metadata

license: apache-2.0
library_name: peft
tags:
  - trl
  - sft
  - generated_from_trainer
base_model: mistralai/Mistral-7B-v0.1
model-index:
  - name: results
    results: []

results

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.9724

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: 0.0002
train_batch_size: 4
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 0.03
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.0248	0.03	50	1.0145
1.0168	0.06	100	1.0078
1.008	0.09	150	1.0058
1.0082	0.12	200	1.0030
0.9846	0.14	250	1.0005
0.9807	0.17	300	0.9998
0.9968	0.2	350	0.9992
0.9834	0.23	400	0.9967
1.0267	0.26	450	0.9953
1.0119	0.29	500	0.9937
0.9759	0.32	550	0.9939
0.9978	0.35	600	0.9921
1.0145	0.38	650	0.9901
1.0064	0.4	700	0.9897
0.9949	0.43	750	0.9890
0.9936	0.46	800	0.9865
0.9944	0.49	850	0.9852
0.9819	0.52	900	0.9845
0.9991	0.55	950	0.9826
0.9874	0.58	1000	0.9812
0.981	0.61	1050	0.9798
0.9807	0.64	1100	0.9789
0.9639	0.67	1150	0.9776
0.9645	0.69	1200	0.9767
0.9788	0.72	1250	0.9758
0.9823	0.75	1300	0.9751
0.9906	0.78	1350	0.9745
0.9536	0.81	1400	0.9738
0.9635	0.84	1450	0.9732
0.9754	0.87	1500	0.9729
0.9785	0.9	1550	0.9727
0.9828	0.93	1600	0.9725
0.9951	0.95	1650	0.9724
0.983	0.98	1700	0.9724

Framework versions

PEFT 0.9.0
Transformers 4.38.2
Pytorch 2.2.1+cu121
Datasets 2.18.0
Tokenizers 0.15.2