.0269631147384644,
4
+ "eval_logits/rejected": -1.061506986618042,
5
+ "eval_logps/chosen": -112.56452178955078,
6
+ "eval_logps/rejected": -272.78466796875,
7
+ "eval_loss": 0.000979769160039723,
8
+ "eval_rewards/accuracies": 0.9996050596237183,
9
+ "eval_rewards/chosen": -3.615877628326416,
10
+ "eval_rewards/margins": 16.76466178894043,
11
+ "eval_rewards/rejected": -20.380538940429688,
12
+ "eval_runtime": 151.8806,
13
+ "eval_samples_per_second": 33.302,
14
+ "eval_steps_per_second": 4.168
15
+ }
eval_results.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.9998604326587579,
3
+ "eval_logits/chosen": -1.0269631147384644,
4
+ "eval_logits/rejected": -1.061506986618042,
5
+ "eval_logps/chosen": -112.56452178955078,
6
+ "eval_logps/rejected": -272.78466796875,
7
+ "eval_loss": 0.000979769160039723,
8
+ "eval_rewards/accuracies": 0.9996050596237183,
9
+ "eval_rewards/chosen": -3.615877628326416,
10
+ "eval_rewards/margins": 16.76466178894043,
11
+ "eval_rewards/rejected": -20.380538940429688,
12
+ "eval_runtime": 151.8806,
13
+ "eval_samples_per_second": 33.302,
14
+ "eval_steps_per_second": 4.168
15
+ }