falcon7b-chunks-10k-v1_1

This model is a fine-tuned version of tiiuae/falcon-7b on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.6306

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: 0.0002
• train_batch_size: 4
• eval_batch_size: 8
• 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: constant
• lr_scheduler_warmup_ratio: 0.03
• training_steps: 1500

Training results

Training Loss	Epoch	Step	Validation Loss
1.9498	0.57	10	1.8499
1.8859	1.14	20	1.8045
1.7494	1.71	30	1.7604
1.6834	2.29	40	1.7328
1.4871	2.86	50	1.7014
1.3791	3.43	60	1.7721
1.315	4.0	70	1.7826
1.116	4.57	80	1.8231
1.0471	5.14	90	1.9451
0.871	5.71	100	1.8934
0.8318	6.29	110	1.9911
0.7367	6.86	120	2.0047
0.6208	7.43	130	2.1256
0.5716	8.0	140	2.0944
0.4414	8.57	150	2.1695
0.415	9.14	160	2.3015
0.3441	9.71	170	2.3119
0.2823	10.29	180	2.3515
0.2576	10.86	190	2.3800
0.2074	11.43	200	2.4828
0.1984	12.0	210	2.4218
0.151	12.57	220	2.5316
0.144	13.14	230	2.6191
0.1255	13.71	240	2.6013
0.1086	14.29	250	2.6306

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.13.0
• Tokenizers 0.13.3