Hugging Face's logo

stojchet
/
k6

Text Generation
Transformers
Safetensors
llama
trl
kto
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Community
k6 / README.md
stojchet's picture
stojchet
End of training
4161868 verified
preview code
|
raw
history blame contribute delete
1.97 kB
metadata
license: other
base_model: deepseek-ai/deepseek-coder-1.3b-base
tags:
  - trl
  - kto
  - generated_from_trainer
model-index:
  - name: k6
    results: []

Visualize in Weights & Biases

k6

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: 0.1628
  • Eval/rewards/chosen: 1.9672
  • Eval/logps/chosen: -122.9227
  • Eval/rewards/rejected: -13.4303
  • Eval/logps/rejected: -315.3628
  • Eval/rewards/margins: 15.3975
  • Eval/kl: 0.9837

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: 5e-05
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 16
  • total_train_batch_size: 256
  • 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: 4
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
0.1152 1.7094 100 0.1699 10.6285
0.0151 3.4188 200 0.1628 0.9837

Framework versions

  • Transformers 4.43.0.dev0
  • Pytorch 2.2.2+cu121
  • Datasets 2.19.2
  • Tokenizers 0.19.1