This model is a fine-tuned version of tsavage68/Summary4500_L3_100steps_1e6rate_SFT on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.0014
Rewards/chosen: -1.4084
Rewards/rejected: -18.4001
Rewards/accuracies: 0.9980
Rewards/margins: 16.9917
Logps/rejected: -317.1989
Logps/chosen: -98.2741
Logits/rejected: -1.0846
Logits/chosen: -1.0076

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-06
train_batch_size: 1
eval_batch_size: 1
seed: 42
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
0.0114	0.0112	50	0.0093	-0.2009	-5.9855	0.9980	5.7846	-193.0523	-86.1985	-1.1075	-1.0649
0.0	0.0224	100	0.0024	-0.9378	-10.4848	0.9980	9.5470	-238.0455	-93.5676	-1.1001	-1.0461
0.0	0.0336	150	0.0017	-1.0803	-12.7703	0.9980	11.6899	-260.8999	-94.9929	-1.0979	-1.0362
0.0	0.0448	200	0.0015	-2.1051	-16.1714	0.9980	14.0663	-294.9110	-105.2404	-1.0968	-1.0306
0.0	0.0559	250	0.0015	-1.2418	-15.6144	0.9980	14.3726	-289.3413	-96.6073	-1.0946	-1.0268
0.0	0.0671	300	0.0015	-1.2850	-16.0588	0.9980	14.7738	-293.7853	-97.0396	-1.0920	-1.0240
0.0	0.0783	350	0.0014	-1.5607	-17.5217	0.9980	15.9609	-308.4142	-99.7972	-1.0919	-1.0200
0.0	0.0895	400	0.0014	-1.5463	-17.5816	0.9980	16.0353	-309.0129	-99.6524	-1.0908	-1.0187
0.0	0.1007	450	0.0014	-1.5768	-17.6781	0.9980	16.1012	-309.9779	-99.9583	-1.0908	-1.0182
0.0	0.1119	500	0.0014	-1.4380	-17.9331	0.9980	16.4952	-312.5286	-98.5695	-1.0817	-1.0071
0.0	0.1231	550	0.0014	-1.4831	-18.1851	0.9980	16.7020	-315.0485	-99.0211	-1.0852	-1.0099
0.0	0.1343	600	0.0014	-1.4779	-18.1900	0.9980	16.7121	-315.0977	-98.9690	-1.0853	-1.0100
0.0	0.1454	650	0.0014	-1.4375	-18.2718	0.9980	16.8342	-315.9149	-98.5652	-1.0861	-1.0096
0.0	0.1566	700	0.0014	-1.4049	-18.3712	0.9980	16.9664	-316.9096	-98.2383	-1.0854	-1.0084
0.0004	0.1678	750	0.0014	-1.4073	-18.3876	0.9980	16.9803	-317.0729	-98.2626	-1.0845	-1.0075
0.0	0.1790	800	0.0014	-1.4175	-18.4190	0.9980	17.0016	-317.3878	-98.3644	-1.0846	-1.0076
0.0001	0.1902	850	0.0014	-1.4088	-18.4040	0.9980	16.9952	-317.2370	-98.2774	-1.0844	-1.0074
0.0	0.2014	900	0.0014	-1.4115	-18.4067	0.9980	16.9952	-317.2642	-98.3050	-1.0845	-1.0074
0.0	0.2126	950	0.0014	-1.4069	-18.4091	0.9980	17.0022	-317.2884	-98.2590	-1.0845	-1.0075
0.0	0.2238	1000	0.0014	-1.4084	-18.4001	0.9980	16.9917	-317.1989	-98.2741	-1.0846	-1.0076

Framework versions

Transformers 4.42.4
Pytorch 2.0.0+cu117
Datasets 2.20.0
Tokenizers 0.19.1

tsavage68/Summary4500_L3_1000steps_1e6rate_01beta_CSFTDPO

Hyponatremia_L3_1000steps_1e6rate_01beta_CSFTDPO

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

tsavage68/
Summary4500_L3_1000steps_1e6rate_01beta_CSFTDPO