GenAIJob

Params

name

A human readable description of the object.

service

An instance of GenAIModel.

output_token_count

Generative ai model job output token count from e-footprint hypothesis in dimensionless.

Backwards links

• UsageJourneyStep

Calculated attributes

output_token_weights

ExplainableQuantity in kilobyte, representing the Generative ai model job output token weights.

Example value: 3.0 kilobyte

Depends directly on:

• Generative AI model job output token count from e-footprint hypothesis
• Number of bits per token from e-footprint hypothesis

through the following calculations:

You can also visit the link to Generative AI model job output token weights’s full calculation graph.

data_stored

ExplainableQuantity in kilobyte, representing the Generative ai model job data stored.

Example value: 103.0 kilobyte

Depends directly on:

• Generative AI model job output token weights

through the following calculations:

You can also visit the link to Generative AI model job data stored’s full calculation graph.

data_transferred

ExplainableQuantity in kilobyte, representing the Generative ai model job data transferred.

Example value: 103.0 kilobyte

Depends directly on:

• Generative AI model job output token weights

through the following calculations:

You can also visit the link to Generative AI model job data transferred’s full calculation graph.

request_duration

ExplainableQuantity in nanosecond, representing the Generative ai model job request duration.

Example value: 28154600000.0 nanosecond

Depends directly on:

• Generative AI model job output token count from e-footprint hypothesis
• GPU latency per active parameter and output token from Ecologits
• open-mistral-7b from mistralai nb of active parameters from Ecologits
• Base GPU latency per output_token from Ecologits

through the following calculations:

You can also visit the link to Generative AI model job request duration’s full calculation graph.

ram_needed

ExplainableQuantity in gigabyte, representing the No additional gpu ram needed because model is already loaded in memory from e-footprint hypothesis.

Example value: 0.0 gigabyte

Depends directly on:

through the following calculations:

You can also visit the link to No additional GPU RAM needed because model is already loaded in memory from e-footprint hypothesis’s full calculation graph.

compute_needed

ExplainableQuantity in gpu, representing the Generative ai model job nb of required gpus during inference.

Example value: 0.22 gpu

Depends directly on:

• Generative AI model ratio between GPU memory footprint and model size from Ecologits
• open-mistral-7b from mistralai nb of active parameters from Ecologits
• Generative AI model nb of bits per parameter from e-footprint hypothesis
• on premise GPU server RAM per GPU from Estimating the Carbon Footprint of BLOOM

through the following calculations:

You can also visit the link to Generative AI model job nb of required GPUs during inference’s full calculation graph.

hourly_occurrences_per_usage_pattern

Dictionary with UsagePattern as keys and Hourly generative ai model job occurrences in usage pattern as values, in occurrence.

Example value: {
35f108: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in occurrence:
first 10 vals [2.0, 6.0, 8.0, 1.0, 6.0, 5.0, 3.0, 5.0, 7.0, 5.0],
last 10 vals [2.0, 1.0, 6.0, 1.0, 9.0, 8.0, 3.0, 3.0, 1.0, 3.0],
}

Depends directly on:

• usage pattern UTC

through the following calculations:

You can also visit the link to Hourly Generative AI model job occurrences in usage pattern’s full calculation graph.

hourly_avg_occurrences_per_usage_pattern

Dictionary with UsagePattern as keys and Average hourly generative ai model job occurrences in usage pattern as values, in concurrent.

Example value: {
35f108: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in concurrent:
first 10 vals [0.02, 0.05, 0.06, 0.01, 0.05, 0.04, 0.02, 0.04, 0.05, 0.04],
last 10 vals [0.02, 0.01, 0.05, 0.01, 0.07, 0.06, 0.02, 0.02, 0.01, 0.02],
}

Depends directly on:

• Hourly Generative AI model job occurrences in usage pattern
• Generative AI model job request duration

through the following calculations:

You can also visit the link to Average hourly Generative AI model job occurrences in usage pattern’s full calculation graph.

hourly_data_transferred_per_usage_pattern

Dictionary with UsagePattern as keys and Hourly data transferred for generative ai model job in usage pattern as values, in concurrent * hour * kilobyte / nanosecond.

Example value: {
35f108: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in concurrent * h * kB / ns:
first 10 vals [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
last 10 vals [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
}

Depends directly on:

• Average hourly Generative AI model job occurrences in usage pattern
• Generative AI model job data transferred
• Generative AI model job request duration

through the following calculations:

You can also visit the link to Hourly data transferred for Generative AI model job in usage pattern’s full calculation graph.

hourly_data_stored_per_usage_pattern

Dictionary with UsagePattern as keys and Hourly data stored for generative ai model job in usage pattern as values, in concurrent * hour * kilobyte / nanosecond.

Example value: {
35f108: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in concurrent * h * kB / ns:
first 10 vals [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
last 10 vals [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
}

Depends directly on:

• Average hourly Generative AI model job occurrences in usage pattern
• Generative AI model job data stored
• Generative AI model job request duration

through the following calculations:

You can also visit the link to Hourly data stored for Generative AI model job in usage pattern’s full calculation graph.

hourly_avg_occurrences_across_usage_patterns

Hourly generative ai model job average occurrences across usage patterns in concurrent.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in concurrent:
first 10 vals [0.02, 0.05, 0.06, 0.01, 0.05, 0.04, 0.02, 0.04, 0.05, 0.04],
last 10 vals [0.02, 0.01, 0.05, 0.01, 0.07, 0.06, 0.02, 0.02, 0.01, 0.02]

Depends directly on:

• Average hourly Generative AI model job occurrences in usage pattern

through the following calculations:

You can also visit the link to Hourly Generative AI model job average occurrences across usage patterns’s full calculation graph.

hourly_data_transferred_across_usage_patterns

Hourly generative ai model job data transferred across usage patterns in concurrent * hour * kilobyte / nanosecond.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in concurrent * h * kB / ns:
first 10 vals [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
last 10 vals [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

Depends directly on:

• Hourly data transferred for Generative AI model job in usage pattern

through the following calculations:

You can also visit the link to Hourly Generative AI model job data transferred across usage patterns’s full calculation graph.

hourly_data_stored_across_usage_patterns

Hourly generative ai model job data stored across usage patterns in concurrent * hour * kilobyte / nanosecond.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in concurrent * h * kB / ns:
first 10 vals [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
last 10 vals [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

Depends directly on:

• Hourly data stored for Generative AI model job in usage pattern

through the following calculations:

You can also visit the link to Hourly Generative AI model job data stored across usage patterns’s full calculation graph.