kiritoxkiriko/llava-1.5-7b-hf-100k

llava-1.5-7b-hf-finetuned-100k

This model is a fine-tuned version of llava-hf/llava-1.5-7b-hf on the first 100k rows of theblackcat102/llava-instruct-mix dataset. It achieves the following results on the evaluation set:

• Loss: 0.6758

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: 5e-05
• train_batch_size: 24
• eval_batch_size: 8
• seed: 666
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 192
• total_eval_batch_size: 32
• 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: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.6707	2.1279	500	0.6810

Framework versions

• PEFT 0.15.1
• Transformers 4.51.3
• Pytorch 2.6.0a0+df5bbc09d1.nv24.12
• Datasets 3.5.0
• Tokenizers 0.21.1