metadata

license: gemma
library_name: peft
tags:
  - trl
  - reward-trainer
  - generated_from_trainer
metrics:
  - accuracy
base_model: google/gemma-2b
model-index:
  - name: RM-HH-Mix_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenFalse
    results: []

RM-HH-Mix_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenFalse

This model is a fine-tuned version of google/gemma-2b on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.0445
Accuracy: 0.9815

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: 1.41e-05
train_batch_size: 1
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.8191	0.06	250	0.5824	0.695
0.6294	0.11	500	0.1346	0.953
0.5811	0.17	750	0.0888	0.9705
0.5753	0.22	1000	0.0684	0.975
0.5539	0.28	1250	0.0588	0.979
0.5764	0.33	1500	0.0595	0.9785
0.5261	0.39	1750	0.0558	0.979
0.5423	0.44	2000	0.0533	0.9795
0.5261	0.5	2250	0.0501	0.98
0.5363	0.56	2500	0.0485	0.98
0.5051	0.61	2750	0.0472	0.981
0.5157	0.67	3000	0.0509	0.9795
0.5368	0.72	3250	0.0507	0.9785
0.5281	0.78	3500	0.0467	0.981
0.5005	0.83	3750	0.0450	0.9815
0.5239	0.89	4000	0.0445	0.9815
0.5111	0.94	4250	0.0445	0.9815

Framework versions

PEFT 0.9.0
Transformers 4.38.2
Pytorch 2.1.2
Datasets 2.18.0
Tokenizers 0.15.2