Instructions to use stabilityai/stable-diffusion-3.5-medium-amdnpu with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use stabilityai/stable-diffusion-3.5-medium-amdnpu with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-3.5-medium-amdnpu", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
stable-diffusion-3.5-medium-amdnpu / normal /tea_caching /dynamic /dd /cache /pos_embedprojConv_meta.json
| { | |
| "dd_meta_major_version": 1, | |
| "dd_meta_minor_version": 4, | |
| "state_table_updates": [], | |
| "op_list": [ | |
| { | |
| "name": "/pos_embed/proj/Conv", | |
| "type": "SDConv", | |
| "in_args": [ | |
| "hidden_states_nhwc.out8_0_0" | |
| ], | |
| "const_args": [ | |
| "existing_model.pos_embed.proj.weight" | |
| ], | |
| "out_args": [ | |
| "/pos_embed/Transpose_output_0.out8_0_0" | |
| ], | |
| "attrs": { | |
| "auto_pad": { | |
| "type": "str", | |
| "value": [ | |
| "NOTSET" | |
| ] | |
| }, | |
| "dilations": { | |
| "type": "int", | |
| "value": [ | |
| "1", | |
| "1" | |
| ] | |
| }, | |
| "group": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| }, | |
| "kernel_shape": { | |
| "type": "int", | |
| "value": [ | |
| "2", | |
| "2" | |
| ] | |
| }, | |
| "pads": { | |
| "type": "int", | |
| "value": [ | |
| "0", | |
| "0", | |
| "0", | |
| "0" | |
| ] | |
| }, | |
| "strides": { | |
| "type": "int", | |
| "value": [ | |
| "2", | |
| "2" | |
| ] | |
| }, | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "w", | |
| "h", | |
| "16" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "floor(w/2)", | |
| "floor(h/2)", | |
| "1536" | |
| ] | |
| }, | |
| "weight_shape": { | |
| "type": "int", | |
| "value": [ | |
| "1536", | |
| "2", | |
| "2", | |
| "16" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16", | |
| "bfp16ebs8", | |
| "float" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/pos_embed/Add_2", | |
| "type": "SDAdd", | |
| "in_args": [ | |
| "/pos_embed/Transpose_output_0.out8_0_0", | |
| "/pos_embed/Reshape_1_output_0.out_13_1_2" | |
| ], | |
| "const_args": [], | |
| "out_args": [ | |
| "/pos_embed/Add_2_output_0.out_13_1_2" | |
| ], | |
| "attrs": { | |
| "a_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "state_dim1", | |
| "1536" | |
| ] | |
| }, | |
| "b_shape": { | |
| "type": "str", | |
| "value": [ | |
| "1", | |
| "state_dim1", | |
| "1536" | |
| ] | |
| }, | |
| "c_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "state_dim1", | |
| "1536" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16", | |
| "bfloat16" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "is_bias_add": { | |
| "type": "int", | |
| "value": [ | |
| "0" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/time_text_embed/timestep_embedder/linear_1/Gemm", | |
| "type": "SDGemm", | |
| "in_args": [ | |
| "/time_text_embed/Cast_output_0.out11_4_2" | |
| ], | |
| "const_args": [ | |
| "existing_model.time_text_embed.timestep_embedder.linear_1.weight_5_1_0_15_0_2" | |
| ], | |
| "out_args": [ | |
| "/time_text_embed/timestep_embedder/linear_1/Gemm_output_0.out11_4_2" | |
| ], | |
| "attrs": { | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1", | |
| "256" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1", | |
| "1536" | |
| ] | |
| }, | |
| "weight_shape": { | |
| "type": "int", | |
| "value": [ | |
| "256", | |
| "1536" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "bias_enable": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16", | |
| "bfp16ebs8", | |
| "bfloat16" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/time_text_embed/timestep_embedder/act/Sigmoid", | |
| "type": "SDSilu", | |
| "in_args": [ | |
