Try searching for a different model name
Choose multiple models from the dropdown to see a side-by-side comparison
Try selecting different models
| Metric | ' for i, model_name in enumerate(selected_models): # Truncate long names short_name = model_name if len(model_name) <= 20 else model_name[:18] + "…" html += f'{short_name} | ' html += '|
|---|---|---|
| Average | ' valid_avgs_list = [model_avgs.get(m) for m in selected_models if model_avgs.get(m) is not None] max_avg_val = max(valid_avgs_list) if valid_avgs_list else None for model_name in selected_models: avg = model_avgs.get(model_name) if avg is not None: cell_class = "best" if avg == max_avg_val and len(valid_avgs_list) > 1 else "" html += f'{avg:.2f} | ' else: html += '— | ' html += '
| {metric_name} | ' # Get all scores for this metric metric_scores = {} for model_name in selected_models: if model_name in model_data_dict and leaderboard_name in model_data_dict[model_name]: results = model_data_dict[model_name][leaderboard_name].get("results", {}) metric_scores[model_name] = results.get(metric_name) valid_scores = [v for v in metric_scores.values() if v is not None] if valid_scores: max_score = max(valid_scores) min_score = min(valid_scores) score_range = max_score - min_score if max_score > min_score else 1 else: max_score = min_score = score_range = None for model_name in selected_models: score = metric_scores.get(model_name) if score is not None and score_range is not None: # Determine color class based on relative position if len(valid_scores) > 1: pct = (score - min_score) / score_range if score_range > 0 else 1 if score == max_score: cell_class = "best" elif pct >= 0.75: cell_class = "good" elif pct >= 0.5: cell_class = "mid" elif pct >= 0.25: cell_class = "low" else: cell_class = "worst" else: cell_class = "" html += f'{score:.2f} | ' else: html += '— | ' html += '