[INFO] QAIRT_SDK_ROOT=/home/boyucwsl/executorch/qairt/2.37.0.250724
[WARN] QNN_SDK_ROOT/SNPE_ROOT set to QAIRT_SDK_ROOT for backwards compatibility and will be deprecated in a future release.
[INFO] QAIRT SDK environment setup complete
[QNN] Using QNN SDK at /home/boyucwsl/executorch/qairt/2.37.0.250724 (from QNN_SDK_ROOT)
[PipelineGraphCollector] Installed patch: prepare_pt2e
[PipelineGraphCollector] Installed patch: convert_pt2e
[PipelineGraphCollector] Installed patch: to_edge_transform_and_lower
[PipelineGraphCollector] Installed ETRecord patches
libquantized_ops_aot_lib is not loaded
The backend_opinfo module couldn't be imported, so the abstract implementation will be used instead. This might be because $QNN_SDK_ROOT/lib/python isn't included in your PYTHONPATH, or the `BackendOpInfo` API isn't available in your QNN SDK version. Note that the `BackendOpInfo` API is supported starting from QNN SDK 2.41 and above.
[PipelineGraphCollector] Installed QNN patch: _ptq_calibrate
[AccuracyLens] Installed patch: get_imagenet_dataset
[AccuracyLens] Installed patch: get_masked_language_model_dataset
[Observatory CLI] Running as module: examples.qualcomm.oss_scripts.distilbert
PyTorch version 2.11.0+cpu available.
Using parser's config
HTTP Request: HEAD https://huggingface.co/distilbert/distilbert-base-uncased/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect"
Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads.
Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads.
HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/distilbert/distilbert-base-uncased/12040accade4e8a0f71eabdb258fecc2e7e948be/tokenizer_config.json "HTTP/1.1 200 OK"
HTTP Request: GET https://huggingface.co/api/models/distilbert/distilbert-base-uncased/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
HTTP Request: GET https://huggingface.co/api/models/distilbert/distilbert-base-uncased/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK"
HTTP Request: HEAD https://huggingface.co/distilbert/distilbert-base-uncased/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/distilbert/distilbert-base-uncased/12040accade4e8a0f71eabdb258fecc2e7e948be/config.json "HTTP/1.1 200 OK"

Loading weights:   0%|          | 0/105 [00:00<?, ?it/s]
Loading weights:   1%|          | 1/105 [00:00<00:00, 6853.44it/s, Materializing param=distilbert.embeddings.LayerNorm.bias]
Loading weights:   1%|          | 1/105 [00:00<00:00, 1847.71it/s, Materializing param=distilbert.embeddings.LayerNorm.bias]
Loading weights:   2%|▏         | 2/105 [00:00<00:00, 908.94it/s, Materializing param=distilbert.embeddings.LayerNorm.weight]
Loading weights:   2%|▏         | 2/105 [00:00<00:00, 724.34it/s, Materializing param=distilbert.embeddings.LayerNorm.weight]
Loading weights:   3%|▎         | 3/105 [00:00<00:00, 816.07it/s, Materializing param=distilbert.embeddings.position_embeddings.weight]
Loading weights:   3%|▎         | 3/105 [00:00<00:00, 568.08it/s, Materializing param=distilbert.embeddings.position_embeddings.weight]
Loading weights:   4%|▍         | 4/105 [00:00<00:00, 532.76it/s, Materializing param=distilbert.embeddings.word_embeddings.weight]    
Loading weights:   4%|▍         | 4/105 [00:00<00:00, 460.85it/s, Materializing param=distilbert.embeddings.word_embeddings.weight]
Loading weights:   5%|▍         | 5/105 [00:00<00:00, 540.54it/s, Materializing param=distilbert.transformer.layer.0.attention.k_lin.bias]
Loading weights:   5%|▍         | 5/105 [00:00<00:00, 512.53it/s, Materializing param=distilbert.transformer.layer.0.attention.k_lin.bias]
Loading weights:   6%|▌         | 6/105 [00:00<00:00, 549.17it/s, Materializing param=distilbert.transformer.layer.0.attention.k_lin.weight]
Loading weights:   6%|▌         | 6/105 [00:00<00:00, 479.68it/s, Materializing param=distilbert.transformer.layer.0.attention.k_lin.weight]
Loading weights:   7%|▋         | 7/105 [00:00<00:00, 446.01it/s, Materializing param=distilbert.transformer.layer.0.attention.out_lin.bias]
Loading weights:   7%|▋         | 7/105 [00:00<00:00, 395.39it/s, Materializing param=distilbert.transformer.layer.0.attention.out_lin.bias]
Loading weights:   8%|▊         | 8/105 [00:00<00:00, 409.33it/s, Materializing param=distilbert.transformer.layer.0.attention.out_lin.weight]
Loading weights:   8%|▊         | 8/105 [00:00<00:00, 401.03it/s, Materializing param=distilbert.transformer.layer.0.attention.out_lin.weight]
Loading weights:   9%|▊         | 9/105 [00:00<00:00, 431.01it/s, Materializing param=distilbert.transformer.layer.0.attention.q_lin.bias]    
Loading weights:   9%|▊         | 9/105 [00:00<00:00, 427.65it/s, Materializing param=distilbert.transformer.layer.0.attention.q_lin.bias]
Loading weights:  10%|▉         | 10/105 [00:00<00:00, 412.69it/s, Materializing param=distilbert.transformer.layer.0.attention.q_lin.weight]
Loading weights:  10%|▉         | 10/105 [00:00<00:00, 410.22it/s, Materializing param=distilbert.transformer.layer.0.attention.q_lin.weight]
Loading weights:  10%|█         | 11/105 [00:00<00:00, 418.34it/s, Materializing param=distilbert.transformer.layer.0.attention.v_lin.bias]  
Loading weights:  10%|█         | 11/105 [00:00<00:00, 413.70it/s, Materializing param=distilbert.transformer.layer.0.attention.v_lin.bias]
Loading weights:  11%|█▏        | 12/105 [00:00<00:00, 438.43it/s, Materializing param=distilbert.transformer.layer.0.attention.v_lin.weight]
Loading weights:  11%|█▏        | 12/105 [00:00<00:00, 435.09it/s, Materializing param=distilbert.transformer.layer.0.attention.v_lin.weight]
Loading weights:  12%|█▏        | 13/105 [00:00<00:00, 436.68it/s, Materializing param=distilbert.transformer.layer.0.ffn.lin1.bias]         
Loading weights:  12%|█▏        | 13/105 [00:00<00:00, 412.23it/s, Materializing param=distilbert.transformer.layer.0.ffn.lin1.bias]
Loading weights:  13%|█▎        | 14/105 [00:00<00:00, 433.37it/s, Materializing param=distilbert.transformer.layer.0.ffn.lin1.weight]
Loading weights:  13%|█▎        | 14/105 [00:00<00:00, 413.81it/s, Materializing param=distilbert.transformer.layer.0.ffn.lin1.weight]
Loading weights:  14%|█▍        | 15/105 [00:00<00:00, 435.57it/s, Materializing param=distilbert.transformer.layer.0.ffn.lin2.bias]  
Loading weights:  14%|█▍        | 15/105 [00:00<00:00, 430.39it/s, Materializing param=distilbert.transformer.layer.0.ffn.lin2.bias]
Loading weights:  15%|█▌        | 16/105 [00:00<00:00, 448.81it/s, Materializing param=distilbert.transformer.layer.0.ffn.lin2.weight]
Loading weights:  15%|█▌        | 16/105 [00:00<00:00, 427.04it/s, Materializing param=distilbert.transformer.layer.0.ffn.lin2.weight]
Loading weights:  16%|█▌        | 17/105 [00:00<00:00, 387.08it/s, Materializing param=distilbert.transformer.layer.0.output_layer_norm.bias]
Loading weights:  16%|█▌        | 17/105 [00:00<00:00, 380.94it/s, Materializing param=distilbert.transformer.layer.0.output_layer_norm.bias]
Loading weights:  17%|█▋        | 18/105 [00:00<00:00, 397.72it/s, Materializing param=distilbert.transformer.layer.0.output_layer_norm.weight]
Loading weights:  17%|█▋        | 18/105 [00:00<00:00, 396.85it/s, Materializing param=distilbert.transformer.layer.0.output_layer_norm.weight]
Loading weights:  18%|█▊        | 19/105 [00:00<00:00, 393.15it/s, Materializing param=distilbert.transformer.layer.0.sa_layer_norm.bias]      
Loading weights:  18%|█▊        | 19/105 [00:00<00:00, 377.92it/s, Materializing param=distilbert.transformer.layer.0.sa_layer_norm.bias]
Loading weights:  19%|█▉        | 20/105 [00:00<00:00, 379.21it/s, Materializing param=distilbert.transformer.layer.0.sa_layer_norm.weight]
Loading weights:  19%|█▉        | 20/105 [00:00<00:00, 377.10it/s, Materializing param=distilbert.transformer.layer.0.sa_layer_norm.weight]
Loading weights:  20%|██        | 21/105 [00:00<00:00, 375.16it/s, Materializing param=distilbert.transformer.layer.1.attention.k_lin.bias]
