mistral-journal-finetune

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.3365

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: 2.5e-05
• train_batch_size: 2
• 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: 1
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss
1.9309	2.0833	25	1.6414
0.5146	4.1667	50	1.8892
0.2481	6.25	75	1.9643
0.1804	8.3333	100	1.9184
0.1683	10.4167	125	1.9770
0.1582	12.5	150	2.1538
0.1591	14.5833	175	2.1592
0.1509	16.6667	200	2.1474
0.1478	18.75	225	2.1839
0.1465	20.8333	250	2.2255
0.1465	22.9167	275	2.2356
0.1427	25.0	300	2.2581
0.144	27.0833	325	2.2707
0.139	29.1667	350	2.2694
0.1437	31.25	375	2.2956
0.141	33.3333	400	2.3087
0.1359	35.4167	425	2.3056
0.1385	37.5	450	2.3213
0.1363	39.5833	475	2.3332
0.1359	41.6667	500	2.3365

Framework versions

• PEFT 0.11.1
• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1