| "/time_text_embed/timestep_embedder/linear_1/Gemm_output_0.out11_4_2" | |
| ], | |
| "const_args": [ | |
| "/time_text_embed/timestep_embedder/act/Sigmoid.weights5_1" | |
| ], | |
| "out_args": [ | |
| "/time_text_embed/timestep_embedder/act/Sigmoid_output_0.out5_1" | |
| ], | |
| "attrs": { | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1536" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1536" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "weight_shape": { | |
| "type": "int", | |
| "value": [ | |
| "128" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/time_text_embed/timestep_embedder/linear_2/Gemm", | |
| "type": "SDGemm", | |
| "in_args": [ | |
| "/time_text_embed/timestep_embedder/act/Sigmoid_output_0.out5_1" | |
| ], | |
| "const_args": [ | |
| "existing_model.time_text_embed.timestep_embedder.linear_2.weight_5_1_1_15_0_3" | |
| ], | |
| "out_args": [ | |
| "/time_text_embed/timestep_embedder/linear_2/Gemm_output_0.out11_4_3" | |
| ], | |
| "attrs": { | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1", | |
| "1536" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1", | |
| "1536" | |
| ] | |
| }, | |
| "weight_shape": { | |
| "type": "int", | |
| "value": [ | |
| "1536", | |
| "1536" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "bias_enable": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16", | |
| "bfp16ebs8", | |
| "bfloat16" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/time_text_embed/text_embedder/linear_1/Gemm", | |
| "type": "SDGemm", | |
| "in_args": [ | |
| "pooled_projections.out11_4_0" | |
| ], | |
| "const_args": [ | |
| "existing_model.time_text_embed.text_embedder.linear_1.weight_5_1_2_15_0_0" | |
| ], | |
| "out_args": [ | |
| "/time_text_embed/text_embedder/linear_1/Gemm_output_0.out11_4_0" | |
| ], | |
| "attrs": { | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1", | |
| "2048" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1", | |
| "1536" | |
| ] | |
| }, | |
| "weight_shape": { | |
| "type": "int", | |
| "value": [ | |
| "2048", | |
| "1536" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "bias_enable": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16", | |
| "bfp16ebs8", | |
| "bfloat16" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/time_text_embed/text_embedder/act_1/Sigmoid", | |
| "type": "SDSilu", | |
| "in_args": [ | |
| "/time_text_embed/text_embedder/linear_1/Gemm_output_0.out11_4_0" | |
| ], | |
| "const_args": [ | |
| "/time_text_embed/text_embedder/act_1/Sigmoid.weights5_0" | |
| ], | |
| "out_args": [ | |
| "/time_text_embed/text_embedder/act_1/Sigmoid_output_0.out5_0" | |
| ], | |
| "attrs": { | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1536" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1536" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "weight_shape": { | |
| "type": "int", | |
| "value": [ | |
| "128" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/time_text_embed/text_embedder/linear_2/Gemm", | |
| "type": "SDGemm", | |
| "in_args": [ | |
| "/time_text_embed/text_embedder/act_1/Sigmoid_output_0.out5_0" | |
| ], | |
| "const_args": [ | |
| "existing_model.time_text_embed.text_embedder.linear_2.weight_5_1_3_15_0_1" | |
| ], | |
| "out_args": [ | |
| "/time_text_embed/text_embedder/linear_2/Gemm_output_0.out11_4_1" | |
| ], | |
| "attrs": { | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1", | |
| "1536" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1", | |
| "1536" | |
| ] | |
| }, | |
| "weight_shape": { | |
| "type": "int", | |
| "value": [ | |
| "1536", | |
| "1536" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "bias_enable": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16", | |