Loading weights:  20%|██        | 21/105 [00:00<00:00, 364.18it/s, Materializing param=distilbert.transformer.layer.1.attention.k_lin.bias]
Loading weights:  21%|██        | 22/105 [00:00<00:00, 371.25it/s, Materializing param=distilbert.transformer.layer.1.attention.k_lin.weight]
Loading weights:  21%|██        | 22/105 [00:00<00:00, 367.13it/s, Materializing param=distilbert.transformer.layer.1.attention.k_lin.weight]
Loading weights:  22%|██▏       | 23/105 [00:00<00:00, 363.52it/s, Materializing param=distilbert.transformer.layer.1.attention.out_lin.bias]
Loading weights:  22%|██▏       | 23/105 [00:00<00:00, 362.17it/s, Materializing param=distilbert.transformer.layer.1.attention.out_lin.bias]
Loading weights:  23%|██▎       | 24/105 [00:00<00:00, 357.59it/s, Materializing param=distilbert.transformer.layer.1.attention.out_lin.weight]
Loading weights:  23%|██▎       | 24/105 [00:00<00:00, 355.18it/s, Materializing param=distilbert.transformer.layer.1.attention.out_lin.weight]
Loading weights:  24%|██▍       | 25/105 [00:00<00:00, 354.36it/s, Materializing param=distilbert.transformer.layer.1.attention.q_lin.bias]    
Loading weights:  24%|██▍       | 25/105 [00:00<00:00, 353.59it/s, Materializing param=distilbert.transformer.layer.1.attention.q_lin.bias]
Loading weights:  25%|██▍       | 26/105 [00:00<00:00, 366.28it/s, Materializing param=distilbert.transformer.layer.1.attention.q_lin.weight]
Loading weights:  25%|██▍       | 26/105 [00:00<00:00, 365.75it/s, Materializing param=distilbert.transformer.layer.1.attention.q_lin.weight]
Loading weights:  26%|██▌       | 27/105 [00:00<00:00, 371.93it/s, Materializing param=distilbert.transformer.layer.1.attention.v_lin.bias]  
Loading weights:  26%|██▌       | 27/105 [00:00<00:00, 362.35it/s, Materializing param=distilbert.transformer.layer.1.attention.v_lin.bias]
Loading weights:  27%|██▋       | 28/105 [00:00<00:00, 361.69it/s, Materializing param=distilbert.transformer.layer.1.attention.v_lin.weight]
Loading weights:  27%|██▋       | 28/105 [00:00<00:00, 359.59it/s, Materializing param=distilbert.transformer.layer.1.attention.v_lin.weight]
Loading weights:  28%|██▊       | 29/105 [00:00<00:00, 349.37it/s, Materializing param=distilbert.transformer.layer.1.ffn.lin1.bias]         
Loading weights:  28%|██▊       | 29/105 [00:00<00:00, 347.01it/s, Materializing param=distilbert.transformer.layer.1.ffn.lin1.bias]
Loading weights:  29%|██▊       | 30/105 [00:00<00:00, 345.11it/s, Materializing param=distilbert.transformer.layer.1.ffn.lin1.weight]
Loading weights:  29%|██▊       | 30/105 [00:00<00:00, 342.00it/s, Materializing param=distilbert.transformer.layer.1.ffn.lin1.weight]
Loading weights:  30%|██▉       | 31/105 [00:00<00:00, 343.98it/s, Materializing param=distilbert.transformer.layer.1.ffn.lin2.bias]  
Loading weights:  30%|██▉       | 31/105 [00:00<00:00, 340.02it/s, Materializing param=distilbert.transformer.layer.1.ffn.lin2.bias]
Loading weights:  30%|███       | 32/105 [00:00<00:00, 349.84it/s, Materializing param=distilbert.transformer.layer.1.ffn.lin2.weight]
Loading weights:  30%|███       | 32/105 [00:00<00:00, 346.59it/s, Materializing param=distilbert.transformer.layer.1.ffn.lin2.weight]
Loading weights:  31%|███▏      | 33/105 [00:00<00:00, 352.61it/s, Materializing param=distilbert.transformer.layer.1.output_layer_norm.bias]
Loading weights:  31%|███▏      | 33/105 [00:00<00:00, 351.76it/s, Materializing param=distilbert.transformer.layer.1.output_layer_norm.bias]
Loading weights:  32%|███▏      | 34/105 [00:00<00:00, 359.94it/s, Materializing param=distilbert.transformer.layer.1.output_layer_norm.weight]
Loading weights:  32%|███▏      | 34/105 [00:00<00:00, 353.92it/s, Materializing param=distilbert.transformer.layer.1.output_layer_norm.weight]
Loading weights:  33%|███▎      | 35/105 [00:00<00:00, 362.09it/s, Materializing param=distilbert.transformer.layer.1.sa_layer_norm.bias]      
Loading weights:  33%|███▎      | 35/105 [00:00<00:00, 360.91it/s, Materializing param=distilbert.transformer.layer.1.sa_layer_norm.bias]
Loading weights:  34%|███▍      | 36/105 [00:00<00:00, 368.42it/s, Materializing param=distilbert.transformer.layer.1.sa_layer_norm.weight]
Loading weights:  34%|███▍      | 36/105 [00:00<00:00, 367.79it/s, Materializing param=distilbert.transformer.layer.1.sa_layer_norm.weight]
Loading weights:  35%|███▌      | 37/105 [00:00<00:00, 377.50it/s, Materializing param=distilbert.transformer.layer.2.attention.k_lin.bias]
Loading weights:  35%|███▌      | 37/105 [00:00<00:00, 377.21it/s, Materializing param=distilbert.transformer.layer.2.attention.k_lin.bias]
Loading weights:  36%|███▌      | 38/105 [00:00<00:00, 386.80it/s, Materializing param=distilbert.transformer.layer.2.attention.k_lin.weight]
Loading weights:  36%|███▌      | 38/105 [00:00<00:00, 386.39it/s, Materializing param=distilbert.transformer.layer.2.attention.k_lin.weight]
Loading weights:  37%|███▋      | 39/105 [00:00<00:00, 396.13it/s, Materializing param=distilbert.transformer.layer.2.attention.out_lin.bias]
Loading weights:  37%|███▋      | 39/105 [00:00<00:00, 395.87it/s, Materializing param=distilbert.transformer.layer.2.attention.out_lin.bias]
Loading weights:  38%|███▊      | 40/105 [00:00<00:00, 405.59it/s, Materializing param=distilbert.transformer.layer.2.attention.out_lin.weight]
Loading weights:  38%|███▊      | 40/105 [00:00<00:00, 405.34it/s, Materializing param=distilbert.transformer.layer.2.attention.out_lin.weight]
Loading weights:  39%|███▉      | 41/105 [00:00<00:00, 415.05it/s, Materializing param=distilbert.transformer.layer.2.attention.q_lin.bias]    
Loading weights:  39%|███▉      | 41/105 [00:00<00:00, 414.80it/s, Materializing param=distilbert.transformer.layer.2.attention.q_lin.bias]
Loading weights:  40%|████      | 42/105 [00:00<00:00, 424.48it/s, Materializing param=distilbert.transformer.layer.2.attention.q_lin.weight]
Loading weights:  40%|████      | 42/105 [00:00<00:00, 424.21it/s, Materializing param=distilbert.transformer.layer.2.attention.q_lin.weight]
Loading weights:  41%|████      | 43/105 [00:00<00:00, 433.88it/s, Materializing param=distilbert.transformer.layer.2.attention.v_lin.bias]  
Loading weights:  41%|████      | 43/105 [00:00<00:00, 433.62it/s, Materializing param=distilbert.transformer.layer.2.attention.v_lin.bias]
Loading weights:  42%|████▏     | 44/105 [00:00<00:00, 443.25it/s, Materializing param=distilbert.transformer.layer.2.attention.v_lin.weight]
Loading weights:  42%|████▏     | 44/105 [00:00<00:00, 442.99it/s, Materializing param=distilbert.transformer.layer.2.attention.v_lin.weight]
Loading weights:  43%|████▎     | 45/105 [00:00<00:00, 452.61it/s, Materializing param=distilbert.transformer.layer.2.ffn.lin1.bias]         
Loading weights:  43%|████▎     | 45/105 [00:00<00:00, 452.34it/s, Materializing param=distilbert.transformer.layer.2.ffn.lin1.bias]
Loading weights:  44%|████▍     | 46/105 [00:00<00:00, 461.90it/s, Materializing param=distilbert.transformer.layer.2.ffn.lin1.weight]
Loading weights:  44%|████▍     | 46/105 [00:00<00:00, 461.51it/s, Materializing param=distilbert.transformer.layer.2.ffn.lin1.weight]
Loading weights:  45%|████▍     | 47/105 [00:00<00:00, 471.07it/s, Materializing param=distilbert.transformer.layer.2.ffn.lin2.bias]  
Loading weights:  45%|████▍     | 47/105 [00:00<00:00, 470.79it/s, Materializing param=distilbert.transformer.layer.2.ffn.lin2.bias]
Loading weights:  46%|████▌     | 48/105 [00:00<00:00, 480.32it/s, Materializing param=distilbert.transformer.layer.2.ffn.lin2.weight]
Loading weights:  46%|████▌     | 48/105 [00:00<00:00, 480.04it/s, Materializing param=distilbert.transformer.layer.2.ffn.lin2.weight]
Loading weights:  47%|████▋     | 49/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.2.ffn.lin2.weight]
Loading weights:  47%|████▋     | 49/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.2.output_layer_norm.bias]
