CoolFace
Modelpublic

tsavage68/Transaminitis_L3_1000steps_1e7rate_03beta_CSFTDPO

sourceHugging Facellama3updated 2y agoView on Hugging Face
0likes13downloads
Model Card

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

TransaminitisL31000steps1e7rate03beta_CSFTDPO

This model is a fine-tuned version of tsavage68/Transaminitis_L3_1000rate_1e7_SFT on an unknown dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.1392
  • —Rewards/chosen: 1.7582
  • —Rewards/rejected: -2.4140
  • —Rewards/accuracies: 0.9300
  • —Rewards/margins: 4.1722
  • —Logps/rejected: -26.6014
  • —Logps/chosen: -12.6736
  • —Logits/rejected: -1.0630
  • —Logits/chosen: -1.0536

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-07
  • —trainbatchsize: 2
  • —evalbatchsize: 1
  • —seed: 42
  • —gradientaccumulationsteps: 2
  • —totaltrainbatch_size: 4
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: cosine
  • —lrschedulerwarmup_steps: 100
  • —training_steps: 1000

Training results

Training LossEpochStepValidation LossRewards/chosenRewards/rejectedRewards/accuraciesRewards/marginsLogps/rejectedLogps/chosenLogits/rejectedLogits/chosen
0.69490.2250.6922-0.0089-0.01130.56000.0023-18.5922-18.5640-1.0661-1.0649
0.6890.4500.6902-0.0743-0.08080.57000.0065-18.8241-18.7820-1.0662-1.0650
0.69660.6750.6990-0.0262-0.02040.4500-0.0058-18.6227-18.6216-1.0666-1.0653
0.66280.81000.7163-0.1175-0.11320.4600-0.0042-18.9322-18.9257-1.0693-1.0680
0.70151.01250.6776-0.4612-0.55100.54000.0898-20.3914-20.0715-1.0730-1.0717
0.66681.21500.68000.06350.00840.46000.0551-18.5267-18.3224-1.0700-1.0688
0.64091.41750.6361-0.0704-0.19360.77000.1232-19.2001-18.7690-1.0737-1.0723
0.61231.62000.60740.0180-0.18490.73000.2030-19.1711-18.4741-1.0741-1.0726
0.60431.82250.59880.31890.09430.750.2245-18.2403-17.4714-1.0739-1.0724
0.56082.02500.56250.46390.16020.83000.3037-18.0205-16.9878-1.0739-1.0723
0.52062.22750.50840.54760.09710.87000.4504-18.2310-16.7090-1.0751-1.0732
0.47092.43000.46140.71690.12300.88000.5939-18.1448-16.1447-1.0753-1.0732
0.4332.63250.43881.12170.36900.85000.7527-17.3245-14.7951-1.0729-1.0703
0.28022.83500.31651.1549-0.08960.92001.2445-18.8533-14.6845-1.0737-1.0701
0.30443.03750.27441.3071-0.19210.90001.4992-19.1949-14.1771-1.0729-1.0688
0.2843.24000.23141.5290-0.33170.91001.8607-19.6603-13.4374-1.0713-1.0663
0.17713.44250.19471.6285-0.60310.93002.2317-20.5652-13.1057-1.0703-1.0645
0.21673.64500.18401.5135-0.96780.92002.4813-21.7808-13.4893-1.0704-1.0643
0.13953.84750.17711.5622-1.13040.92002.6926-22.3226-13.3267-1.0702-1.0636
0.21484.05000.18741.6351-1.20630.93002.8414-22.5757-13.0837-1.0669-1.0600
0.02864.25250.15771.6211-1.60790.92003.2289-23.9142-13.1307-1.0665-1.0588
0.13474.45500.15871.6299-1.77630.92003.4062-24.4757-13.1011-1.0654-1.0575
0.05754.65750.14731.6935-1.97900.92003.6725-25.1514-12.8892-1.0648-1.0564
0.1584.86000.15091.6992-2.02960.93003.7288-25.3200-12.8703-1.0651-1.0566
0.06075.06250.14751.6735-2.17520.94003.8488-25.8055-12.9557-1.0647-1.0558
0.11225.26500.14041.7217-2.20660.94003.9283-25.9101-12.7953-1.0637-1.0549
0.15175.46750.14481.7525-2.25150.94004.0040-26.0596-12.6925-1.0628-1.0538
0.04135.67000.14181.7609-2.30020.94004.0611-26.2220-12.6647-1.0635-1.0545
0.05285.87250.13581.7517-2.35390.92004.1056-26.4012-12.6954-1.0634-1.0542
0.2436.07500.13561.7676-2.38040.93004.1480-26.4894-12.6424-1.0623-1.0531
0.13096.27750.15221.7393-2.39230.93004.1315-26.5290-12.7367-1.0632-1.0540
0.19436.48000.14251.7703-2.39970.93004.1699-26.5535-12.6333-1.0628-1.0534
0.07546.68250.13081.7709-2.41430.94004.1852-26.6025-12.6314-1.0623-1.0529
0.06036.88500.14011.7587-2.40400.93004.1627-26.5682-12.6720-1.0630-1.0537
0.06037.08750.14021.7651-2.39760.93004.1627-26.5466-12.6505-1.0636-1.0543
0.0467.29000.13971.7519-2.42540.93004.1773-26.6393-12.6945-1.0631-1.0538
0.21027.49250.13901.7602-2.41680.93004.1770-26.6105-12.6669-1.0631-1.0537
0.21167.69500.13921.7582-2.41400.93004.1722-26.6014-12.6736-1.0630-1.0536
0.0947.89750.13921.7582-2.41400.93004.1722-26.6014-12.6736-1.0630-1.0536
0.11218.010000.13921.7582-2.41400.93004.1722-26.6014-12.6736-1.0630-1.0536

Framework versions

  • —Transformers 4.40.2
  • —Pytorch 2.0.0+cu117
  • —Datasets 2.19.1
  • —Tokenizers 0.19.1