| "bfp16ebs8", | |
| "bfloat16" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/time_text_embed/Add", | |
| "type": "SDAdd", | |
| "in_args": [ | |
| "/time_text_embed/timestep_embedder/linear_2/Gemm_output_0.out11_4_3", | |
| "/time_text_embed/text_embedder/linear_2/Gemm_output_0.out11_4_1" | |
| ], | |
| "const_args": [], | |
| "out_args": [ | |
| "/time_text_embed/Add_output_0.out_13_1_3" | |
| ], | |
| "attrs": { | |
| "a_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1536" | |
| ] | |
| }, | |
| "b_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1536" | |
| ] | |
| }, | |
| "c_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1536" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16", | |
| "bfloat16" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "is_bias_add": { | |
| "type": "int", | |
| "value": [ | |
| "0" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/transformer_blocks.0/norm1/norm/LayerNormalization", | |
| "type": "SDLayerNorm_bfbfp", | |
| "in_args": [ | |
| "/pos_embed/Add_2_output_0.out_13_1_2" | |
| ], | |
| "const_args": [ | |
| "/transformer_blocks.0/norm1/norm/Constant_output_0_9_0", | |
| "/transformer_blocks.0/norm1/norm/Constant_1_output_0_9_0" | |
| ], | |
| "out_args": [ | |
| "/transformer_blocks.0/norm1/norm/LayerNormalization_output_0.out9_0_bfp.out10_0" | |
| ], | |
| "attrs": { | |
| "axis": { | |
| "type": "int", | |
| "value": [ | |
| "-1" | |
| ] | |
| }, | |
| "epsilon": { | |
| "type": "float", | |
| "value": [ | |
| "9.999999974752427e-07" | |
| ] | |
| }, | |
| "stash_type": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| }, | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "state_dim1", | |
| "1536" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "state_dim1", | |
| "1536" | |
| ] | |
| }, | |
| "gamma_shape": { | |
| "type": "int", | |
| "value": [ | |
| "1536" | |
| ] | |
| }, | |
| "beta_shape": { | |
| "type": "int", | |
| "value": [ | |
| "1536" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16", | |
| "bfloat16" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfp16ebs8" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/transformer_blocks.0/norm1/silu/Sigmoid", | |
| "type": "SDSilu", | |
| "in_args": [ | |
| "/time_text_embed/Add_output_0.out_13_1_3" | |
| ], | |
| "const_args": [ | |
| "/transformer_blocks.0/norm1/silu/Sigmoid.weights5_2" | |
| ], | |
| "out_args": [ | |
| "/transformer_blocks.0/norm1/silu/Sigmoid_output_0.out5_2" | |
| ], | |
| "attrs": { | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1536" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1536" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "weight_shape": { | |
| "type": "int", | |
| "value": [ | |
| "128" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/transformer_blocks.0/norm1/Add_2", | |
| "type": "SDGemmGemmMulAdd_bfp", | |
| "in_args": [ | |
| "/transformer_blocks.0/norm1/silu/Sigmoid_output_0.out5_2", | |
| "/transformer_blocks.0/norm1/norm/LayerNormalization_output_0.out9_0_bfp.out10_0" | |
| ], | |
| "const_args": [ | |
| "existing_model.transformer_blocks.0.norm1.linear.weight_5_1_4_27_0_0_existing_model.transformer_blocks.0.norm1.linear.weight_5_1_4_27_0_1" | |
| ], | |
| "out_args": [ | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0_bfp.out4_0" | |
| ], | |
| "attrs": { | |
| "input_shape_0": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "1", | |
| "1536" | |
| ] | |
| }, | |
| "input_shape_1": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "state_dim1", | |
| "1536" | |
| ] | |
| }, | |
| "weight_shape_0": { | |
