ONNX Model Zoo

amd/stable-diffusion-3.5-medium-amdnpu-onnx

By amd · ↓ 0 · ♥ 0 · Updated 2026-03-18

What's in this repo

This page lists the files in amd/stable-diffusion-3.5-medium-amdnpu-onnx, including model.onnx, replaced.onnx and more.

  • 9 ONNX models (.onnx files)
  • 117 files, including 24 .ctrlpkt, 24 .meta, 24 .super, 19 .json, 9 .onnx, 3 .fconst
  • Last updated 2026-03-18

ONNX files

File Size Download
common/text_encoder/model.onnx — Download
common/text_encoder_2/model.onnx — Download
common/text_encoder_3/model.onnx — Download
common/tokenizer/model.onnx — Download
common/tokenizer_2/model.onnx — Download
common/tokenizer_3/model.onnx — Download
common/vae_decoder/dynamic/dd/replaced.onnx — Download
common/vae_encoder/dynamic/dd/replaced.onnx — Download
normal/transformer/dynamic/dd/replaced.onnx — Download

How to use

These ONNX files can be loaded with ONNX Runtime. To inspect a model graph, upload the file to the Netron viewer. Download links point to the original files on Hugging Face.

README


license: other license_name: stabilityai-ai-community license_link: LICENSE.md tags:

  • text-to-image
  • stable-diffusion
  • diffusers inference: true extra_gated_prompt: >- By clicking "Agree", you agree to the License Agreement and acknowledge Stability AI's Privacy Policy. extra_gated_fields: Name: text Email: text Country: country Organization or Affiliation: text Receive email updates and promotions on Stability AI products, services, and research?: type: select options:
    • 'Yes'
    • 'No' What do you intend to use the model for?: type: select options:
    • Research
    • Personal use
    • Creative Professional
    • Startup
    • Enterprise I agree to the License Agreement and acknowledge Stability AI's Privacy Policy: checkbox

language:

  • en pipeline_tag: text-to-image

🚀 Stable Diffusion 3.5 Medium on AMD NPU

"Stable Diffusion 3.5 Medium is an improved version of SD3 Medium, featuring better prompt adherence, better generation quality, and more diverse outputs. SD3.5 Medium is a Multimodal Diffusion Transformer (MMDiT) text-to-image model that features improved image quality, typography, and prompt understanding." (stabilityai/stable-diffusion-3.5-medium)

This version of the model has been optimized to run on AMD NPUs.

For more information, refer to the original model card under stabilityai/stable-diffusion-3.5-medium.

Model Details

Model Details Description
Person or organization developing model AMD
Original model authors Stability AI
Model date February 2026
Model version 1.7.1
Model type Diffusion-based text-to-image generation model
Information about training algorithms, parameters, fairness constraints or other applied approaches, and features This is a model that can be used to generate and modify images based on text prompts.
License Stability AI Community
Where to send questions or comments about the model Community Tab and AMD Developer Community Discord

⚡ Intended Use

Getting Started

To get started with this model, visit github.com/amd/sd-sandbox.

⚓ Ethical Considerations

AMD is committed to conducting our business in a fair, ethical and honest manner and in compliance with all applicable laws, rules and regulations. You can find out more at the AMD Ethics and Compliance page.

⚠️ Caveats and Recommendations

Please visit the original model card for more details: stabilityai/stable-diffusion-3.5-medium.

Model info

text-to-image diffusers
diffusers onnx text-to-image stable-diffusion en license:other region:us
↓ 0 ♥ 0

Created: 2026-03-18

Updated: 2026-03-18

View on Hugging Face

Files (117)

