Na_M2_1000steps_1e7_SFT

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3111

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: 1e-07
• train_batch_size: 2
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss
1.9785	0.2667	50	1.8984
0.815	0.5333	100	0.6662
0.4371	0.8	150	0.4306
0.3721	1.0667	200	0.3807
0.3439	1.3333	250	0.3367
0.3251	1.6	300	0.3266
0.3215	1.8667	350	0.3233
0.3156	2.1333	400	0.3205
0.3124	2.4	450	0.3183
0.3165	2.6667	500	0.3161
0.3128	2.9333	550	0.3130
0.3093	3.2	600	0.3120
0.311	3.4667	650	0.3109
0.3073	3.7333	700	0.3112
0.306	4.0	750	0.3115
0.307	4.2667	800	0.3112
0.3052	4.5333	850	0.3111
0.3048	4.8	900	0.3105
0.3034	5.0667	950	0.3111
0.3057	5.3333	1000	0.3111

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1