For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Python class

PenaltyInputs

PenaltyInputs

class max.pipelines.sampling.PenaltyInputs(frequency_data, frequency_penalty, presence_penalty, repetition_penalty)

source

Bases: object

Container for penalty inputs.

Parameters:

• frequency_data (list[FrequencyData])
• frequency_penalty (Buffer)
• presence_penalty (Buffer)
• repetition_penalty (Buffer)

as_list()

as_list()

source

Returns the penalty inputs as a list of buffers.

Return type:

list[Buffer]

create()

classmethod create(batch, device, num_steps=1)

source

Create penalty input tensors from context batch.

Parameters:

• batch (list[TextContext]) – List of context objects containing sampling parameters.
• device (Device) – Device to place the tensors on.
• num_steps (int) – Number of generation steps for frequency CSR padding.

Returns:

PenaltyInputs containing the penalty input tensors

Return type:

Self

frequency_data

frequency_data: list[FrequencyData]

source

frequency_penalty

frequency_penalty: Buffer

source

presence_penalty

presence_penalty: Buffer

source

repetition_penalty

repetition_penalty: Buffer

source

Was this page helpful?

Thank you! We'll create more content like this.

Thank you for helping us improve!

😔 What went wrong?

Some code doesn’t work

It includes inaccurate information

It's missing information I need

It was difficult to understand

Other

Submit