Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e-08
train_batch_size: 1
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 100
training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.6914	0.2004	50	0.6919	0.0004	-0.0022	0.0900	0.0026	-15.2856	-9.3787	-1.0954	-1.0968
0.6938	0.4008	100	0.6918	0.0000	-0.0027	0.1050	0.0027	-15.2908	-9.3826	-1.0961	-1.0975
0.6926	0.6012	150	0.6915	0.0016	-0.0018	0.0900	0.0034	-15.2822	-9.3672	-1.0959	-1.0973
0.6948	0.8016	200	0.6918	-0.0002	-0.0030	0.1000	0.0028	-15.2940	-9.3849	-1.0955	-1.0969
0.6909	1.0020	250	0.6917	-0.0000	-0.0030	0.0850	0.0030	-15.2939	-9.3829	-1.0959	-1.0973
0.6935	1.2024	300	0.6926	0.0000	-0.0011	0.0800	0.0011	-15.2744	-9.3825	-1.0964	-1.0978
0.6939	1.4028	350	0.6918	-0.0003	-0.0031	0.0850	0.0028	-15.2946	-9.3858	-1.0962	-1.0976
0.6949	1.6032	400	0.6911	0.0007	-0.0036	0.0950	0.0043	-15.2994	-9.3754	-1.0962	-1.0977
0.6924	1.8036	450	0.6920	-0.0003	-0.0028	0.1000	0.0025	-15.2913	-9.3856	-1.0961	-1.0975
0.6929	2.0040	500	0.6915	0.0000	-0.0034	0.1000	0.0035	-15.2981	-9.3826	-1.0961	-1.0975
0.6922	2.2044	550	0.6931	-0.0011	-0.0012	0.0800	0.0001	-15.2760	-9.3935	-1.0962	-1.0976
0.694	2.4048	600	0.6926	-0.0001	-0.0014	0.0850	0.0013	-15.2774	-9.3837	-1.0961	-1.0975
0.6915	2.6052	650	0.6921	0.0003	-0.0019	0.0850	0.0021	-15.2825	-9.3800	-1.0957	-1.0972
0.6937	2.8056	700	0.6927	-0.0010	-0.0021	0.0850	0.0011	-15.2845	-9.3929	-1.0961	-1.0975
0.6954	3.0060	750	0.6925	0.0010	-0.0004	0.0800	0.0015	-15.2682	-9.3724	-1.0960	-1.0974
0.6945	3.2064	800	0.6929	0.0000	-0.0006	0.0700	0.0006	-15.2695	-9.3826	-1.0956	-1.0971
0.6911	3.4068	850	0.6924	0.0002	-0.0014	0.0750	0.0016	-15.2780	-9.3805	-1.0955	-1.0970
0.6944	3.6072	900	0.6922	-0.0000	-0.0020	0.0850	0.0020	-15.2842	-9.3833	-1.0956	-1.0970
0.6925	3.8076	950	0.6922	-0.0000	-0.0020	0.0850	0.0020	-15.2842	-9.3833	-1.0956	-1.0970
0.6951	4.0080	1000	0.6922	-0.0000	-0.0020	0.0850	0.0020	-15.2842	-9.3833	-1.0956	-1.0970

Framework versions

Transformers 4.41.2
Pytorch 2.0.0+cu117
Datasets 2.20.0
Tokenizers 0.19.1

tsavage68/Summary_L3_1000steps_1e8rate_01beta_CSFTDPO

Summary_L3_1000steps_1e8rate_01beta_CSFTDPO

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

tsavage68/
Summary_L3_1000steps_1e8rate_01beta_CSFTDPO