metadata
tags:
- trl
- sft
- generated_from_trainer
base_model: microsoft/codebert-base-mlm
datasets:
- apps
model-index:
- name: SFTCodeBertbase-mlm-APPS5k
results: []
SFTCodeBertbase-mlm-APPS5k
This model is a fine-tuned version of microsoft/codebert-base-mlm on the apps dataset. It achieves the following results on the evaluation set:
- Loss: 1.9251
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: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 5000
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 3.2308 | 1.61 | 500 | 2.3196 |
| 2.1364 | 3.22 | 1000 | 2.1168 |
| 1.9211 | 4.83 | 1500 | 2.0125 |
| 1.776 | 6.44 | 2000 | 1.9736 |
| 1.6872 | 8.05 | 2500 | 1.9470 |
| 1.6137 | 9.65 | 3000 | 1.9344 |
| 1.5724 | 11.26 | 3500 | 1.9288 |
| 1.533 | 12.87 | 4000 | 1.9267 |
| 1.5211 | 14.48 | 4500 | 1.9246 |
| 1.5115 | 16.09 | 5000 | 1.9251 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2