Mistral-7B-Instruct-v0.3-advisegpt-v0.2

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

• Loss: 0.0780
• Bleu: {'bleu': 0.9576887647563643, 'precisions': [0.9773669728326163, 0.9618578951912286, 0.9507543139197927, 0.9415539534224628], 'brevity_penalty': 0.9998937877305732, 'length_ratio': 0.9998937933706971, 'translation_length': 696681, 'reference_length': 696755}
• Rouge: {'rouge1': 0.9756788083049949, 'rouge2': 0.9583995226740446, 'rougeL': 0.9744286269286386, 'rougeLsum': 0.9754176834545093}
• Exact Match: {'exact_match': 0.0}

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: 2e-05
• train_batch_size: 3
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 10
• total_train_batch_size: 30
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Bleu	Exact Match	Validation Loss	Rouge
0.0674	1.0	1272	{'bleu': 0.9527886605929342, 'brevity_penalty': 1.0, 'length_ratio': 1.0000143520618172, 'precisions': [0.9750579097383083, 0.9573526809414263, 0.9448030782898647, 0.9344211337471139], 'reference_length': 696764, 'translation_length': 696774}	{'exact_match': 0.0}	0.0879	{'rouge1': 0.9733658563180658, 'rouge2': 0.9537375602435975, 'rougeL': 0.9714782202593812, 'rougeLsum': 0.97295717219818}
0.0515	2.0	2544	{'bleu': 0.9576841610471547, 'brevity_penalty': 0.9998937891025745, 'length_ratio': 0.9998937947425527, 'precisions': [0.9774131966871923, 0.9617936398993365, 0.9507289899724213, 0.9415792951573699], 'reference_length': 696764, 'translation_length': 696690}	{'exact_match': 0.0}	0.0783	{'rouge1': 0.9757072387880681, 'rouge2': 0.9584139466483359, 'rougeL': 0.9743902945474832, 'rougeLsum': 0.9754213243935133}
0.0574	2.9993	2997	{'bleu': 0.9566916740680499, 'brevity_penalty': 1.0, 'length_ratio': 1.00018514398892, 'precisions': [0.9768282813208511, 0.960876488636805, 0.9494536267704137, 0.9400012431679968], 'reference_length': 696755, 'translation_length': 696884}	{'exact_match': 0.0}	0.0809	{'rouge1': 0.9754024081831265, 'rouge2': 0.9579286248562431, 'rougeL': 0.9741313460430334, 'rougeLsum': 0.9751613463738352}
0.0482	3.9993	3996	0.0808	{'bleu': 0.9574684676357755, 'precisions': [0.9771731036056137, 0.9615197629595535, 0.950377700460969, 0.9411784261633503], 'brevity_penalty': 1.0, 'length_ratio': 1.0002669517979779, 'translation_length': 696941, 'reference_length': 696755}	{'rouge1': 0.9757795166466966, 'rouge2': 0.9586013928880327, 'rougeL': 0.9745320041915129, 'rougeLsum': 0.9755165129747526}	{'exact_match': 0.0}
0.0458	4.9986	4995	0.0847	{'bleu': 0.9570671016785056, 'precisions': [0.976919456982413, 0.9611588208136713, 0.9499201259530098, 0.9406510563080023], 'brevity_penalty': 1.0, 'length_ratio': 1.0003430187081541, 'translation_length': 696994, 'reference_length': 696755}	{'rouge1': 0.9755632054095464, 'rouge2': 0.9582426903380377, 'rougeL': 0.9743228923598912, 'rougeLsum': 0.9753134364311447}	{'exact_match': 0.0}

Framework versions

• PEFT 0.10.0
• Transformers 4.40.2
• Pytorch 2.2.0
• Datasets 2.19.1
• Tokenizers 0.19.1