dpnmw/agents
1
0
Fork 0
mirror of https://github.com/wshobson/agents.git synced 2026-03-18 09:37:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

⚡ Bolt: optimize prompt evaluation loop to skip redundant calls (#152)

Browse Source 
- Avoid re-evaluating the current prompt if metrics are already available from the previous iteration.
- Pass metrics from the best variation to the next iteration.
- Reduces N-1 expensive LLM calls in an N-iteration optimization loop.

Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
This commit is contained in:
google-labs-jules[bot]
2025-12-21 19:02:37 -05:00
committed by GitHub
parent fda45604b7
commit a86384334b
1 changed files with 10 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
plugins/llm-application-dev/skills/prompt-engineering-patterns/scripts/optimize-prompt.py
View File

@@ -106,12 +106,18 @@ class PromptOptimizer:
current_prompt = base_prompt current_prompt = base_prompt
best_prompt = base_prompt best_prompt = base_prompt
best_score = 0 best_score = 0
current_metrics = None
for iteration in range(max_iterations): for iteration in range(max_iterations):
print(f"\nIteration {iteration + 1}/{max_iterations}") print(f"\nIteration {iteration + 1}/{max_iterations}")
# Evaluate current prompt # Evaluate current prompt
metrics = self.evaluate_prompt(current_prompt) # Bolt Optimization: Avoid re-evaluating if we already have metrics from previous iteration
if current_metrics:
metrics = current_metrics
else:
metrics = self.evaluate_prompt(current_prompt)
print(f"Accuracy: {metrics['avg_accuracy']:.2f}, Latency: {metrics['avg_latency']:.2f}s") print(f"Accuracy: {metrics['avg_accuracy']:.2f}, Latency: {metrics['avg_latency']:.2f}s")
# Track results # Track results
@@ -137,14 +143,17 @@ class PromptOptimizer:
# Test variations and pick best # Test variations and pick best
best_variation = current_prompt best_variation = current_prompt
best_variation_score = metrics['avg_accuracy'] best_variation_score = metrics['avg_accuracy']
best_variation_metrics = metrics
for variation in variations: for variation in variations:
var_metrics = self.evaluate_prompt(variation) var_metrics = self.evaluate_prompt(variation)
if var_metrics['avg_accuracy'] > best_variation_score: if var_metrics['avg_accuracy'] > best_variation_score:
best_variation_score = var_metrics['avg_accuracy'] best_variation_score = var_metrics['avg_accuracy']
best_variation = variation best_variation = variation
best_variation_metrics = var_metrics
current_prompt = best_variation current_prompt = best_variation
current_metrics = best_variation_metrics
return { return {
'best_prompt': best_prompt, 'best_prompt': best_prompt,