#0: Update UNet Shallow README

tenstorrent · Sep 14, 2024 · 2d0d38a · 2d0d38a
1 parent b4a026f
commit 2d0d38a
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 2 deletions.
diff --git a/README.md b/README.md
@@ -49,7 +49,7 @@
 | [ResNet-50 (224x224) (data parallel)](./models/demos/tgg/resnet50)          | 1024  | [Two Galaxies](https://tenstorrent.com/hardware/galaxy)  | 128,800 | 448,000    |             |
 | [ViT](./models/demos/grayskull/vit)                                         | 9     | [e150](https://tenstorrent.com/hardware/grayskull)       | 1,360   | 2,000      |             |
 | [Stable Diffusion 1.4 (512x512)](./models/demos/wormhole/stable_diffusion)  | 1     | [n150](https://tenstorrent.com/hardware/wormhole)        | 0.167   | 0.3        |             |
-| [Unet (shallow)](./models/experimental/functional_unet)                     | 2     | [n150](https://tenstorrent.com/hardware/wormhole)        | 51      | 1000       |             |
+| [Unet (shallow)](./models/experimental/functional_unet)                     | 2     | [n150](https://tenstorrent.com/hardware/wormhole)        | 190     | 1000       |             |
 
 ## NLPs
 | Model                                               | Batch | Hardware                                           | sen/sec   | Target sen/sec | Release     |

diff --git a/models/experimental/functional_unet/README.md b/models/experimental/functional_unet/README.md
@@ -5,13 +5,27 @@
 To run the demo, make sure to build the project, activate the environment, and set the appropriate environment variables.
 For more information, refer [installation and build guide](https://docs.tenstorrent.com/tt-metalium/latest/get_started/get_started.html#install-and-build).
 
-Use `pytest --disable-warnings models/experimental/functional_unet/tests/test_unet_model.py` to run the full UNet Shallow model.
+To run UNet Shallow for multiple iterations on single-chip at the best performance:
+
+```sh
+pytest --disable-warnings models/experimental/functional_unet/tests/test_unet_perf.py::test_unet_perf_e2e
+```
+
+To run UNet Shallow for multiple iterations on N300 and T3000 at the best performance:
+
+```sh
+pytest --disable-warnings models/experimental/functional_unet/tests/test_unet_perf.py::test_unet_data_parallel_perf_e2e
+````
+
+Use `pytest models/experimental/functional_unet/tests/test_unet_model.py` to run the functional UNet Shallow model on a single-chip.
 
 ## Supported Hardware
 
 - N150
 - N300
   - Make sure to place dispatch on ethernet cores with `export WH_ARCH_YAML=wormhole_b0_80_arch_eth_dispatch.yaml` for optimal performance
+- T3K
+  - Make sure to place dispatch on ethernet cores with `export WH_ARCH_YAML=wormhole_b0_80_arch_eth_dispatch.yaml` for optimal performance
 
 ## Other Details