mistral-7b-instruct-lora-v3.0

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.2 on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 2.2224

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

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	4	2.3675
No log	2.0	8	2.3372
2.3612	3.0	12	2.3082
2.3612	4.0	16	2.2819
2.296	5.0	20	2.2598
2.296	6.0	24	2.2430
2.296	7.0	28	2.2317
2.2141	8.0	32	2.2254
2.2141	9.0	36	2.2229
2.252	10.0	40	2.2224

Framework versions

• PEFT 0.11.1
• Transformers 4.43.0.dev0
• Pytorch 2.1.0+cu118
• Datasets 2.20.0
• Tokenizers 0.19.1