This model is a fine-tuned version of tsavage68/UTI_L3_1000steps_1e5rate_SFT on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.0261
Rewards/chosen: 2.3344
Rewards/rejected: -5.4705
Rewards/accuracies: 0.9800
Rewards/margins: 7.8050
Logps/rejected: -54.2106
Logps/chosen: -24.5560
Logits/rejected: -1.1516
Logits/chosen: -1.1414

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: 2
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 2
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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.5446	0.3333	25	0.2409	0.7934	-0.6030	0.9800	1.3964	-44.4754	-27.6381	-1.1424	-1.1365
0.0009	0.6667	50	0.0261	2.3344	-5.4705	0.9800	7.8050	-54.2106	-24.5560	-1.1516	-1.1414

Framework versions

Transformers 4.41.2
Pytorch 2.0.0+cu117
Datasets 2.19.2
Tokenizers 0.19.1

tsavage68/UTI2_L3_50steps_1e6rate_05beta_CSFTDPO

UTI2_L3_50steps_1e6rate_05beta_CSFTDPO

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

tsavage68/
UTI2_L3_50steps_1e6rate_05beta_CSFTDPO