| "type": "int", | |
| "value": [ | |
| "1536", | |
| "1536" | |
| ] | |
| }, | |
| "weight_shape_1": { | |
| "type": "int", | |
| "value": [ | |
| "1536", | |
| "1536" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "state_dim1", | |
| "1536" | |
| ] | |
| }, | |
| "bias_enable": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16", | |
| "bfp16ebs8", | |
| "bfp16ebs8" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfp16ebs8" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "name": "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0_bfp.out4_0_SDCastBfp2Bf", | |
| "type": "SDCastBfp2Bf", | |
| "in_args": [ | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0_bfp.out4_0" | |
| ], | |
| "const_args": [ | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0_bfp.out4_0_bfp.wts" | |
| ], | |
| "out_args": [ | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0" | |
| ], | |
| "attrs": { | |
| "input_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "state_dim1", | |
| "1536" | |
| ] | |
| }, | |
| "output_shape": { | |
| "type": "str", | |
| "value": [ | |
| "batch_2", | |
| "state_dim1", | |
| "1536" | |
| ] | |
| }, | |
| "in_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfp16ebs8" | |
| ] | |
| }, | |
| "out_dtypes": { | |
| "type": "str", | |
| "value": [ | |
| "bfloat16" | |
| ] | |
| }, | |
| "ctrl_packet": { | |
| "type": "int", | |
| "value": [ | |
| "1" | |
| ] | |
| } | |
| } | |
| } | |
| ], | |
| "fused_tensors": { | |
| "in": { | |
| "buffer_size": 7712, | |
| "xrt_arg_id": 0, | |
| "packed_tensors": [ | |
| "hidden_states_nhwc.out8_0_0", | |
| "/pos_embed/Reshape_1_output_0.out_13_1_2", | |
| "/time_text_embed/Cast_output_0.out11_4_2", | |
| "pooled_projections.out11_4_0" | |
| ] | |
| }, | |
| "out": { | |
| "buffer_size": 3072, | |
| "xrt_arg_id": 1, | |
| "packed_tensors": [ | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0" | |
| ] | |
| }, | |
| "scratch": { | |
| "buffer_size": 33792, | |
| "xrt_arg_id": 2, | |
| "packed_tensors": [ | |
| "/pos_embed/Transpose_output_0.out8_0_0", | |
| "/pos_embed/Add_2_output_0.out_13_1_2", | |
| "/time_text_embed/timestep_embedder/linear_1/Gemm_output_0.out11_4_2", | |
| "/time_text_embed/timestep_embedder/act/Sigmoid_output_0.out5_1", | |
| "/time_text_embed/timestep_embedder/linear_2/Gemm_output_0.out11_4_3", | |
| "/time_text_embed/text_embedder/linear_1/Gemm_output_0.out11_4_0", | |
| "/time_text_embed/text_embedder/act_1/Sigmoid_output_0.out5_0", | |
| "/time_text_embed/text_embedder/linear_2/Gemm_output_0.out11_4_1", | |
| "/time_text_embed/Add_output_0.out_13_1_3", | |
| "/transformer_blocks.0/norm1/norm/LayerNormalization_output_0.out9_0_bfp.out10_0", | |
| "/transformer_blocks.0/norm1/silu/Sigmoid_output_0.out5_2", | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0_bfp.out4_0" | |
| ] | |
| }, | |
| "const": { | |
| "buffer_size": 14943104, | |
| "xrt_arg_id": 3, | |
| "packed_tensors": [ | |
| "existing_model.pos_embed.proj.weight", | |
| "existing_model.time_text_embed.timestep_embedder.linear_1.weight_5_1_0_15_0_2", | |
| "/time_text_embed/timestep_embedder/act/Sigmoid.weights5_1", | |
| "existing_model.time_text_embed.timestep_embedder.linear_2.weight_5_1_1_15_0_3", | |
| "existing_model.time_text_embed.text_embedder.linear_1.weight_5_1_2_15_0_0", | |
| "/time_text_embed/text_embedder/act_1/Sigmoid.weights5_0", | |
| "existing_model.time_text_embed.text_embedder.linear_2.weight_5_1_3_15_0_1", | |
| "/transformer_blocks.0/norm1/norm/Constant_output_0_9_0", | |
| "/transformer_blocks.0/norm1/norm/Constant_1_output_0_9_0", | |
| "/transformer_blocks.0/norm1/silu/Sigmoid.weights5_2", | |
