This model is a fine-tuned version of deepseek-ai/deepseek-coder-1.3b-base on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.5
Eval/rewards/chosen: -1111.5853
Eval/logps/chosen: -11153.2350
Eval/rewards/rejected: -1076.7920
Eval/logps/rejected: -10842.3708
Eval/rewards/margins: -34.7933
Eval/kl: 0.0

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: 0.07
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 16
total_train_batch_size: 128
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
lr_scheduler_warmup_steps: 200
num_epochs: 7
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.5004	1.1552	100	0.5
0.5005	2.3105	200	0.5
0.5002	3.4657	300	0.5
0.4984	4.6209	400	0.5
0.4992	5.7762	500	0.5
0.5008	6.9314	600	0.5

Framework versions

Transformers 4.45.0
Pytorch 2.5.1+cu124
Datasets 2.19.2
Tokenizers 0.20.3

stojchet/kto5

kto5

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

stojchet/
kto5