Loading weights:  47%|████▋     | 49/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.2.output_layer_norm.bias]
Loading weights:  48%|████▊     | 50/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.2.output_layer_norm.weight]
Loading weights:  48%|████▊     | 50/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.2.output_layer_norm.weight]
Loading weights:  49%|████▊     | 51/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.2.sa_layer_norm.bias]      
Loading weights:  49%|████▊     | 51/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.2.sa_layer_norm.bias]
Loading weights:  50%|████▉     | 52/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.2.sa_layer_norm.weight]
Loading weights:  50%|████▉     | 52/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.2.sa_layer_norm.weight]
Loading weights:  50%|█████     | 53/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.k_lin.bias]
Loading weights:  50%|█████     | 53/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.k_lin.bias]
Loading weights:  51%|█████▏    | 54/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.k_lin.weight]
Loading weights:  51%|█████▏    | 54/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.k_lin.weight]
Loading weights:  52%|█████▏    | 55/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.out_lin.bias]
Loading weights:  52%|█████▏    | 55/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.out_lin.bias]
Loading weights:  53%|█████▎    | 56/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.out_lin.weight]
Loading weights:  53%|█████▎    | 56/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.out_lin.weight]
Loading weights:  54%|█████▍    | 57/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.q_lin.bias]    
Loading weights:  54%|█████▍    | 57/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.q_lin.bias]
Loading weights:  55%|█████▌    | 58/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.q_lin.weight]
Loading weights:  55%|█████▌    | 58/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.q_lin.weight]
Loading weights:  56%|█████▌    | 59/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.v_lin.bias]  
Loading weights:  56%|█████▌    | 59/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.v_lin.bias]
Loading weights:  57%|█████▋    | 60/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.v_lin.weight]
Loading weights:  57%|█████▋    | 60/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.attention.v_lin.weight]
Loading weights:  58%|█████▊    | 61/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.ffn.lin1.bias]         
Loading weights:  58%|█████▊    | 61/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.ffn.lin1.bias]
Loading weights:  59%|█████▉    | 62/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.ffn.lin1.weight]
Loading weights:  59%|█████▉    | 62/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.ffn.lin1.weight]
Loading weights:  60%|██████    | 63/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.ffn.lin2.bias]  
Loading weights:  60%|██████    | 63/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.ffn.lin2.bias]
Loading weights:  61%|██████    | 64/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.ffn.lin2.weight]
Loading weights:  61%|██████    | 64/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.ffn.lin2.weight]
Loading weights:  62%|██████▏   | 65/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.output_layer_norm.bias]
Loading weights:  62%|██████▏   | 65/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.output_layer_norm.bias]
Loading weights:  63%|██████▎   | 66/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.output_layer_norm.weight]
Loading weights:  63%|██████▎   | 66/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.output_layer_norm.weight]
Loading weights:  64%|██████▍   | 67/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.sa_layer_norm.bias]      
Loading weights:  64%|██████▍   | 67/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.sa_layer_norm.bias]
Loading weights:  65%|██████▍   | 68/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.sa_layer_norm.weight]
Loading weights:  65%|██████▍   | 68/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.3.sa_layer_norm.weight]
Loading weights:  66%|██████▌   | 69/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.k_lin.bias]
Loading weights:  66%|██████▌   | 69/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.k_lin.bias]
Loading weights:  67%|██████▋   | 70/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.k_lin.weight]
Loading weights:  67%|██████▋   | 70/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.k_lin.weight]
Loading weights:  68%|██████▊   | 71/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.out_lin.bias]
Loading weights:  68%|██████▊   | 71/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.out_lin.bias]
Loading weights:  69%|██████▊   | 72/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.out_lin.weight]
Loading weights:  69%|██████▊   | 72/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.out_lin.weight]
Loading weights:  70%|██████▉   | 73/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.q_lin.bias]    
Loading weights:  70%|██████▉   | 73/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.q_lin.bias]
Loading weights:  70%|███████   | 74/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.q_lin.weight]
Loading weights:  70%|███████   | 74/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.q_lin.weight]
Loading weights:  71%|███████▏  | 75/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.v_lin.bias]  
Loading weights:  71%|███████▏  | 75/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.v_lin.bias]
Loading weights:  72%|███████▏  | 76/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.v_lin.weight]
Loading weights:  72%|███████▏  | 76/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.attention.v_lin.weight]
Loading weights:  73%|███████▎  | 77/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.ffn.lin1.bias]         
Loading weights:  73%|███████▎  | 77/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.ffn.lin1.bias]
Loading weights:  74%|███████▍  | 78/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.ffn.lin1.weight]
Loading weights:  74%|███████▍  | 78/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.ffn.lin1.weight]
Loading weights:  75%|███████▌  | 79/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.ffn.lin2.bias]  
Loading weights:  75%|███████▌  | 79/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.ffn.lin2.bias]
Loading weights:  76%|███████▌  | 80/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.ffn.lin2.weight]
Loading weights:  76%|███████▌  | 80/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.ffn.lin2.weight]
Loading weights:  77%|███████▋  | 81/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.output_layer_norm.bias]
Loading weights:  77%|███████▋  | 81/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.output_layer_norm.bias]
Loading weights:  78%|███████▊  | 82/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.output_layer_norm.weight]
Loading weights:  78%|███████▊  | 82/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.output_layer_norm.weight]
Loading weights:  79%|███████▉  | 83/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.sa_layer_norm.bias]      
Loading weights:  79%|███████▉  | 83/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.sa_layer_norm.bias]
Loading weights:  80%|████████  | 84/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.sa_layer_norm.weight]
Loading weights:  80%|████████  | 84/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.4.sa_layer_norm.weight]
Loading weights:  81%|████████  | 85/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.k_lin.bias]
Loading weights:  81%|████████  | 85/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.k_lin.bias]
