Documentation Index
Fetch the complete documentation index at: https://docs.predibase.com/llms.txt
Use this file to discover all available pages before exploring further.
TrainingComputeSpec
The TrainingComputeSpec class provides configuration options for training compute.
General Hyperparameters
| Parameter | Type | Required | Default | Description |
|---|
region | string | Yes | | The region to use for training |
requests | TrainingComputeRequests | Yes | | The number of requests to use for training |
TrainingComputeRequests
The TrainingComputeRequests class contains specifications for training compute.
General Hyperparameters
| Parameter | Type | Required | Default | Description |
|---|
trainer | ComputeRequest | Yes | | The instance type to use for training |
ComputeRequest
The ComputeRequest class allows the user to request specific compute resources for training.
| Parameter | Type | Required | Default | Description |
|---|
sku | string | Yes | | The SKU to use for training |
Example Usage
from predibase import TrainingComputeSpec, TrainingComputeRequests, ComputeRequest
compute_spec = TrainingComputeSpec(
region="us-west-2",
requests=TrainingComputeRequests(
trainer=ComputeRequest(
sku="a100_80gb_100"
)
)
)
pb.finetuning.jobs.create(
config=SFTConfig(
base_model="qwen3-8b",
epochs=1,
rank=8,
target_modules=["q_proj", "v_proj", "k_proj"],
),
dataset="tldr_news",
compute_spec=compute_spec
)
