---
library_name: transformers
license: apache-2.0
base_model: Qwen/Qwen3.5-35B-A3B
tags:
- llama-factory
- full
- generated_from_trainer
model-index:
- name: Kallisti-35B-A3B
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->


# 🍎 Kallisti-35B-A3B

This model is a fine-tuned version of [Qwen/Qwen3.5-35B-A3B](https://huggingface.co/Qwen/Qwen3.5-35B-A3B).

- 🌐 [Project Website](https://vab.bakelab.ai/) - Learn more about Visual Aesthetic Benchmark
- 📖 [Technical Report](https://arxiv.org/abs/2605.12684) - Discover technical details behind VAB
- 🔧 [GitHub Repo](https://github.com/BakeLab/Visual-Aesthetic-Benchmark) - Evaluation scripts and benchmark tooling
- 🤗 [Visual Aesthetic Benchmark](https://huggingface.co/datasets/BakeLab/Visual-Aesthetic-Benchmark) - HF Datasets
- 🤗 [Kallisti-35B-A3B](https://huggingface.co/BakeLab/Kallisti-35B-A3B) - Finetune model [📍| You are here!]

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 1e-06
- train_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10.0
- num_epochs: 3.0

### Framework versions

- Transformers 5.2.0
- Pytorch 2.10.0+cu128
- Datasets 4.0.0
- Tokenizers 0.22.2

## License

[Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)

## Contact

Please contact [Yichen](mailto:yfeng42@uw.edu) by email.