chaymaemerhrioui/mistral-Brain_Model_ACC_Trainer

Brain_Model_ACC_Trainer

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1245

Model description

The system in question incorporates a sophisticated model designed to process user-provided inputs, specifically descriptions of various systems, and subsequently generate comprehensive and detailed descriptions of those systems. By analyzing the initial input, the model effectively elaborates on the provided information, producing an extensive and formalized depiction that enhances the understanding and documentation of the system described.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• 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: linear
• lr_scheduler_warmup_steps: 50
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.0754	0.1	20	0.9337
0.5144	0.2	40	0.3974
0.2526	0.3	60	0.2197
0.193	0.4	80	0.1900
0.1939	0.5	100	0.1733
0.1695	0.6	120	0.1645
0.1739	0.7	140	0.1567
0.1511	0.8	160	0.1504
0.1484	0.9	180	0.1462
0.1419	1.0	200	0.1435
0.1467	1.1	220	0.1410
0.1315	1.2	240	0.1384
0.1344	1.3	260	0.1370
0.1411	1.4	280	0.1355
0.1338	1.5	300	0.1346
0.128	1.6	320	0.1324
0.1312	1.7	340	0.1328
0.1191	1.8	360	0.1312
0.1228	1.9	380	0.1301
0.1317	2.0	400	0.1291
0.1152	2.1	420	0.1291
0.1167	2.2	440	0.1285
0.1178	2.3	460	0.1283
0.1184	2.4	480	0.1268
0.118	2.5	500	0.1267
0.1104	2.6	520	0.1258
0.1128	2.7	540	0.1254
0.1113	2.8	560	0.1249
0.1174	2.9	580	0.1247
0.1081	3.0	600	0.1245

Framework versions

• PEFT 0.15.2
• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.1