Loading weights:  82%|████████▏ | 86/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.k_lin.weight]
Loading weights:  82%|████████▏ | 86/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.k_lin.weight]
Loading weights:  83%|████████▎ | 87/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.out_lin.bias]
Loading weights:  83%|████████▎ | 87/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.out_lin.bias]
Loading weights:  84%|████████▍ | 88/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.out_lin.weight]
Loading weights:  84%|████████▍ | 88/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.out_lin.weight]
Loading weights:  85%|████████▍ | 89/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.q_lin.bias]    
Loading weights:  85%|████████▍ | 89/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.q_lin.bias]
Loading weights:  86%|████████▌ | 90/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.q_lin.weight]
Loading weights:  86%|████████▌ | 90/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.q_lin.weight]
Loading weights:  87%|████████▋ | 91/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.v_lin.bias]  
Loading weights:  87%|████████▋ | 91/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.v_lin.bias]
Loading weights:  88%|████████▊ | 92/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.v_lin.weight]
Loading weights:  88%|████████▊ | 92/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.attention.v_lin.weight]
Loading weights:  89%|████████▊ | 93/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.ffn.lin1.bias]         
Loading weights:  89%|████████▊ | 93/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.ffn.lin1.bias]
Loading weights:  90%|████████▉ | 94/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.ffn.lin1.weight]
Loading weights:  90%|████████▉ | 94/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.ffn.lin1.weight]
Loading weights:  90%|█████████ | 95/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.ffn.lin2.bias]  
Loading weights:  90%|█████████ | 95/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.ffn.lin2.bias]
Loading weights:  91%|█████████▏| 96/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.ffn.lin2.weight]
Loading weights:  91%|█████████▏| 96/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.ffn.lin2.weight]
Loading weights:  92%|█████████▏| 97/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.output_layer_norm.bias]
Loading weights:  92%|█████████▏| 97/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.output_layer_norm.bias]
Loading weights:  93%|█████████▎| 98/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.output_layer_norm.weight]
Loading weights:  93%|█████████▎| 98/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.output_layer_norm.weight]
Loading weights:  94%|█████████▍| 99/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.sa_layer_norm.bias]      
Loading weights:  94%|█████████▍| 99/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.sa_layer_norm.bias]
Loading weights:  95%|█████████▌| 100/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.sa_layer_norm.weight]
Loading weights:  95%|█████████▌| 100/105 [00:00<00:00, 489.39it/s, Materializing param=distilbert.transformer.layer.5.sa_layer_norm.weight]
Loading weights:  96%|█████████▌| 101/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_layer_norm.bias]                              
Loading weights:  96%|█████████▌| 101/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_layer_norm.bias]
Loading weights:  97%|█████████▋| 102/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_layer_norm.weight]
Loading weights:  97%|█████████▋| 102/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_layer_norm.weight]
Loading weights:  98%|█████████▊| 103/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_projector.bias]   
Loading weights:  98%|█████████▊| 103/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_projector.bias]
Loading weights:  99%|█████████▉| 104/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_transform.bias]
Loading weights:  99%|█████████▉| 104/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_transform.bias]
Loading weights: 100%|██████████| 105/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_transform.weight]
Loading weights: 100%|██████████| 105/105 [00:00<00:00, 489.39it/s, Materializing param=vocab_transform.weight]
Loading weights: 100%|██████████| 105/105 [00:00<00:00, 943.82it/s, Materializing param=vocab_transform.weight]
[AccuracyLens] Captured MLM targets (9 samples)
/home/boyucwsl/.local/share/uv/python/cpython-3.11.14-linux-x86_64-gnu/lib/python3.11/copyreg.py:105: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
Creating new layout from 274 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 161.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[AccuracyLens] Using backend-captured fallback dataset from PipelineGraphCollectorLens
[AccuracyLens] Auto-configured mlm evaluator with 5 metrics
[Observatory] Collected Exported Float in 6977.6 ms
Creating new layout from 815 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 259.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[Observatory] Collected Annotated Model in 2800.7 ms
Creating new layout from 815 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 259.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[Observatory] Collected Calibrated Model in 2850.8 ms
Creating new layout from 1118 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 350.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[Observatory] Collected Quantized Model in 7444.4 ms
[INFO] [Qnn ExecuTorch]: Creating new backend bundle.
[INFO] [Qnn ExecuTorch]: create QNN Logger with log_level 1
Creating new layout from 1118 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 350.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[Observatory] Collected Pre-EdgeTransform/forward in 5731.5 ms
/home/boyucwsl/.local/share/uv/python/cpython-3.11.14-linux-x86_64-gnu/lib/python3.11/copyreg.py:105: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
Op aten.unbind.int was requested for preservation by partitioner.  This request is ignored because it is in a blocklist.
/home/boyucwsl/.local/share/uv/python/cpython-3.11.14-linux-x86_64-gnu/lib/python3.11/copyreg.py:105: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
Op aten.unbind.int was requested for preservation by partitioner.  This request is ignored because it is in a blocklist.
/home/boyucwsl/.local/share/uv/python/cpython-3.11.14-linux-x86_64-gnu/lib/python3.11/copyreg.py:105: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
Creating new layout from 1118 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 350.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[Observatory] Collected ETRecord Exported/forward in 6130.7 ms
/home/boyucwsl/.local/share/uv/python/cpython-3.11.14-linux-x86_64-gnu/lib/python3.11/copyreg.py:105: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
Creating new layout from 1132 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 350.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[Observatory] Collected ETRecord Edge/forward in 9045.0 ms
/home/boyucwsl/.local/share/uv/python/cpython-3.11.14-linux-x86_64-gnu/lib/python3.11/copyreg.py:105: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
Creating new layout from 602 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 154.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[Observatory] Collected ETRecord Extra/edge_after_transform/forward in 8294.7 ms
/home/boyucwsl/.local/share/uv/python/cpython-3.11.14-linux-x86_64-gnu/lib/python3.11/copyreg.py:105: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
Building graph payload model...
Running fast-sugiyama layout...
[FX Graph Viewer] Compiling base graph payload...
[ FX Graph Viewer ] Compiling extension payloads...
