Qwen_32B_brevity_finetuned

This model is a fine-tuned version of Qwen/Qwen1.5-32B-Chat on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4598

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 30
• training_steps: 300

Training results

Training Loss	Epoch	Step	Validation Loss
3.5689	1.0	10	3.4447
3.3897	2.0	20	3.1637
3.0419	3.0	30	2.7805
2.646	4.0	40	2.3927
2.2145	5.0	50	1.9464
1.6159	6.0	60	1.2767
0.8947	7.0	70	0.7462
0.5729	8.0	80	0.6091
0.4858	9.0	90	0.5534
0.4338	10.0	100	0.5046
0.3885	11.0	110	0.4680
0.3457	12.0	120	0.4339
0.3003	13.0	130	0.4091
0.2573	14.0	140	0.3947
0.2167	15.0	150	0.3864
0.1771	16.0	160	0.3882
0.1399	17.0	170	0.3901
0.1083	18.0	180	0.3994
0.0835	19.0	190	0.4049
0.0601	20.0	200	0.4176
0.0483	21.0	210	0.4301
0.0363	22.0	220	0.4378
0.0313	23.0	230	0.4436
0.0278	24.0	240	0.4486
0.023	25.0	250	0.4504
0.0205	26.0	260	0.4549
0.0199	27.0	270	0.4586
0.0191	28.0	280	0.4584
0.0183	29.0	290	0.4593
0.0204	30.0	300	0.4598

Framework versions

• PEFT 0.11.1
• Transformers 4.41.0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1