tsavage68/Summary4500_M2_1000steps_1e8rate_01beta_CSFTDPO

Hyponatremia_M2_1000steps_1e8rate_01beta_CSFTDPO

This model is a fine-tuned version of tsavage68/Summary4500_M2_200steps_1e7rate_SFT on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6747
• Rewards/chosen: 0.0004
• Rewards/rejected: -0.0376
• Rewards/accuracies: 0.7520
• Rewards/margins: 0.0381
• Logps/rejected: -153.1061
• Logps/chosen: -93.7354
• Logits/rejected: -2.3509
• Logits/chosen: -2.3035

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
• 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
0.6837	0.0112	50	0.6927	-0.0020	-0.0036	0.4960	0.0017	-152.7662	-93.7594	-2.3523	-2.3049
0.6845	0.0224	100	0.6912	0.0015	-0.0032	0.5580	0.0047	-152.7614	-93.7247	-2.3530	-2.3055
0.6844	0.0336	150	0.6870	-0.0007	-0.0140	0.5640	0.0133	-152.8696	-93.7470	-2.3520	-2.3046
0.6667	0.0448	200	0.6836	-0.0018	-0.0219	0.6100	0.0201	-152.9489	-93.7576	-2.3522	-2.3048
0.6785	0.0559	250	0.6806	-0.0009	-0.0271	0.6940	0.0261	-153.0003	-93.7490	-2.3522	-2.3048
0.6748	0.0671	300	0.6782	-0.0019	-0.0329	0.6920	0.0311	-153.0590	-93.7583	-2.3512	-2.3038
0.6987	0.0783	350	0.6757	-0.0017	-0.0378	0.7140	0.0361	-153.1073	-93.7563	-2.3508	-2.3034
0.6323	0.0895	400	0.6739	-0.0008	-0.0406	0.7560	0.0398	-153.1360	-93.7480	-2.3511	-2.3037
0.6642	0.1007	450	0.6753	-0.0004	-0.0375	0.7340	0.0371	-153.1046	-93.7441	-2.3508	-2.3034
0.6692	0.1119	500	0.6726	-0.0014	-0.0438	0.7680	0.0424	-153.1682	-93.7542	-2.3499	-2.3025
0.6745	0.1231	550	0.6736	-0.0002	-0.0406	0.7320	0.0404	-153.1359	-93.7421	-2.3513	-2.3039
0.6661	0.1343	600	0.6741	-0.0001	-0.0398	0.7560	0.0396	-153.1274	-93.7413	-2.3514	-2.3040
0.6629	0.1454	650	0.6739	0.0002	-0.0397	0.7400	0.0398	-153.1265	-93.7381	-2.3518	-2.3043
0.6572	0.1566	700	0.6731	-0.0007	-0.0422	0.7460	0.0415	-153.1519	-93.7464	-2.3499	-2.3025
0.6694	0.1678	750	0.6742	-0.0011	-0.0403	0.7380	0.0392	-153.1327	-93.7505	-2.3509	-2.3034
0.6763	0.1790	800	0.6717	0.0021	-0.0424	0.7660	0.0445	-153.1538	-93.7188	-2.3509	-2.3035
0.669	0.1902	850	0.6758	-0.0005	-0.0364	0.7320	0.0358	-153.0933	-93.7452	-2.3509	-2.3035
0.6696	0.2014	900	0.6747	0.0004	-0.0376	0.7520	0.0381	-153.1061	-93.7354	-2.3509	-2.3035
0.6593	0.2126	950	0.6747	0.0004	-0.0376	0.7520	0.0381	-153.1061	-93.7354	-2.3509	-2.3035
0.6831	0.2238	1000	0.6747	0.0004	-0.0376	0.7520	0.0381	-153.1061	-93.7354	-2.3509	-2.3035

Framework versions

• Transformers 4.42.4
• Pytorch 2.0.0+cu117
• Datasets 2.20.0
• Tokenizers 0.19.1