No quant config is implemented for op, aten.dropout.default
No quant config is implemented for op, aten.clone.default
No quant config is implemented for op, aten.dropout.default
No quant config is implemented for op, aten.clone.default
No quant config is implemented for op, aten.dropout.default
No quant config is implemented for op, aten.clone.default
No quant config is implemented for op, aten.dropout.default
No quant config is implemented for op, aten.clone.default
No quant config is implemented for op, aten.dropout.default
No quant config is implemented for op, aten.clone.default
No quant config is implemented for op, aten.dropout.default
No quant config is implemented for op, aten.clone.default
No quant config is implemented for op, aten.dropout.default
Building graph payload model...
Running fast-sugiyama layout...
[FX Graph Viewer] Compiling base graph payload...
[ FX Graph Viewer ] Compiling extension payloads...
Quantizing(PTQ) the model...
Building graph payload model...
Running fast-sugiyama layout...
[FX Graph Viewer] Compiling base graph payload...
[ FX Graph Viewer ] Compiling extension payloads...
Building graph payload model...
Running fast-sugiyama layout...
[FX Graph Viewer] Compiling base graph payload...
[ FX Graph Viewer ] Compiling extension payloads...
Building graph payload model...
Running fast-sugiyama layout...
[FX Graph Viewer] Compiling base graph payload...
[ FX Graph Viewer ] Compiling extension payloads...
Building graph payload model...
Running fast-sugiyama layout...
[FX Graph Viewer] Compiling base graph payload...
[ FX Graph Viewer ] Compiling extension payloads...
Building graph payload model...
Running fast-sugiyama layout...
[FX Graph Viewer] Compiling base graph payload...
[ FX Graph Viewer ] Compiling extension payloads...
Building graph payload model...
Running fast-sugiyama layout...
[FX Graph Viewer] Compiling base graph payload...
[ FX Graph Viewer ] Compiling extension payloads...
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.gelu.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.gelu.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten._safe_softmax.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.where.self | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.gelu.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten._safe_softmax.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.where.self | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.gelu.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten._safe_softmax.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.where.self | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.gelu.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten._safe_softmax.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.where.self | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.gelu.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten._safe_softmax.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.where.self | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | TrueFound 1 subgraphs to be partitioned.
Qnn partitioner will delegate torch mutable buffer with the same I/O address during the runtime, so if your model contains mutable buffer, then you can get the better performance with skip_mutable_buffer=False. If you encounter accuracy issue during the runtime, then please set `skip_mutable_buffer=True` and try again.
/home/boyucwsl/.local/share/uv/python/cpython-3.11.14-linux-x86_64-gnu/lib/python3.11/copyreg.py:105: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
[INFO] [Qnn ExecuTorch]: Initialize Qnn backend parameters for Qnn executorch backend type 2
[INFO] [Qnn ExecuTorch]: Caching: Caching is in SAVE MODE.
[INFO] [Qnn ExecuTorch]: Running level=3 optimization.
Processing Method(0): (1/1)
Visiting: scalar_tensor, scalar_tensor.default
Visiting: scalar_tensor_1, scalar_tensor.default
Visiting: scalar_tensor_2, scalar_tensor.default
Visiting: scalar_tensor_3, scalar_tensor.default
Visiting: scalar_tensor_4, scalar_tensor.default
Visiting: scalar_tensor_5, scalar_tensor.default
Visiting: scalar_tensor_6, scalar_tensor.default
Visiting: scalar_tensor_7, scalar_tensor.default
Visiting: scalar_tensor_8, scalar_tensor.default
Visiting: scalar_tensor_9, scalar_tensor.default
Visiting: scalar_tensor_10, scalar_tensor.default
Visiting: scalar_tensor_11, scalar_tensor.default
Visiting: aten__to_copy_default, aten._to_copy.default
Visiting: aten_where_self, aten.where.self
Visiting: aten_where_self_1, aten.where.self
Visiting: aten_where_self_2, aten.where.self
Visiting: aten_where_self_3, aten.where.self
Visiting: aten_where_self_4, aten.where.self
Visiting: aten_where_self_5, aten.where.self
Visiting: aten_embedding_default, aten.embedding.default
Visiting: aten_view_copy_default_9, aten.view_copy.default
Visiting: aten_view_copy_default_10, aten.view_copy.default
Visiting: aten_view_copy_default_33, aten.view_copy.default
Visiting: aten_view_copy_default_34, aten.view_copy.default
Visiting: aten_view_copy_default_57, aten.view_copy.default
Visiting: aten_view_copy_default_58, aten.view_copy.default
Visiting: aten_view_copy_default_81, aten.view_copy.default
Visiting: aten_view_copy_default_82, aten.view_copy.default
Visiting: aten_view_copy_default_105, aten.view_copy.default
Visiting: aten_view_copy_default_106, aten.view_copy.default
Visiting: aten_view_copy_default_129, aten.view_copy.default
Visiting: aten_view_copy_default_130, aten.view_copy.default
Visiting: aten_slice_copy_tensor, aten.slice_copy.Tensor
Visiting: aten_embedding_default_1, aten.embedding.default
Visiting: aten_add_tensor, aten.add.Tensor
Visiting: aten_native_layer_norm_default, aten.native_layer_norm.default
Visiting: getitem, getitem
Visiting: aten_view_copy_default, aten.view_copy.default
Visiting: aten_view_copy_default_1, aten.view_copy.default
Visiting: aten_view_copy_default_2, aten.view_copy.default
Visiting: aten_linear_default_2, aten.linear.default
Visiting: aten_linear_default_1, aten.linear.default
Visiting: aten_linear_default, aten.linear.default
Visiting: aten_view_copy_default_7, aten.view_copy.default
Visiting: aten_view_copy_default_5, aten.view_copy.default
Visiting: aten_view_copy_default_3, aten.view_copy.default
Visiting: aten_view_copy_default_8, aten.view_copy.default
Visiting: aten_view_copy_default_6, aten.view_copy.default
Visiting: aten_permute_copy_default_42, aten.permute_copy.default
Visiting: aten_view_copy_default_4, aten.view_copy.default
Visiting: aten_permute_copy_default_2, aten.permute_copy.default
Visiting: aten_permute_copy_default, aten.permute_copy.default
Visiting: aten_view_copy_default_15, aten.view_copy.default
Visiting: aten__to_copy_default_1, aten._to_copy.default
Visiting: aten_mul_tensor, aten.mul.Tensor
Visiting: aten_mul_tensor_1, aten.mul.Tensor
Visiting: aten_view_copy_default_11, aten.view_copy.default
Visiting: aten_view_copy_default_12, aten.view_copy.default
Visiting: aten__to_copy_default_2, aten._to_copy.default
Visiting: aten_bmm_default, aten.bmm.default
Visiting: aten_view_copy_default_13, aten.view_copy.default
Visiting: aten_add_tensor_1, aten.add.Tensor
Visiting: aten__safe_softmax_default, aten._safe_softmax.default
Visiting: aten_view_copy_default_14, aten.view_copy.default
Visiting: aten_bmm_default_1, aten.bmm.default
Visiting: aten_view_copy_default_16, aten.view_copy.default
Visiting: aten_permute_copy_default_4, aten.permute_copy.default
