metadata

license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - anli
base_model: EleutherAI/gpt-j-6b
model-index:
  - name: sft-trl-claim-128
    results: []

sft-trl-claim-128

This model is a fine-tuned version of EleutherAI/gpt-j-6b on the anli dataset. It achieves the following results on the evaluation set:

Loss: 0.3201

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: 8
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 100
num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
1.4364	0.08	1000	1.8760
0.9805	0.16	2000	1.2595
0.6629	0.24	3000	1.2970
0.4647	0.32	4000	0.9789
0.4579	0.4	5000	0.8591
0.383	0.48	6000	0.8866
0.4915	0.56	7000	0.4281
0.4139	0.64	8000	0.3946
0.2563	0.72	9000	0.3653
0.3179	0.8	10000	0.3528
0.4199	0.88	11000	0.3602
0.3877	0.96	12000	0.3457
0.2332	1.04	13000	0.3882
0.3817	1.11	14000	0.3604
0.2734	1.19	15000	0.3613
0.213	1.27	16000	0.3722
0.3154	1.35	17000	0.3378
0.2258	1.43	18000	0.3117
0.3198	1.51	19000	0.3213
0.2959	1.59	20000	0.3050
0.2588	1.67	21000	0.3190
0.2279	1.75	22000	0.3065
0.2988	1.83	23000	0.3077
0.3701	1.91	24000	0.3092
0.281	1.99	25000	0.3038
0.1743	2.07	26000	0.3542
0.1374	2.15	27000	0.3550
0.1282	2.23	28000	0.3386
0.1757	2.31	29000	0.3489
0.1371	2.39	30000	0.3316
0.1689	2.47	31000	0.3291
0.1882	2.55	32000	0.3292
0.1685	2.63	33000	0.3196
0.1775	2.71	34000	0.3320
0.1963	2.79	35000	0.3278
0.1733	2.87	36000	0.3221
0.1503	2.95	37000	0.3201

Framework versions

Transformers 4.30.2
Pytorch 2.0.1+cu117
Datasets 2.13.1
Tokenizers 0.13.3