This model is a fine-tuned version of llava-hf/llava-v1.6-mistral-7b-hf on the AA_preference_random_0_20 dataset. It achieves the following results on the evaluation set:

Loss: 0.6312
Rewards/chosen: 1.1237
Rewards/rejected: -0.0848
Rewards/accuracies: 0.75
Rewards/margins: 1.2086
Logps/rejected: -204.8366
Logps/chosen: -232.7576
Logits/rejected: -2.2673
Logits/chosen: -2.3213

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-06
train_batch_size: 8
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 4
total_train_batch_size: 256
total_eval_batch_size: 64
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 10
num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.219	1.8692	50	0.6226	1.0140	-0.1156	0.7396	1.1296	-205.1440	-233.8548	-2.3453	-2.3944

Framework versions

Transformers 4.45.2
Pytorch 2.4.0+cu121
Datasets 2.21.0
Tokenizers 0.20.3

htlou/mm-interp-AA_preference_random_0_20

AA_preference_random_0_20

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

htlou/
mm-interp-AA_preference_random_0_20