Felldude/QWEN_32B_Comfy_MinimaxH3_Pruned_FP32

Qwen3-32B (QP)

QP (Quantization Prediction)

QP is theoretically improving 40-50% of the blocks on the trailing 16 Matnitsa bits when predicting from BF16 to FP32.

For the other 50-60% that it does not improve it should not degrade beyond the rounded vaules of BF16.

This was tested on Full FP32 trainings such as T5.

Citation

If you find our work helpful, feel free to give us a cite.

@misc{qwen3technicalreport,
      title={Qwen3 Technical Report}, 
      author={Qwen Team},
      year={2025},
      eprint={2505.09388},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2505.09388}, 
}