llama_finetune_cs_20_cot

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.3122

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: 0.0001
• train_batch_size: 4
• 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: 5
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.5397	1.0	150	1.4682
1.2842	2.0	300	1.5033
1.0418	3.0	450	1.7145
0.3202	4.0	600	1.9554
0.3109	5.0	750	2.2233
0.3597	6.0	900	2.4190
0.1135	7.0	1050	2.4910
0.0863	8.0	1200	2.5746
0.1082	9.0	1350	2.6889
0.0757	10.0	1500	2.7235
0.0818	11.0	1650	2.8289
0.0698	12.0	1800	2.8434
0.0714	13.0	1950	2.7956
0.0779	14.0	2100	2.9733
0.067	15.0	2250	3.0476
0.065	16.0	2400	3.1342
0.067	17.0	2550	3.1901
0.0777	18.0	2700	3.2490
0.0647	19.0	2850	3.3113
0.0661	20.0	3000	3.3122

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.0+cu121
• Datasets 2.13.1
• Tokenizers 0.14.1