UTI_M2_1000steps_1e8rate_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: 2.6170

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-08
• 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
2.6263	0.3333	25	2.6302
2.6415	0.6667	50	2.6306
2.6336	1.0	75	2.6298
2.6804	1.3333	100	2.6284
2.5885	1.6667	125	2.6291
2.542	2.0	150	2.6286
2.6601	2.3333	175	2.6249
2.6936	2.6667	200	2.6240
2.6593	3.0	225	2.6238
2.6088	3.3333	250	2.6237
2.542	3.6667	275	2.6208
2.5635	4.0	300	2.6185
2.6825	4.3333	325	2.6183
2.5987	4.6667	350	2.6198
2.6421	5.0	375	2.6196
2.6426	5.3333	400	2.6183
2.6517	5.6667	425	2.6189
2.5762	6.0	450	2.6190
2.6801	6.3333	475	2.6190
2.5918	6.6667	500	2.6169
2.6354	7.0	525	2.6163
2.6433	7.3333	550	2.6156
2.7024	7.6667	575	2.6172
2.575	8.0	600	2.6176
2.6109	8.3333	625	2.6165
2.6088	8.6667	650	2.6195
2.698	9.0	675	2.6185
2.612	9.3333	700	2.6166
2.615	9.6667	725	2.6188
2.681	10.0	750	2.6165
2.6118	10.3333	775	2.6190
2.5875	10.6667	800	2.6168
2.6859	11.0	825	2.6170
2.5953	11.3333	850	2.6171
2.5884	11.6667	875	2.6169
2.6721	12.0	900	2.6170
2.5951	12.3333	925	2.6170
2.6112	12.6667	950	2.6170
2.5842	13.0	975	2.6170
2.5962	13.3333	1000	2.6170

Framework versions

• Transformers 4.41.2
• Pytorch 2.0.0+cu117
• Datasets 2.19.2
• Tokenizers 0.19.1