metadata

license: apache-2.0
base_model: alignment-handbook/zephyr-7b-sft-full
tags:
  - trl
  - sft
  - generated_from_trainer
datasets:
  - generator
model-index:
  - name: zephyr-7b-sft-safe-hhrlhf
    results: []

zephyr-7b-sft-safe-hhrlhf

This model is a fine-tuned version of alignment-handbook/zephyr-7b-sft-full on the generator dataset. It achieves the following results on the evaluation set:

Loss: 1.1996

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: 2e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
num_devices: 4
total_train_batch_size: 32
total_eval_batch_size: 32
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
1.4854	0.06	25	1.4540
1.43	0.12	50	1.4291
1.4188	0.18	75	1.4451
1.4465	0.25	100	1.4565
1.4499	0.31	125	1.4525
1.4387	0.37	150	1.4409
1.4368	0.43	175	1.4226
1.4095	0.49	200	1.4227
1.4065	0.55	225	1.4044
1.3747	0.62	250	1.3881
1.3847	0.68	275	1.3720
1.3526	0.74	300	1.3581
1.3072	0.8	325	1.3410
1.3146	0.86	350	1.3239
1.2783	0.92	375	1.3106
1.2773	0.99	400	1.2896
0.7303	1.05	425	1.3327
0.6911	1.11	450	1.3081
0.7044	1.17	475	1.3012
0.6597	1.23	500	1.2845
0.6717	1.29	525	1.2701
0.6538	1.35	550	1.2666
0.6484	1.42	575	1.2481
0.633	1.48	600	1.2404
0.6252	1.54	625	1.2332
0.6423	1.6	650	1.2212
0.6182	1.66	675	1.2115
0.6218	1.72	700	1.2099
0.6293	1.79	725	1.2038
0.5978	1.85	750	1.2014
0.6176	1.91	775	1.1995
0.6475	1.97	800	1.1996

Framework versions

Transformers 4.39.0.dev0
Pytorch 2.3.0+cu121
Datasets 2.14.6
Tokenizers 0.15.2