Configuration Parsing Warning:In UNKNOWN_FILENAME: "quantization_config.config_groups.group_0.format" must be a string

Model Details

This model is an MXFP8 (W8A8, group_size 32) model of tencent/Hy4-flash-FP8 generated by intel/auto-round using RTN. Please follow the license of the original model.

  • Quantized (MXFP8): routed experts, attention query up-projection, attention gate and output projections, DFly draft layers
  • Kept in BF16: dense layer, attention latent projections, shared experts, expert latent projections, router, sparse-attention indexer, Engram embeddings, vision tower, embeddings, lm_head

Generate the Model

git clone https://github.com/intel/auto-round.git && cd auto-round
git checkout 1586bc6f64b4702981ff249810971c347fdc67ad
curl -L https://huggingface.co/intel-ai/Hy4-flash-FP8-MXFP8-CT-AutoRound/resolve/main/autoround_hy4flash_fused_scale_inv.patch | git apply
uv pip install -e . compressed-tensors
uv pip install git+https://github.com/ManaEstras/transformers.git@c2ec850510a37978ea83df226589623d5f2be776
auto-round \
    --model_name tencent/Hy4-flash-FP8 \
    --scheme MXFP8 \
    --model_free \
    --ignore_layers q_a_proj,kv_a_proj_with_mqa,kv_b_proj,model.layers.0.self_attn,model.layers.0.mlp,shared_experts,routed_expert_down_proj,routed_expert_up_proj,compressed_indexer,engram_layer,vit.,hidden_correction,mtp_fusion,confidence_head,lm_head \
    --format llm_compressor \
    --output_dir Hy4-flash-FP8-MXFP8-CT-AutoRound

Ethical Considerations and Limitations

The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs. Therefore, before deploying any applications of the model, developers should perform safety testing.

Caveats and Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Here are a couple of useful links to learn more about Intel's AI software:

Disclaimer

The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.

Cite

@article{cheng2023optimize,
  title={Optimize weight rounding via signed gradient descent for the quantization of llms},
  author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi},
  journal={arXiv preprint arXiv:2309.05516},
  year={2023}
}

arxiv github

Signed-off-by: Sun, Xuehao xuehao.sun@intel.com

Downloads last month
100
Safetensors
Model size
220B params
Tensor type
F32
·
BF16
·
I32
·
F8_E4M3
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for intel-ai/Hy4-flash-FP8-MXFP8-CT-AutoRound

Quantized
(2)
this model

Paper for intel-ai/Hy4-flash-FP8-MXFP8-CT-AutoRound