| "existing_model.transformer_blocks.0.norm1.linear.weight_5_1_4_27_0_0_existing_model.transformer_blocks.0.norm1.linear.weight_5_1_4_27_0_1", | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0_bfp.out4_0_bfp.wts" | |
| ] | |
| }, | |
| "super_instr": { | |
| "buffer_size": 0, | |
| "xrt_arg_id": 4, | |
| "packed_tensors": [] | |
| } | |
| }, | |
| "tensor_map": { | |
| "hidden_states_nhwc.out8_0_0": { | |
| "packed_buffer_label": "in", | |
| "xrt_arg_id": 0, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1, | |
| 1, | |
| 16 | |
| ], | |
| "size_in_bytes": 32, | |
| "op_tensor_size": 32, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "w", | |
| "h", | |
| "False" | |
| ], | |
| "offset": 0 | |
| }, | |
| "/pos_embed/Reshape_1_output_0.out_13_1_2": { | |
| "packed_buffer_label": "in", | |
| "xrt_arg_id": 0, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "False", | |
| "state_dim1", | |
| "False" | |
| ], | |
| "offset": 32 | |
| }, | |
| "/time_text_embed/Cast_output_0.out11_4_2": { | |
| "packed_buffer_label": "in", | |
| "xrt_arg_id": 0, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 256 | |
| ], | |
| "size_in_bytes": 512, | |
| "op_tensor_size": 512, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 3104 | |
| }, | |
| "pooled_projections.out11_4_0": { | |
| "packed_buffer_label": "in", | |
| "xrt_arg_id": 0, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 2048 | |
| ], | |
| "size_in_bytes": 4096, | |
| "op_tensor_size": 4096, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 3616 | |
| }, | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0": { | |
| "packed_buffer_label": "out", | |
| "xrt_arg_id": 1, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "state_dim1", | |
| "False" | |
| ], | |
| "offset": 0 | |
| }, | |
| "/pos_embed/Transpose_output_0.out8_0_0": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "state_dim1", | |
| "False" | |
| ], | |
| "offset": 0 | |
| }, | |
| "/pos_embed/Add_2_output_0.out_13_1_2": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "state_dim1", | |
| "False" | |
| ], | |
| "offset": 3072 | |
| }, | |
| "/time_text_embed/timestep_embedder/linear_1/Gemm_output_0.out11_4_2": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 6144 | |
| }, | |
| "/time_text_embed/timestep_embedder/act/Sigmoid_output_0.out5_1": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 9216 | |
| }, | |
| "/time_text_embed/timestep_embedder/linear_2/Gemm_output_0.out11_4_3": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 12288 | |
| }, | |
| "/time_text_embed/text_embedder/linear_1/Gemm_output_0.out11_4_0": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 15360 | |
| }, | |
| "/time_text_embed/text_embedder/act_1/Sigmoid_output_0.out5_0": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 18432 | |
| }, | |
| "/time_text_embed/text_embedder/linear_2/Gemm_output_0.out11_4_1": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 21504 | |
| }, | |
| "/time_text_embed/Add_output_0.out_13_1_3": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 24576 | |
| }, | |
| "/transformer_blocks.0/norm1/norm/LayerNormalization_output_0.out9_0_bfp.out10_0": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "uint8", | |
| "shape": [ | |
| 1, | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 1536, | |
| "op_tensor_size": 1536, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "state_dim1", | |
| "False" | |
| ], | |
| "offset": 27648 | |
| }, | |
| "/transformer_blocks.0/norm1/silu/Sigmoid_output_0.out5_2": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "False" | |
| ], | |
| "offset": 29184 | |
