update lab2 to new formats

lab2/main.py

@@ -0,0 +1,49 @@
+import logging
+import os
+
+from zigzag.api import get_hardware_performance_zigzag
+from zigzag.visualization.results.plot_cme import (
+    bar_plot_cost_model_evaluations_breakdown,
+)
+from zigzag.visualization.results.print_mapping import print_mapping
+
+
+# Initialize the logger
+logging_level = logging.INFO
+logging_format = "%(asctime)s - %(name)s.%(funcName)s +%(lineno)s - %(levelname)s - %(message)s"
+logging.basicConfig(level=logging_level, format=logging_format)
+
+# Define the experiment id and pickle name
+hw_name = "tpu_like"
+workload_name = "resnet18_first_layer"
+experiment_id = f"{hw_name}-{workload_name}"
+pickle_name = f"{experiment_id}-saved_list_of_cmes"
+
+# Define main input paths
+accelerator = "zigzag/inputs/hardware/tpu_like.yaml"
+workload = "lab2/resnet18_first_layer.onnx"
+mapping = "lab2/mapping.yaml"
+assert os.path.exists(mapping), "Copy mapping.yaml from lab1 and modify it for lab2."
+
+# Define other inputs of api call
+temporal_mapping_search_engine = "loma"
+optimization_criterion = "latency"
+dump_folder = f"lab2/outputs/{experiment_id}"
+pickle_filename = f"lab2/outputs/{pickle_name}.pickle"
+
+# Get the hardware performance through api call
+energy, latency, results = get_hardware_performance_zigzag(
+    accelerator=accelerator,
+    workload=workload,
+    mapping=mapping,
+    temporal_mapping_search_engine=temporal_mapping_search_engine,
+    opt=optimization_criterion,
+    dump_folder=dump_folder,
+    pickle_filename=pickle_filename,
+)
+
+# Save a bar plot of the cost model evaluations breakdown
+cmes = [result[0] for result in results[0][1]]
+save_path = "lab2/outputs/breakdown.png"
+bar_plot_cost_model_evaluations_breakdown(cmes, save_path=save_path)
+print_mapping(cmes[0])

lab2/solutions/mapping.yaml → lab2/mapping.yaml

@@ -1,23 +1,21 @@
-- name: /conv1/Conv
-  core_allocation: [1]
-  # spatial_mapping:
-  #   D1:
-  #     - K, 32
-  #   D2:
-  #     - C, 32
+- name: Conv1  # Name of the first conv layer in the onnx model
+  spatial_mapping:
+    D1:
+      - C, 32
+    D2:
+      - K, 32
   memory_operand_links:
     O: O
     W: I2
     I: I1
 
 # Default entry in case layer name and operator type are not available
 - name: default
-  core_allocation: [1]
   spatial_mapping:
     D1:
-      - K, 32
-    D2:
       - C, 32
+    D2:
+      - K, 32
   memory_operand_links:
     O: O
     W: I2