phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-31 05:53:53 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge remote-tracking branch 'origin/main' into if_eval

Browse source
This commit is contained in:
Botao Chen 2025-03-19 15:17:05 -07:00
commit 507e4e17e6
8 changed files with 289 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

12
llama_stack/templates/open-benchmark/open_benchmark.py
View file

@ -210,6 +210,13 @@ def get_distribution_template() -> DistributionTemplate:
uri="huggingface://datasets/llamastack/IfEval?split=train",
),
),
DatasetInput(
dataset_id="docvqa",
purpose=DatasetPurpose.eval_messages_answer,
source=URIDataSource(
uri="huggingface://datasets/llamastack/docvqa?split=val",
)
)
]
default_benchmarks = [
@ -243,6 +250,11 @@ def get_distribution_template() -> DistributionTemplate:
dataset_id="ifeval",
scoring_functions=["basic::ifeval"],
),
BenchmarkInput(
benchmark_id="meta-reference-docvqa",
dataset_id="docvqa",
scoring_functions=["basic::docvqa"],
)
]
return DistributionTemplate(
name=name,