| }, | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0_bfp.out4_0": { | |
| "packed_buffer_label": "scratch", | |
| "xrt_arg_id": 2, | |
| "dtype": "uint8", | |
| "shape": [ | |
| 1, | |
| 1, | |
| 1536 | |
| ], | |
| "size_in_bytes": 1536, | |
| "op_tensor_size": 1536, | |
| "dynamic_shapes": [ | |
| "batch_2", | |
| "state_dim1", | |
| "False" | |
| ], | |
| "offset": 32256 | |
| }, | |
| "existing_model.pos_embed.proj.weight": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfp16ebs8", | |
| "shape": [ | |
| 135168 | |
| ], | |
| "size_in_bytes": 135168, | |
| "op_tensor_size": 135168, | |
| "offset": 0, | |
| "file_name": "cache/pos_embedprojConv_0.const", | |
| "file_size": 135168 | |
| }, | |
| "existing_model.time_text_embed.timestep_embedder.linear_1.weight_5_1_0_15_0_2": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfp16ebs8", | |
| "shape": [ | |
| 448512 | |
| ], | |
| "size_in_bytes": 448512, | |
| "op_tensor_size": 448512, | |
| "offset": 135168, | |
| "file_name": "cache/pos_embedprojConv_1.const", | |
| "file_size": 448512 | |
| }, | |
| "/time_text_embed/timestep_embedder/act/Sigmoid.weights5_1": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 128 | |
| ], | |
| "size_in_bytes": 256, | |
| "op_tensor_size": 256, | |
| "offset": 583680, | |
| "file_name": "cache/pos_embedprojConv_2.const", | |
| "file_size": 256 | |
| }, | |
| "existing_model.time_text_embed.timestep_embedder.linear_2.weight_5_1_1_15_0_3": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfp16ebs8", | |
| "shape": [ | |
| 2691072 | |
| ], | |
| "size_in_bytes": 2691072, | |
| "op_tensor_size": 2691072, | |
| "offset": 583936, | |
| "file_name": "cache/pos_embedprojConv_3.const", | |
| "file_size": 2691072 | |
| }, | |
| "existing_model.time_text_embed.text_embedder.linear_1.weight_5_1_2_15_0_0": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfp16ebs8", | |
| "shape": [ | |
| 3588096 | |
| ], | |
| "size_in_bytes": 3588096, | |
| "op_tensor_size": 3588096, | |
| "offset": 3275008, | |
| "file_name": "cache/pos_embedprojConv_4.const", | |
| "file_size": 3588096 | |
| }, | |
| "/time_text_embed/text_embedder/act_1/Sigmoid.weights5_0": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 128 | |
| ], | |
| "size_in_bytes": 256, | |
| "op_tensor_size": 256, | |
| "offset": 6863104, | |
| "file_name": "cache/pos_embedprojConv_5.const", | |
| "file_size": 256 | |
| }, | |
| "existing_model.time_text_embed.text_embedder.linear_2.weight_5_1_3_15_0_1": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfp16ebs8", | |
| "shape": [ | |
| 2691072 | |
| ], | |
| "size_in_bytes": 2691072, | |
| "op_tensor_size": 2691072, | |
| "offset": 6863360, | |
| "file_name": "cache/pos_embedprojConv_6.const", | |
| "file_size": 2691072 | |
| }, | |
| "/transformer_blocks.0/norm1/norm/Constant_output_0_9_0": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "offset": 9554432, | |
| "file_name": "cache/pos_embedprojConv_7.const", | |
| "file_size": 3072 | |
| }, | |
| "/transformer_blocks.0/norm1/norm/Constant_1_output_0_9_0": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 1536 | |
| ], | |
| "size_in_bytes": 3072, | |
| "op_tensor_size": 3072, | |
| "offset": 9557504, | |
| "file_name": "cache/pos_embedprojConv_8.const", | |
| "file_size": 3072 | |
| }, | |
| "/transformer_blocks.0/norm1/silu/Sigmoid.weights5_2": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 128 | |
| ], | |
| "size_in_bytes": 256, | |
| "op_tensor_size": 256, | |
| "offset": 9560576, | |
| "file_name": "cache/pos_embedprojConv_9.const", | |
| "file_size": 256 | |
| }, | |