Visiting: aten__to_copy_default_3, aten._to_copy.default
Visiting: aten_permute_copy_default_43, aten.permute_copy.default
Visiting: aten_view_copy_default_17, aten.view_copy.default
Visiting: aten_view_copy_default_18, aten.view_copy.default
Visiting: aten_linear_default_3, aten.linear.default
Visiting: aten_view_copy_default_19, aten.view_copy.default
Visiting: aten_add_tensor_2, aten.add.Tensor
Visiting: aten_native_layer_norm_default_1, aten.native_layer_norm.default
Visiting: getitem_1, getitem
Visiting: aten_view_copy_default_20, aten.view_copy.default
Visiting: aten_linear_default_4, aten.linear.default
Visiting: aten_view_copy_default_21, aten.view_copy.default
Visiting: aten_gelu_default, aten.gelu.default
Visiting: aten_view_copy_default_22, aten.view_copy.default
Visiting: aten_linear_default_5, aten.linear.default
Visiting: aten_view_copy_default_23, aten.view_copy.default
Visiting: aten_add_tensor_3, aten.add.Tensor
Visiting: aten_native_layer_norm_default_2, aten.native_layer_norm.default
Visiting: getitem_2, getitem
Visiting: aten_view_copy_default_24, aten.view_copy.default
Visiting: aten_view_copy_default_25, aten.view_copy.default
Visiting: aten_view_copy_default_26, aten.view_copy.default
Visiting: aten_linear_default_8, aten.linear.default
Visiting: aten_linear_default_7, aten.linear.default
Visiting: aten_linear_default_6, aten.linear.default
Visiting: aten_view_copy_default_31, aten.view_copy.default
Visiting: aten_view_copy_default_29, aten.view_copy.default
Visiting: aten_view_copy_default_27, aten.view_copy.default
Visiting: aten_view_copy_default_32, aten.view_copy.default
Visiting: aten_view_copy_default_30, aten.view_copy.default
Visiting: aten_permute_copy_default_44, aten.permute_copy.default
Visiting: aten_view_copy_default_28, aten.view_copy.default
Visiting: aten_permute_copy_default_9, aten.permute_copy.default
Visiting: aten_permute_copy_default_7, aten.permute_copy.default
Visiting: aten_view_copy_default_39, aten.view_copy.default
Visiting: aten__to_copy_default_4, aten._to_copy.default
Visiting: aten_mul_tensor_2, aten.mul.Tensor
Visiting: aten_mul_tensor_3, aten.mul.Tensor
Visiting: aten_view_copy_default_35, aten.view_copy.default
Visiting: aten_view_copy_default_36, aten.view_copy.default
Visiting: aten__to_copy_default_5, aten._to_copy.default
Visiting: aten_bmm_default_2, aten.bmm.default
Visiting: aten_view_copy_default_37, aten.view_copy.default
Visiting: aten_add_tensor_4, aten.add.Tensor
Visiting: aten__safe_softmax_default_1, aten._safe_softmax.default
Visiting: aten_view_copy_default_38, aten.view_copy.default
Visiting: aten_bmm_default_3, aten.bmm.default
Visiting: aten_view_copy_default_40, aten.view_copy.default
Visiting: aten_permute_copy_default_11, aten.permute_copy.default
Visiting: aten__to_copy_default_6, aten._to_copy.default
Visiting: aten_permute_copy_default_45, aten.permute_copy.default
Visiting: aten_view_copy_default_41, aten.view_copy.default
Visiting: aten_view_copy_default_42, aten.view_copy.default
Visiting: aten_linear_default_9, aten.linear.default
Visiting: aten_view_copy_default_43, aten.view_copy.default
Visiting: aten_add_tensor_5, aten.add.Tensor
Visiting: aten_native_layer_norm_default_3, aten.native_layer_norm.default
Visiting: getitem_3, getitem
Visiting: aten_view_copy_default_44, aten.view_copy.default
Visiting: aten_linear_default_10, aten.linear.default
Visiting: aten_view_copy_default_45, aten.view_copy.default
Visiting: aten_gelu_default_1, aten.gelu.default
Visiting: aten_view_copy_default_46, aten.view_copy.default
Visiting: aten_linear_default_11, aten.linear.default
Visiting: aten_view_copy_default_47, aten.view_copy.default
Visiting: aten_add_tensor_6, aten.add.Tensor
Visiting: aten_native_layer_norm_default_4, aten.native_layer_norm.default
Visiting: getitem_4, getitem
Visiting: aten_view_copy_default_48, aten.view_copy.default
Visiting: aten_view_copy_default_49, aten.view_copy.default
Visiting: aten_view_copy_default_50, aten.view_copy.default
Visiting: aten_linear_default_14, aten.linear.default
Visiting: aten_linear_default_13, aten.linear.default
Visiting: aten_linear_default_12, aten.linear.default
Visiting: aten_view_copy_default_55, aten.view_copy.default
Visiting: aten_view_copy_default_53, aten.view_copy.default
Visiting: aten_view_copy_default_51, aten.view_copy.default
Visiting: aten_view_copy_default_56, aten.view_copy.default
Visiting: aten_view_copy_default_54, aten.view_copy.default
Visiting: aten_permute_copy_default_46, aten.permute_copy.default
Visiting: aten_view_copy_default_52, aten.view_copy.default
Visiting: aten_permute_copy_default_16, aten.permute_copy.default
Visiting: aten_permute_copy_default_14, aten.permute_copy.default
Visiting: aten_view_copy_default_63, aten.view_copy.default
Visiting: aten__to_copy_default_7, aten._to_copy.default
Visiting: aten_mul_tensor_4, aten.mul.Tensor
Visiting: aten_mul_tensor_5, aten.mul.Tensor
Visiting: aten_view_copy_default_59, aten.view_copy.default
Visiting: aten_view_copy_default_60, aten.view_copy.default
Visiting: aten__to_copy_default_8, aten._to_copy.default
Visiting: aten_bmm_default_4, aten.bmm.default
Visiting: aten_view_copy_default_61, aten.view_copy.default
Visiting: aten_add_tensor_7, aten.add.Tensor
Visiting: aten__safe_softmax_default_2, aten._safe_softmax.default
Visiting: aten_view_copy_default_62, aten.view_copy.default
Visiting: aten_bmm_default_5, aten.bmm.default
Visiting: aten_view_copy_default_64, aten.view_copy.default
Visiting: aten_permute_copy_default_18, aten.permute_copy.default
Visiting: aten__to_copy_default_9, aten._to_copy.default
Visiting: aten_permute_copy_default_47, aten.permute_copy.default
Visiting: aten_view_copy_default_65, aten.view_copy.default
Visiting: aten_view_copy_default_66, aten.view_copy.default
Visiting: aten_linear_default_15, aten.linear.default
Visiting: aten_view_copy_default_67, aten.view_copy.default
Visiting: aten_add_tensor_8, aten.add.Tensor
Visiting: aten_native_layer_norm_default_5, aten.native_layer_norm.default
Visiting: getitem_5, getitem
Visiting: aten_view_copy_default_68, aten.view_copy.default
Visiting: aten_linear_default_16, aten.linear.default
Visiting: aten_view_copy_default_69, aten.view_copy.default
Visiting: aten_gelu_default_2, aten.gelu.default
Visiting: aten_view_copy_default_70, aten.view_copy.default
Visiting: aten_linear_default_17, aten.linear.default
Visiting: aten_view_copy_default_71, aten.view_copy.default
Visiting: aten_add_tensor_9, aten.add.Tensor
Visiting: aten_native_layer_norm_default_6, aten.native_layer_norm.default
Visiting: getitem_6, getitem
Visiting: aten_view_copy_default_72, aten.view_copy.default
Visiting: aten_view_copy_default_73, aten.view_copy.default
Visiting: aten_view_copy_default_74, aten.view_copy.default
Visiting: aten_linear_default_20, aten.linear.default
Visiting: aten_linear_default_19, aten.linear.default
Visiting: aten_linear_default_18, aten.linear.default
Visiting: aten_view_copy_default_79, aten.view_copy.default
Visiting: aten_view_copy_default_77, aten.view_copy.default
Visiting: aten_view_copy_default_75, aten.view_copy.default
Visiting: aten_view_copy_default_80, aten.view_copy.default
Visiting: aten_view_copy_default_78, aten.view_copy.default
Visiting: aten_permute_copy_default_48, aten.permute_copy.default
Visiting: aten_view_copy_default_76, aten.view_copy.default
Visiting: aten_permute_copy_default_23, aten.permute_copy.default
