smb-vision-large-1202

This model is trained from scratch using VideoMAE on over 55k CT volumes.

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: 3e-04
• train_batch_size: 16
• eval_batch_size: 1
• seed: 42
• 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
• num_epochs: 10.0

Training results

{ "_runtime": 2641.091489502, "_step": 399, "_timestamp": 1733187755.3146417, "_wandb.runtime": 2660, "train/epoch": 8.425414364640885, "train/global_step": 18300, "train/grad_norm": 0.04110511764883995, "train/learning_rate": 0.0001624558726951691, "train/loss": 0.4292 }

Framework versions

• Transformers 4.46.0
• Pytorch 2.5.0
• Datasets 3.0.2
• Tokenizers 0.20.1

How to use

# load data using `dataload.py`

model = VideoMAEForPreTraining.from_pretrained(
    standardmodelbio/smb-vision-large,
    trust_remote_code=True,
)

embedding = model.videomae(batch["image"])