.gitattributes
LICENSE
README.md
common/scheduler/scheduler_config.json
common/text_encoder/config.json
common/text_encoder/model.onnx ONNX
common/text_encoder_2/config.json
common/text_encoder_2/model.onnx ONNX
common/text_encoder_2/model.onnx_data
common/text_encoder_3/config.json
common/text_encoder_3/model.onnx ONNX
common/text_encoder_3/model.onnx_data
common/text_encoder_3_gptq_v2/serialized_quantized_t5-v1_1-xxl_w4_g128_gptq.bin
common/text_encoder_3_gptq_v2/serialized_quantized_t5-v1_1-xxl_w4_g128_gptq.pth
common/tokenizer/merges.txt
common/tokenizer/model.onnx ONNX
common/tokenizer/special_tokens_map.json
common/tokenizer/tokenizer_config.json
common/tokenizer/vocab.json
common/tokenizer_2/merges.txt
common/tokenizer_2/model.onnx ONNX
common/tokenizer_2/special_tokens_map.json
common/tokenizer_2/tokenizer_config.json
common/tokenizer_2/vocab.json
common/tokenizer_3/model.onnx ONNX
common/tokenizer_3/special_tokens_map.json
common/tokenizer_3/spiece.model
common/tokenizer_3/tokenizer.json
common/tokenizer_3/tokenizer_config.json
common/vae_decoder/config.json
common/vae_decoder/dynamic/dd/cache/NhwcConv_0-decoderconv_inConv_meta.json
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv.fconst
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv.state
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_128_width_128_.ctrlpkt
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_128_width_128_.meta
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_128_width_128_.super
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_128_width_72_.ctrlpkt
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_128_width_72_.meta
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_128_width_72_.super
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_128_width_96_.ctrlpkt
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_128_width_96_.meta
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_128_width_96_.super
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_64_width_64_.ctrlpkt
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_64_width_64_.meta
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_64_width_64_.super
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_64_width_96_.ctrlpkt
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_64_width_96_.meta
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_64_width_96_.super
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_72_width_128_.ctrlpkt
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_72_width_128_.meta
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_72_width_128_.super
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_96_width_128_.ctrlpkt
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_96_width_128_.meta
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_96_width_128_.super
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_96_width_64_.ctrlpkt
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_96_width_64_.meta
common/vae_decoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-decoderconv_inConv_height_96_width_64_.super
common/vae_decoder/dynamic/dd/replaced.onnx ONNX
common/vae_encoder/config.json
common/vae_encoder/dynamic/dd/cache/NhwcConv_0-encoderconv_inConv_meta.json
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv.fconst
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv.state
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_1024_width_1024_.ctrlpkt
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_1024_width_1024_.meta
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_1024_width_1024_.super
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_1024_width_576_.ctrlpkt
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_1024_width_576_.meta
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_1024_width_576_.super
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_1024_width_768_.ctrlpkt
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_1024_width_768_.meta
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_1024_width_768_.super
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_512_width_512_.ctrlpkt
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_512_width_512_.meta
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_512_width_512_.super
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_512_width_768_.ctrlpkt
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_512_width_768_.meta
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_512_width_768_.super
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_576_width_1024_.ctrlpkt
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_576_width_1024_.meta
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_576_width_1024_.super
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_768_width_1024_.ctrlpkt
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_768_width_1024_.meta
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_768_width_1024_.super
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_768_width_512_.ctrlpkt
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_768_width_512_.meta
common/vae_encoder/dynamic/dd/dd_metastate_Sd30_VAE_NhwcConv_0-encoderconv_inConv_height_768_width_512_.super
common/vae_encoder/dynamic/dd/replaced.onnx ONNX
normal/transformer/config.json
normal/transformer/dynamic/dd/cache/pos_embedprojConv_meta.json
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv.fconst
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv.state
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_128_h_128_state_dim1_4096_max_length_160_.ctrlpkt
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_128_h_128_state_dim1_4096_max_length_160_.meta
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_128_h_128_state_dim1_4096_max_length_160_.super
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_128_h_72_state_dim1_2304_max_length_160_.ctrlpkt
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_128_h_72_state_dim1_2304_max_length_160_.meta
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_128_h_72_state_dim1_2304_max_length_160_.super
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_128_h_96_state_dim1_3072_max_length_160_.ctrlpkt
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_128_h_96_state_dim1_3072_max_length_160_.meta
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_128_h_96_state_dim1_3072_max_length_160_.super
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_64_h_64_state_dim1_1024_max_length_160_.ctrlpkt
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_64_h_64_state_dim1_1024_max_length_160_.meta
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_64_h_64_state_dim1_1024_max_length_160_.super
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_64_h_96_state_dim1_1536_max_length_160_.ctrlpkt
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_64_h_96_state_dim1_1536_max_length_160_.meta
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_64_h_96_state_dim1_1536_max_length_160_.super
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_72_h_128_state_dim1_2304_max_length_160_.ctrlpkt
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_72_h_128_state_dim1_2304_max_length_160_.meta
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_72_h_128_state_dim1_2304_max_length_160_.super
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_96_h_128_state_dim1_3072_max_length_160_.ctrlpkt
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_96_h_128_state_dim1_3072_max_length_160_.meta
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_96_h_128_state_dim1_3072_max_length_160_.super
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_96_h_64_state_dim1_1536_max_length_160_.ctrlpkt
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_96_h_64_state_dim1_1536_max_length_160_.meta
normal/transformer/dynamic/dd/dd_metastate_Sd30_Mmdit_pos_embedprojConv_batch_2_2_w_96_h_64_state_dim1_1536_max_length_160_.super
normal/transformer/dynamic/dd/replaced.onnx ONNX
normal/transformer/dynamic/dd/replaced.onnx.data

Related models