Searched full:dlrm (Results 1 – 25 of 32) sorted by relevance
12
| /aosp_15_r20/external/pytorch/torch/ao/pruning/_experimental/data_sparsifier/benchmarks/ |
| H A D | README.md | 1 # Data Sparsifier Benchmarking using the DLRM Model
4 …ta sparsifier to prune the embedding bags of the [DLRM Model](https://github.com/facebookresearch/…
11 …sifier/data_norm_sparsifier.py) is used to sparsify the embeddings of the DLRM model. The model is…
17 The benchmarks are created for the dlrm model on the Kaggle CriteoDataset which can be downloaded f…
28 …s are introduced on the features before feeding into the top layer of the dlrm model. Post that, w…
29 The takeaway is that the dlrm model with sparse coo tensor is slower (roughly 2x). This is because …
35 The benchmark codes depend on the [DLRM codebase](https://github.com/facebookresearch/dlrm).
36 1. Clone the dlrm git repository
38 3. The DLRM model can be trained using the following script
47 4. Copy the scripts data sparsifier benchmark scripts into to the dlrm directory.
[all …]
|
| H A D | evaluate_model_metrics.py | 18 def inference_and_evaluation(dlrm, test_dataloader, device): argument
25 dlrm (nn.Module)
26 dlrm model object
47 Z_test = dlrm(X_test, lS_o_test, lS_i_test)
83 """Evaluates the metrics the sparsified metrics for the dlrm model on various sparsity levels,
|
| H A D | evaluate_forward_time.py | 103 print("Forward Time for Sparse DLRM")
109 print("Forward Time for Normal DLRM")
|
| /aosp_15_r20/external/pytorch/benchmarks/dynamo/ |
| H A D | all_torchbench_models_list.txt | 23 dlrm
|
| H A D | torchbench_models_list_cpu.txt | 6 dlrm,2048
|
| /aosp_15_r20/external/pytorch/benchmarks/dynamo/ci_expected_accuracy/ |
| H A D | dynamic_aot_eager_torchbench_training.csv | 61 dlrm,pass,6
|
| H A D | aot_eager_torchbench_training.csv | 61 dlrm,pass,6
|
| H A D | dynamic_inductor_torchbench_training.csv | 61 dlrm,pass,6
|
| H A D | inductor_torchbench_training.csv | 61 dlrm,pass,6
|
| H A D | dynamo_eager_torchbench_training.csv | 61 dlrm,pass,6
|
| H A D | cpu_aot_inductor_freezing_torchbench_inference.csv | 93 dlrm,pass,0
|
| H A D | dynamic_cpu_aot_inductor_amp_freezing_torchbench_inference.csv | 77 dlrm,pass,0
|
| H A D | dynamic_cpu_aot_inductor_freezing_torchbench_inference.csv | 77 dlrm,pass,0
|
| H A D | cpu_aot_inductor_amp_freezing_torchbench_inference.csv | 93 dlrm,pass,0
|
| H A D | cpu_inductor_amp_freezing_torchbench_inference.csv | 97 dlrm,pass,0
|
| H A D | dynamic_cpu_inductor_torchbench_inference.csv | 81 dlrm,pass,0
|
| H A D | cpu_inductor_freezing_torchbench_inference.csv | 97 dlrm,pass,0
|
| H A D | dynamic_inductor_torchbench_inference.csv | 105 dlrm,pass,0
|
| H A D | dynamic_aot_eager_torchbench_inference.csv | 105 dlrm,pass,0
|
| H A D | cpu_inductor_torchbench_inference.csv | 97 dlrm,pass,0
|
| H A D | aot_inductor_torchbench_inference.csv | 97 dlrm,pass,0
|
| H A D | inductor_torchbench_inference.csv | 105 dlrm,pass,0
|
| H A D | dynamo_eager_torchbench_inference.csv | 105 dlrm,pass,0
|
| H A D | aot_eager_torchbench_inference.csv | 105 dlrm,pass,0
|
| /aosp_15_r20/external/pytorch/benchmarks/dynamo/ci_expected_accuracy/cu124/ |
| H A D | aot_inductor_torchbench_inference.csv | 101 dlrm,pass,0
|
12