mistral_7b_0_3-summarize-gpt4o-128k

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

• Loss: 2.0012

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: 0.0002
• train_batch_size: 4
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• total_eval_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
0.6982	0.9980	245	1.8248
0.6596	2.0	491	1.8338
0.6197	2.9980	736	1.8432
0.6011	4.0	982	1.8707
0.5805	4.9980	1227	1.9009
0.5585	6.0	1473	1.9298
0.5413	6.9980	1718	1.9540
0.5295	8.0	1964	1.9814
0.5154	8.9980	2209	1.9979
0.508	9.9796	2450	2.0012

Framework versions

• PEFT 0.12.0
• Transformers 4.44.0
• Pytorch 2.2.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1