mirror of
https://github.com/GammaTauAI/reflexion-human-eval
synced 2024-11-11 19:10:53 +00:00
34 lines
1.1 KiB
Python
34 lines
1.1 KiB
Python
import matplotlib.pyplot as plt
|
|
|
|
def plot_bar_graph(data_dict):
|
|
# Get the labels and accuracy values from the dictionary
|
|
labels = list(data_dict.keys())
|
|
accuracies = list(data_dict.values())
|
|
|
|
# Set the custom colors for the first three bars using hex codes
|
|
other_color = '#8f99b5'
|
|
sota_color = '#3e465c'
|
|
|
|
# Create a list of colors for all the bars, using the custom colors for the first three bars
|
|
colors = [other_color for _ in range(len(labels) - 1)] + [sota_color]
|
|
|
|
# Create the bar plot
|
|
plt.bar(labels, accuracies, color=colors)
|
|
plt.xticks(fontweight="bold")
|
|
|
|
# Set the y-axis range and tick marks
|
|
plt.ylim(0, 1)
|
|
plt.yticks([i/10 for i in range(0, 11)], fontweight="bold")
|
|
|
|
# Add the accuracy labels above each bar
|
|
for i in range(len(labels)):
|
|
plt.text(i, accuracies[i]+0.02, str(round(accuracies[i], 2)), ha="center")
|
|
|
|
# Show the plot
|
|
plt.savefig("performance.png", dpi=500)
|
|
plt.show()
|
|
|
|
if __name__ == "__main__":
|
|
data_dict = {'PaLM': 0.262, 'CodeT+GPT-3.5': 0.658, 'GPT-4': 0.67, 'Reflexion+GPT-4': 0.88}
|
|
plot_bar_graph(data_dict)
|