ctrltokyo/llm_prompt_mask_fill_model

This model is a fine-tuned version of distilbert-base-uncased on the code_instructions_120k dataset. It achieves the following results on the evaluation set:

Train Loss: 2.1215
Validation Loss: 1.5672
Epoch: 0

Model description

It's just distilbert-base-uncased with some fine tuning.

Intended uses & limitations

This model could be used for live autocompletion of PROMPTS in a coding-specific chatbot. Don't try this on code, because it won't work.

Training and evaluation data

Evaluated on 5% of training data. No further evaluation performed at this point. Trained on NVIDIA V100.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

optimizer: {'inner_optimizer': {'class_name': 'AdamWeightDecay', 'config': {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'WarmUp', 'config': {'initial_learning_rate': 2e-05, 'decay_schedule_fn': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 108, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'passive_serialization': True}, 'warmup_steps': 1000, 'power': 1.0, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01}}, 'dynamic': True, 'initial_scale': 32768.0, 'dynamic_growth_steps': 2000}
training_precision: mixed_float16

Training results

Train Loss	Validation Loss	Epoch
2.1215	1.5672	0

Framework versions

Transformers 4.31.0
TensorFlow 2.12.0
Datasets 2.14.1
Tokenizers 0.13.3