Howard881010/heat_transfer_sft_10000

sft_10000

This model is a fine-tuned version of mistralai/Mistral-Nemo-Instruct-2407 on the heat_transfer_10000 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0003

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: 12
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• total_train_batch_size: 24
• total_eval_batch_size: 2
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
0.0082	0.0533	20	0.0030
0.0028	0.1067	40	0.0027
0.0029	0.16	60	0.0026
0.0025	0.2133	80	0.0025
0.003	0.2667	100	0.0026
0.0025	0.32	120	0.0024
0.002	0.3733	140	0.0022
0.0014	0.4267	160	0.0012
0.0011	0.48	180	0.0011
0.0009	0.5333	200	0.0009
0.0006	0.5867	220	0.0008
0.001	0.64	240	0.0008
0.0009	0.6933	260	0.0010
0.0008	0.7467	280	0.0008
0.0009	0.8	300	0.0007
0.0008	0.8533	320	0.0008
0.0007	0.9067	340	0.0010
0.0007	0.96	360	0.0007
0.0006	1.0133	380	0.0006
0.0006	1.0667	400	0.0005
0.0006	1.12	420	0.0006
0.0005	1.1733	440	0.0004
0.0006	1.2267	460	0.0006
0.0007	1.28	480	0.0006
0.0005	1.3333	500	0.0005
0.0005	1.3867	520	0.0006
0.0006	1.44	540	0.0005
0.0004	1.4933	560	0.0004
0.0003	1.5467	580	0.0004
0.0005	1.6	600	0.0004
0.0004	1.6533	620	0.0004
0.0002	1.7067	640	0.0004
0.0003	1.76	660	0.0004
0.0004	1.8133	680	0.0003
0.0003	1.8667	700	0.0003
0.0003	1.92	720	0.0003
0.0003	1.9733	740	0.0003

Framework versions

• PEFT 0.12.0
• Transformers 4.46.0
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.20.1