SafetensorsLLMInContextGenerator

Guide & worked example · Source

SafetensorsLLMInContextGenerator(
    model_path,
    tokenizer_path=None,
    examples=32,
    max_new_tokens=512,
    temperature=0.7,
    do_sample=True,
    load_in_4bit=False,
    max_attempts=20,
    rows_per_prompt=None,
    progress_path=None,
    device=None,
)

Generate CSV-like tabular rows from a local safetensors language model.

This wrapper is intentionally thin. It keeps model choice, prompt budget, and parsing policy visible to the benchmark instead of hiding them behind a broad synthetic-data abstraction.

Attributes

Name	Description
base_model_path
device
do_sample
examples
load_in_4bit
max_attempts
max_new_tokens
model_path
params
progress_path
rows_per_prompt
temperature
tokenizer_path

Methods

Name	Description
fit
sample
summary	Print a summary of the model results.
to	Move all model parameters to specified device.

fit

SafetensorsLLMInContextGenerator.fit(X, column_names=None)

sample

SafetensorsLLMInContextGenerator.sample(n)

summary

SafetensorsLLMInContextGenerator.summary()

Print a summary of the model results.

to

SafetensorsLLMInContextGenerator.to(device)

Move all model parameters to specified device.

Parameters

Name	Type	Description	Default
device	torch.device or str	Target device (‘cuda’, ‘cpu’, etc.)	required

Returns

Name	Type	Description
	BaseEstimator	Self for method chaining