chore: update cifar training readme + accuracy eval (#572)

zama-ai · Apr 3, 2024 · 5306f6c · 5306f6c
1 parent 00dcfc8
commit 5306f6c
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 2 deletions.
diff --git a/use_case_examples/cifar/cifar_brevitas_training/README.md b/use_case_examples/cifar/cifar_brevitas_training/README.md
@@ -85,7 +85,7 @@ Here, an image from the CIFAR10 data-set is randomly chosen and preprocessed. Th
 
 ## Hardware Used for the Experiment
 
-Experiments were conducted on an m6i.metal machine offering 128 CPU cores and 512GB of memory. The choice of hardware can significantly influence the execution time and performance of the model.
+Experiments were conducted on an m6i.metal machine offering 128 CPU cores and 512GB of memory. The choice of hardware can significantly influence the execution time of the model.
 
 ## Accuracy and performance
 

diff --git a/use_case_examples/cifar/cifar_brevitas_training/evaluate_torch_cml.py b/use_case_examples/cifar/cifar_brevitas_training/evaluate_torch_cml.py
@@ -118,7 +118,9 @@ def main(args):
             input_set,
             n_bits={"model_inputs": 8, "model_outputs": 8},
             configuration=cfg,
-            rounding_threshold_bits=rounding_threshold_bits,
+            rounding_threshold_bits={"n_bits": rounding_threshold_bits, "method": "EXACT"}
+            if rounding_threshold_bits is not None
+            else None,
         )
 
         # Print max bit-width in the circuit