A set of parameters are supported by API to help users provide additional details per project.
| Parameter | Description |
|---|---|
| project_name | Name of the project, defaults to codecarbon |
| experiment_id | Id of the experiment |
| measure_power_secs | Interval (in seconds) to measure hardware power usage, defaults to 15 |
| tracking_mode | machine measure the power consumptions of the entire machine (default)process try and isolate the tracked processes in isolation |
| gpu_ids | Comma-separated list of GPU ids to track, defaults to
NoneThese can either be integer indexes of GPUs on the system, or prefixes
to match against GPU identifiers as described here
|
| log_level | Global codecarbon log level (by order of verbosity): "debug", "info" (default),
"warning", "error", or "critical"
|
| co2_signal_api_token | API token for co2signal.com (requires sign-up for free beta)
|
| pue | PUE (Power Usage Effectiveness) of the data center
where the experiment is being run.
|
| force_cpu_power | Force the CPU max power consumption in watts,
use this if you know the TDP of your machine.
(POWER_CONSTANT x CONSUMPTION_PERCENTAGE)
|
| force_ram_power | Force the RAM power consumption in watts,
use this if you know the power consumption of your RAM.
Estimate it with
sudo lshw -C memory -short | grep DIMMto get the number of RAM slots used, then do
RAM power in W = Number of RAM Slots * 5 Watts
|
| force_mode_cpu_load | Force the use of CPU load mode for measuring CPU power consumption,
defaults to
False. When enabled, uses psutil to monitor CPU loadand estimates power consumption based on TDP and current CPU usage.
|
| force_mode_constant | Force the use of constant mode for CPU power consumption measurement,
defaults to
False. When enabled, bypasses psutil completely anduses a constant power consumption based on CPU TDP. Useful when
psutil overhead is significant or psutil is unavailable.
|
| allow_multiple_runs | Boolean variable indicating if multiple instance of CodeCarbon
on the same machine is allowed,
defaults to
True since v3. Used to be False in v2. |
PUE is a multiplication factor provided by the user, so it is up to the user to get it from their cloud provider. Old data-centers have a PUE up to 2.2, where new greener ones could be as low as 1.1.
If you, or your provider, use CUDA_VISIBLE_DEVICES to set the GPUs you want to use, CodeCarbon will automaticly populate this value into gpu_ids.
If you set gpu_ids manually, it will override the CUDA_VISIBLE_DEVICES for CodeCarbon measures.
| Parameter | Description |
|---|---|
| save_to_file | Boolean variable indicating if the emission artifacts should be logged
to a CSV file, defaults to
True |
| output_dir | Directory path to which the experiment details are logged
defaults to current directory
|
| output_file | Name of output CSV file
defaults to
emissions.csv |
| on_csv_write | When calling
tracker.flush() manually choose if-
update the existing run_id row (erasing former data)-
append add a new row to CSV file (defaults) |
| save_to_api | Boolean variable indicating if emissions artifacts should be logged
to the CodeCarbon API, defaults to
False |
| api_endpoint: | Optional URL of CodeCarbon API endpoint for sending emissions data
defaults to "https://api.codecarbon.io"
|
| api_key | API key for the CodeCarbon API (mandatory to use this API!) |
| api_call_interval | Number of measurements between API calls (defaults to 8):
-1 : call API on flush() and at the end
1 : at every measure
2 : at every 2 measure, and so on
|
| save_to_logger | Boolean variable indicating if the emission artifacts should be written
to a dedicated logger, defaults to
False |
| logging_logger | LoggerOutput object encapsulating a logging.Logger or a Google Cloud logger |
| logger_preamble | String to systematically include in the logger's messages (defaults to "") |
| save_to_prometheus | Boolean variable indicating if the emission artifacts should be written
to a Prometheus server, defaults to
False |
| prometheus_url | URL of the Prometheus server
|
| save_to_logfire | Boolean variable indicating if the emission artifacts should be written
to a LogFire server, defaults to
False |
| output_handlers | List of output handlers to use for saving the emissions data
defaults to
[] |
| Parameter | Description |
|---|---|
| country_iso_code | 3-letter ISO Code of the country
where the experiment is being run.
Available countries are listed in global_energy_mix.json
|
| region | Optional name of the Province/State/City, where the infrastructure is hosted
Currently, supported only for US States and Canada
for example - California or New York, from the list
|
| cloud_provider | The cloud provider specified for estimating emissions intensity,
defaults to
None. See impact.csv for a list of cloud providers |
| cloud_region | The region of the cloud data center, defaults to
NoneSee impact.csv for a list of cloud regions.
|
| country_2letter_iso_code | For use with the Electricity Maps emissions API.
See Electricity Maps zones for a list of codes and their locations.
|
Decorator track_emissions in addition to standard arguments, requires the following parameters:
| Parameter | Description |
|---|---|
| fn | function to be decorated |
| offline | Boolean variable indicating if the tracker should be run in offline mode
defaults to
False |
| country_iso_code | 3 letter ISO Code of the country where the experiment is being run.
Available countries are listed in global_energy_mix.json
|
| region | Optional Name of the Province/State/City, where the infrastructure is hosted
Currently, supported only for US States
for example - California or New York, from the list
|
| cloud_provider | The cloud provider specified for estimating emissions intensity,
defaults to
None. See impact.csv for a list of cloud providers |
| cloud_region | The region of the cloud data center, defaults to
NoneSee impact.csv for a list of cloud regions.
|