| "existing_model.transformer_blocks.0.norm1.linear.weight_5_1_4_27_0_0_existing_model.transformer_blocks.0.norm1.linear.weight_5_1_4_27_0_1": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "uint8", | |
| "shape": [ | |
| 5382144 | |
| ], | |
| "size_in_bytes": 5382144, | |
| "op_tensor_size": 5382144, | |
| "offset": 9560832, | |
| "file_name": "cache/pos_embedprojConv_10.const", | |
| "file_size": 5382144 | |
| }, | |
| "/transformer_blocks.0/norm1/Add_2_output_0.out3_0_0_bfp.out4_0_bfp.wts": { | |
| "packed_buffer_label": "const", | |
| "xrt_arg_id": 3, | |
| "dtype": "bfloat16", | |
| "shape": [ | |
| 64 | |
| ], | |
| "size_in_bytes": 128, | |
| "op_tensor_size": 128, | |
| "offset": 14942976, | |
| "file_name": "cache/pos_embedprojConv_11.const", | |
| "file_size": 128 | |
| } | |
| }, | |
| "dynamic_shape_subgraph": true, | |
| "dynamic_shape_list": [ | |
| { | |
| "floor(w/2)": 32, | |
| "h": 64, | |
| "state_dim1": 1024, | |
| "w": 64, | |
| "batch_2": 2, | |
| "floor(h/2)": 32 | |
| }, | |
| { | |
| "floor(w/2)": 40, | |
| "h": 80, | |
| "state_dim1": 1600, | |
| "w": 80, | |
| "batch_2": 2, | |
| "floor(h/2)": 40 | |
| }, | |
| { | |
| "floor(w/2)": 48, | |
| "h": 96, | |
| "state_dim1": 2304, | |
| "w": 96, | |
| "batch_2": 2, | |
| "floor(h/2)": 48 | |
| }, | |
| { | |
| "floor(w/2)": 64, | |
| "h": 128, | |
| "state_dim1": 4096, | |
| "w": 128, | |
| "batch_2": 2, | |
| "floor(h/2)": 64 | |
| }, | |
| { | |
| "floor(w/2)": 24, | |
| "h": 64, | |
| "state_dim1": 768, | |
| "w": 48, | |
| "batch_2": 2, | |
| "floor(h/2)": 32 | |
| }, | |
| { | |
| "floor(w/2)": 30, | |
| "h": 80, | |
| "state_dim1": 1200, | |
| "w": 60, | |
| "batch_2": 2, | |
| "floor(h/2)": 40 | |
| }, | |
| { | |
| "floor(w/2)": 36, | |
| "h": 96, | |
| "state_dim1": 1728, | |
| "w": 72, | |
| "batch_2": 2, | |
| "floor(h/2)": 48 | |
| }, | |
| { | |
| "floor(w/2)": 48, | |
| "h": 128, | |
| "state_dim1": 3072, | |
| "w": 96, | |
| "batch_2": 2, | |
| "floor(h/2)": 64 | |
| }, | |
| { | |
| "floor(w/2)": 32, | |
| "h": 48, | |
| "state_dim1": 768, | |
| "w": 64, | |
| "batch_2": 2, | |
| "floor(h/2)": 24 | |
| }, | |
| { | |
| "floor(w/2)": 40, | |
| "h": 60, | |
| "state_dim1": 1200, | |
| "w": 80, | |
| "batch_2": 2, | |
| "floor(h/2)": 30 | |
| }, | |
| { | |
| "floor(w/2)": 48, | |
| "h": 72, | |
| "state_dim1": 1728, | |
| "w": 96, | |
| "batch_2": 2, | |
| "floor(h/2)": 36 | |
| }, | |
| { | |
| "floor(w/2)": 64, | |
| "h": 96, | |
| "state_dim1": 3072, | |
| "w": 128, | |
| "batch_2": 2, | |
| "floor(h/2)": 48 | |
| }, | |
| { | |
| "floor(w/2)": 18, | |
| "h": 64, | |
| "state_dim1": 576, | |
| "w": 36, | |
| "batch_2": 2, | |
| "floor(h/2)": 32 | |
| }, | |
| { | |
| "floor(w/2)": 24, | |
| "h": 80, | |
| "state_dim1": 960, | |
| "w": 48, | |
| "batch_2": 2, | |
| "floor(h/2)": 40 | |
| }, | |
| { | |
| "floor(w/2)": 28, | |
| "h": 96, | |
| "state_dim1": 1344, | |
| "w": 56, | |
| "batch_2": 2, | |
| "floor(h/2)": 48 | |
| }, | |
| { | |
| "floor(w/2)": 36, | |
| "h": 128, | |
| "state_dim1": 2304, | |
| "w": 72, | |
| "batch_2": 2, | |
| "floor(h/2)": 64 | |
| }, | |
| { | |
| "floor(w/2)": 32, | |
| "h": 36, | |
| "state_dim1": 576, | |
| "w": 64, | |
| "batch_2": 2, | |
| "floor(h/2)": 18 | |
| }, | |
| { | |
| "floor(w/2)": 40, | |
| "h": 48, | |
| "state_dim1": 960, | |
| "w": 80, | |
| "batch_2": 2, | |
| "floor(h/2)": 24 | |
| }, | |
| { | |
| "floor(w/2)": 48, | |
| "h": 56, | |
| "state_dim1": 1344, | |
| "w": 96, | |
| "batch_2": 2, | |
| "floor(h/2)": 28 | |
| }, | |
| { | |
| "floor(w/2)": 64, | |
| "h": 72, | |
| "state_dim1": 2304, | |
| "w": 128, | |
| "batch_2": 2, | |
| "floor(h/2)": 36 | |
| } | |
| ], | |
| "aux_info": {} | |
| } |