Visiting: aten_permute_copy_default_21, aten.permute_copy.default
Visiting: aten_view_copy_default_87, aten.view_copy.default
Visiting: aten__to_copy_default_10, aten._to_copy.default
Visiting: aten_mul_tensor_6, aten.mul.Tensor
Visiting: aten_mul_tensor_7, aten.mul.Tensor
Visiting: aten_view_copy_default_83, aten.view_copy.default
Visiting: aten_view_copy_default_84, aten.view_copy.default
Visiting: aten__to_copy_default_11, aten._to_copy.default
Visiting: aten_bmm_default_6, aten.bmm.default
Visiting: aten_view_copy_default_85, aten.view_copy.default
Visiting: aten_add_tensor_10, aten.add.Tensor
Visiting: aten__safe_softmax_default_3, aten._safe_softmax.default
Visiting: aten_view_copy_default_86, aten.view_copy.default
Visiting: aten_bmm_default_7, aten.bmm.default
Visiting: aten_view_copy_default_88, aten.view_copy.default
Visiting: aten_permute_copy_default_25, aten.permute_copy.default
Visiting: aten__to_copy_default_12, aten._to_copy.default
Visiting: aten_permute_copy_default_49, aten.permute_copy.default
Visiting: aten_view_copy_default_89, aten.view_copy.default
Visiting: aten_view_copy_default_90, aten.view_copy.default
Visiting: aten_linear_default_21, aten.linear.default
Visiting: aten_view_copy_default_91, aten.view_copy.default
Visiting: aten_add_tensor_11, aten.add.Tensor
Visiting: aten_native_layer_norm_default_7, aten.native_layer_norm.default
Visiting: getitem_7, getitem
Visiting: aten_view_copy_default_92, aten.view_copy.default
Visiting: aten_linear_default_22, aten.linear.default
Visiting: aten_view_copy_default_93, aten.view_copy.default
Visiting: aten_gelu_default_3, aten.gelu.default
Visiting: aten_view_copy_default_94, aten.view_copy.default
Visiting: aten_linear_default_23, aten.linear.default
Visiting: aten_view_copy_default_95, aten.view_copy.default
Visiting: aten_add_tensor_12, aten.add.Tensor
Visiting: aten_native_layer_norm_default_8, aten.native_layer_norm.default
Visiting: getitem_8, getitem
Visiting: aten_view_copy_default_96, aten.view_copy.default
Visiting: aten_view_copy_default_97, aten.view_copy.default
Visiting: aten_view_copy_default_98, aten.view_copy.default
Visiting: aten_linear_default_26, aten.linear.default
Visiting: aten_linear_default_25, aten.linear.default
Visiting: aten_linear_default_24, aten.linear.default
Visiting: aten_view_copy_default_103, aten.view_copy.default
Visiting: aten_view_copy_default_101, aten.view_copy.default
Visiting: aten_view_copy_default_99, aten.view_copy.default
Visiting: aten_view_copy_default_104, aten.view_copy.default
Visiting: aten_view_copy_default_102, aten.view_copy.default
Visiting: aten_permute_copy_default_50, aten.permute_copy.default
Visiting: aten_view_copy_default_100, aten.view_copy.default
Visiting: aten_permute_copy_default_30, aten.permute_copy.default
Visiting: aten_permute_copy_default_28, aten.permute_copy.default
Visiting: aten_view_copy_default_111, aten.view_copy.default
Visiting: aten__to_copy_default_13, aten._to_copy.default
Visiting: aten_mul_tensor_8, aten.mul.Tensor
Visiting: aten_mul_tensor_9, aten.mul.Tensor
Visiting: aten_view_copy_default_107, aten.view_copy.default
Visiting: aten_view_copy_default_108, aten.view_copy.default
Visiting: aten__to_copy_default_14, aten._to_copy.default
Visiting: aten_bmm_default_8, aten.bmm.default
Visiting: aten_view_copy_default_109, aten.view_copy.default
Visiting: aten_add_tensor_13, aten.add.Tensor
Visiting: aten__safe_softmax_default_4, aten._safe_softmax.default
Visiting: aten_view_copy_default_110, aten.view_copy.default
Visiting: aten_bmm_default_9, aten.bmm.default
Visiting: aten_view_copy_default_112, aten.view_copy.default
Visiting: aten_permute_copy_default_32, aten.permute_copy.default
Visiting: aten__to_copy_default_15, aten._to_copy.default
Visiting: aten_permute_copy_default_51, aten.permute_copy.default
Visiting: aten_view_copy_default_113, aten.view_copy.default
Visiting: aten_view_copy_default_114, aten.view_copy.default
Visiting: aten_linear_default_27, aten.linear.default
Visiting: aten_view_copy_default_115, aten.view_copy.default
Visiting: aten_add_tensor_14, aten.add.Tensor
Visiting: aten_native_layer_norm_default_9, aten.native_layer_norm.default
Visiting: getitem_9, getitem
Visiting: aten_view_copy_default_116, aten.view_copy.default
Visiting: aten_linear_default_28, aten.linear.default
Visiting: aten_view_copy_default_117, aten.view_copy.default
Visiting: aten_gelu_default_4, aten.gelu.default
Visiting: aten_view_copy_default_118, aten.view_copy.default
Visiting: aten_linear_default_29, aten.linear.default
Visiting: aten_view_copy_default_119, aten.view_copy.default
Visiting: aten_add_tensor_15, aten.add.Tensor
Visiting: aten_native_layer_norm_default_10, aten.native_layer_norm.default
Visiting: getitem_10, getitem
Visiting: aten_view_copy_default_120, aten.view_copy.default
Visiting: aten_view_copy_default_121, aten.view_copy.default
Visiting: aten_view_copy_default_122, aten.view_copy.default
Visiting: aten_linear_default_32, aten.linear.default
Visiting: aten_linear_default_31, aten.linear.default
Visiting: aten_linear_default_30, aten.linear.default
Visiting: aten_view_copy_default_127, aten.view_copy.default
Visiting: aten_view_copy_default_125, aten.view_copy.default
Visiting: aten_view_copy_default_123, aten.view_copy.default
Visiting: aten_view_copy_default_128, aten.view_copy.default
Visiting: aten_view_copy_default_126, aten.view_copy.default
Visiting: aten_permute_copy_default_52, aten.permute_copy.default
Visiting: aten_view_copy_default_124, aten.view_copy.default
Visiting: aten_permute_copy_default_37, aten.permute_copy.default
Visiting: aten_permute_copy_default_35, aten.permute_copy.default
Visiting: aten_view_copy_default_135, aten.view_copy.default
Visiting: aten__to_copy_default_16, aten._to_copy.default
Visiting: aten_mul_tensor_10, aten.mul.Tensor
Visiting: aten_mul_tensor_11, aten.mul.Tensor
Visiting: aten_view_copy_default_131, aten.view_copy.default
Visiting: aten_view_copy_default_132, aten.view_copy.default
Visiting: aten__to_copy_default_17, aten._to_copy.default
Visiting: aten_bmm_default_10, aten.bmm.default
Visiting: aten_view_copy_default_133, aten.view_copy.default
Visiting: aten_add_tensor_16, aten.add.Tensor
Visiting: aten__safe_softmax_default_5, aten._safe_softmax.default
Visiting: aten_view_copy_default_134, aten.view_copy.default
Visiting: aten_bmm_default_11, aten.bmm.default
Visiting: aten_view_copy_default_136, aten.view_copy.default
Visiting: aten_permute_copy_default_39, aten.permute_copy.default
Visiting: aten__to_copy_default_18, aten._to_copy.default
Visiting: aten_permute_copy_default_53, aten.permute_copy.default
Visiting: aten_view_copy_default_137, aten.view_copy.default
Visiting: aten_view_copy_default_138, aten.view_copy.default
Visiting: aten_linear_default_33, aten.linear.default
Visiting: aten_view_copy_default_139, aten.view_copy.default
Visiting: aten_add_tensor_17, aten.add.Tensor
Visiting: aten_native_layer_norm_default_11, aten.native_layer_norm.default
Visiting: getitem_11, getitem
Visiting: aten_view_copy_default_140, aten.view_copy.default
Visiting: aten_linear_default_34, aten.linear.default
Visiting: aten_view_copy_default_141, aten.view_copy.default
Visiting: aten_gelu_default_5, aten.gelu.default
Visiting: aten_view_copy_default_142, aten.view_copy.default
Visiting: aten_linear_default_35, aten.linear.default
Visiting: aten_view_copy_default_143, aten.view_copy.default
Visiting: aten_add_tensor_18, aten.add.Tensor
