🖥️ Local Hardware
Precision
{{ precision | upper }}
Avg Power
{{ mean_power_w }}W
Est. Local Total
{{ est_total_time }}
📏 Benchmark Config
Total Epochs
{{ total_epochs }}
Benchmark Epochs
{{ benchmark_epochs + warmup_epochs }} ({{ warmup_epochs }} warmup)
Data Sampled
{{ sample_data_pct }}%
Location
{{ country_code }}
{% if best_value %}
📊 Savings Overview — Crusoe vs Local
⏱️
{{ time_saved }}
Time saved
💰
{{ cost_diff_fmt }}
{{ cost_diff_label }}
🌍
{{ est_total_co2 }}
CO₂ saved (100%)
⚡
{{ best_value.speedup }}x
Faster ({{ best_value.name }})
{% endif %}
{% if best_value or fastest %}
💡 Recommendation
{% if best_value %}
💰 Best Value
{{ best_value.name }}
Time
{{ best_value_time }}
Cost
{{ best_value_cost }}
Speedup
{{ best_value.speedup }}x
Range
{{ best_value_time_lower }}–{{ best_value_time_upper }}
{% endif %}
{% if fastest %}
⚡ Fastest
{{ fastest.name }}
Speedup
{{ fastest.speedup }}x
Range
{{ fastest_time_lower }}–{{ fastest_time_upper }}
{% endif %}
{% endif %}
{% if crusoe_rows %}
☁️ Crusoe Cloud GPU Comparison
| GPU |
Est. Training Time |
Speedup |
Est. Cost |
CO₂ Emissions |
| 📍 {{ gpu_name }} (local) |
{{ est_total_time }}
({{ est_total_time_lower }}–{{ est_total_time_upper }})
|
1.0x |
{{ est_total_cost }}
(electricity)
|
{{ est_total_co2 }}
|
{% for row in crusoe_rows %}
|
{{ row.name }}
{% if row.is_best_value %}💰 Best Value{% endif %}
{% if row.is_fastest %}⚡ Fastest{% endif %}
|
{{ row.time_fmt }}
({{ row.time_lower_fmt }}–{{ row.time_upper_fmt }})
|
{{ row.speedup }}x
|
{{ row.cost_fmt }}
({{ row.cost_lower_fmt }}–{{ row.cost_upper_fmt }})
({{ row.rate_fmt }}/hr)
|
≈ 0 g CO₂
-100%
|
{% endfor %}
{% endif %}
� Carbon Savings on Crusoe
By training on Crusoe Cloud instead of locally, you would save approximately
{{ est_total_co2 }} of CO₂ emissions.
That's equivalent to:
{% if co2_equivalences %}
{% for eq in co2_equivalences %}
{{ eq.icon }}
{{ eq.value_fmt }}
{{ eq.unit }}
{{ eq.description }}
{% endfor %}
{% endif %}
⏱️ Epoch Timing (Benchmark)
{% for ep in epoch_bars %}
{{ ep.label }}
{{ ep.time_fmt }}
{% endfor %}
Median: {{ median_epoch_time }}s
| Std: {{ std_epoch_time }}s
| CV: {{ cv_epoch_time }}%
({{ variance_rating }})
{% if sample_data_pct_val < 100 %}
| Est. full-data epoch: {{ est_full_epoch_time }}
{% endif %}