george2704/LLM_old

LLM

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

• Loss: 2.1056

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: 0.02
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2
• num_epochs: 27

Training results

Training Loss	Epoch	Step	Validation Loss
2.2807	0.92	6	2.1056
1.9558	2.0	13	2.1056
2.2821	2.92	19	2.1056
1.956	4.0	26	2.1056
2.2816	4.92	32	2.1056
1.9561	6.0	39	2.1056
2.2812	6.92	45	2.1056
1.9556	8.0	52	2.1056
2.2815	8.92	58	2.1056
1.9555	10.0	65	2.1056
2.2824	10.92	71	2.1056
1.9554	12.0	78	2.1056
2.2823	12.92	84	2.1056
1.9566	14.0	91	2.1056
2.2814	14.92	97	2.1056
1.956	16.0	104	2.1056
2.2829	16.92	110	2.1056
1.9559	18.0	117	2.1056
2.2821	18.92	123	2.1056
1.9556	20.0	130	2.1056
2.2817	20.92	136	2.1056
1.9559	22.0	143	2.1056
2.2816	22.92	149	2.1056
1.9561	24.0	156	2.1056
2.1055	24.92	162	2.1056

Framework versions

• PEFT 0.10.0
• Transformers 4.39.3
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2