Visiting: aten_native_layer_norm_default_12, aten.native_layer_norm.default
Visiting: getitem_12, getitem
Visiting: aten_view_copy_default_144, aten.view_copy.default
Visiting: aten_linear_default_36, aten.linear.default
Visiting: aten_view_copy_default_145, aten.view_copy.default
Visiting: aten_gelu_default_6, aten.gelu.default
Visiting: aten_native_layer_norm_default_13, aten.native_layer_norm.default
Visiting: getitem_13, getitem
Visiting: aten_view_copy_default_146, aten.view_copy.default
Visiting: aten_linear_default_37, aten.linear.default
Visiting: aten_view_copy_default_147, aten.view_copy.default
Visiting: quantized_decomposed_dequantize_per_tensor_tensor, quantized_decomposed.dequantize_per_tensor.tensor

====== DDR bandwidth summary ======
spill_bytes=0
fill_bytes=0
write_total_bytes=26126848
read_total_bytes=81601792

[INFO] [Qnn ExecuTorch]: Destroy Qnn context
/home/boyucwsl/.local/share/uv/python/cpython-3.11.14-linux-x86_64-gnu/lib/python3.11/copyreg.py:105: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
Op aten.unbind.int was requested for preservation by partitioner.  This request is ignored because it is in a blocklist.
Creating new layout from 5 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 175.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[Observatory] Collected EdgeProgramManager EP in 8709.8 ms
[INFO] [Qnn ExecuTorch]: Destroy Qnn device
[INFO] [Qnn ExecuTorch]: Destroy Qnn backend
/home/boyucwsl/executorch/exir/tensor.py:83: FutureWarning: guard_size_oblivious will be removed. Consider using explicit unbacked handling     potentially utilizing guard_or_false, guard_or_true, or statically_known_true
  return guard_size_oblivious(self.stride < other.stride)
[PipelineGraphCollector] Uninstalled all patches
[AccuracyLens] Uninstalled all patches
[Observatory] Exported raw data to /home/boyucwsl/Executorch_Observatory_Demo/generated_reports/qualcomm/distilbert/observatory_report.json
Creating new layout from 274 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 161.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
Creating new layout from 815 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 259.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
Creating new layout from 815 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 259.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
Creating new layout from 1118 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 350.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
Creating new layout from 1118 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 350.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
Creating new layout from 1118 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 350.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
Creating new layout from 1132 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 350.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
Creating new layout from 602 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 161.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
Creating new layout from 5 edges
Splitting graph into its connected components
Initializing graphs vertex weights
Start building layout
Configuration is: Config { minimum_length: 1, vertex_spacing: 175.0, dummy_vertices: true, dummy_size: 1.0, ranking_type: MinimizeEdgeLength, c_minimization: Median, transpose: true, check_layout: true, encode_edges: false }
Executing phase 0: Cycle Removal
Graph contains no cycle
Executing phase 1: Ranking
Start ranking, ranking type: MinimizeEdgeLength, minimum_length: 1
Initializing ranks via topological sort.
building feasible tree
Trying to build tight tree.
Initializing cut values
Calculating cut values of tree edges
initialzing low, lim and parent values of vertices
Executing phase 2: Crossing Reduction
Inserting dummy vertices for edges spanning more than 1 ranks
Initializing order of vertices in each rank via dfs.
Reducing crossings via bilayer sweep
Didn't improve after 4 sweeps, returning
Executing phase 3: Coordinate Calculation
Creating individual layouts for coordinate calculation
Marking type one conflicts (edge crossings between dummy vertices and non dummy vertices)
creating layouts for vertical direction: Down, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Down, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Right
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
creating layouts for vertical direction: Up, horizontal direction Left
Creating vertical alignments
calculating coordinates for layout.
Placing vertices in blocks.
move blocks as close together as possible
Aligning all layouts to the one with the smallest width
Calculate relative coordinates, by taking average between two medians of absolute x-coordinates for each layout direction
[Observatory] Exported HTML report to /home/boyucwsl/Executorch_Observatory_Demo/generated_reports/qualcomm/distilbert/observatory_report.html
[Observatory CLI] Reports: html=/home/boyucwsl/Executorch_Observatory_Demo/generated_reports/qualcomm/distilbert/observatory_report.html json=/home/boyucwsl/Executorch_Observatory_Demo/generated_reports/qualcomm/distilbert/observatory_report.json (9 records: Exported Float, Annotated Model, Calibrated Model, Quantized Model, Pre-EdgeTransform/forward, ETRecord Exported/forward, ETRecord Edge/forward, ETRecord Extra/edge_after_transform/forward, EdgeProgramManager EP)

[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.gelu.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten._safe_softmax.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.bmm.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.mul.Tensor | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.where.self | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.permute_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.linear.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.view_copy.default | True
[QNN Partitioner Op Support]: aten.native_layer_norm.default | True
[QNN Partitioner Op Support]: aten.add.Tensor | True
[QNN Partitioner Op Support]: aten.embedding.default | True
[QNN Partitioner Op Support]: aten.slice_copy.Tensor | True
[QNN Partitioner Op Support]: aten.embedding.default | True
[QNN Partitioner Op Support]: aten._to_copy.default | True
Building graph payload model...
Running fast-sugiyama layout...
[FX Graph Viewer] Compiling base graph payload...
[ FX Graph Viewer ] Compiling extension payloads...
Running fast-sugiyama layout...
Running fast-sugiyama layout...
Running fast-sugiyama layout...
Running fast-sugiyama layout...
Running fast-sugiyama layout...
Running fast-sugiyama layout...
Running fast-sugiyama layout...
Running fast-sugiyama layout...
Running fast-sugiyama layout...
$ bash -lc 'source /home/boyucwsl/executorch/qairt/2.37.0.250724/bin/envsetup.sh && python -m executorch.backends.qualcomm.debugger.observatory --output-html /home/boyucwsl/Executorch_Observatory_Demo/generated_reports/qualcomm/distilbert/observatory_report.html --output-json /home/boyucwsl/Executorch_Observatory_Demo/generated_reports/qualcomm/distilbert/observatory_report.json --lens_recipe accuracy examples/qualcomm/oss_scripts/distilbert.py -m SM8650 -b ./build-android --dataset /home/boyucwsl/executorch/wikisent2.txt -H mlgtw-linux -s bebcca9b -a /home/boyucwsl/Executorch_Observatory_Demo/generated_reports/qualcomm/distilbert/artifacts --seed 1126 --compile_only'