phoenix-oss/llama-stack
1
0
Fork 0
forked from phoenix-oss/llama-stack-mirror
Code Pull requests Releases Packages 2 Activity Actions

[Evals API][11/n] huggingface dataset provider + mmlu scoring fn (#392)

Browse source 
* wip

* scoring fn api

* eval api

* eval task

* evaluate api update

* pre commit

* unwrap context -> config

* config field doc

* typo

* naming fix

* separate benchmark / app eval

* api name

* rename

* wip tests

* wip

* datasetio test

* delete unused

* fixture

* scoring resolve

* fix scoring register

* scoring test pass

* score batch

* scoring fix

* fix eval

* test eval works

* huggingface provider

* datasetdef files

* mmlu scoring fn

* test wip

* remove type ignore

* api refactor

* add default task_eval_id for routing

* add eval_id for jobs

* remove type ignore

* huggingface provider

* wip huggingface register

* only keep 1 run_eval

* fix optional

* register task required

* register task required

* delete old tests

* fix

* mmlu loose

* refactor

* msg

* fix tests

* move benchmark task def to file

* msg

* gen openapi

* openapi gen

* move dataset to hf llamastack repo

* remove todo

* refactor

* add register model to unit test

* rename

* register to client

* delete preregistered dataset/eval task

* comments

* huggingface -> remote adapter

* openapi gen
This commit is contained in:
Xi Yan 2024-11-11 14:49:50 -05:00 • committed by GitHub
parent b78ee3a0a5
commit 2b7d70ba86
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
20 changed files with 1607 additions and 718 deletions
Download patch file Download diff file Expand all files Collapse all files

33
llama_stack/providers/inline/meta_reference/datasetio/datasetio.py
View file

@ -3,20 +3,17 @@
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
import io
from typing import List, Optional
import pandas
from llama_models.llama3.api.datatypes import * # noqa: F403
from llama_stack.apis.datasetio import * # noqa: F403
import base64
from abc import ABC, abstractmethod
from dataclasses import dataclass
from urllib.parse import unquote
from llama_stack.providers.datatypes import DatasetsProtocolPrivate
from llama_stack.providers.utils.memory.vector_store import parse_data_url
from llama_stack.providers.utils.datasetio.url_utils import get_dataframe_from_url
from .config import MetaReferenceDatasetIOConfig
@ -73,31 +70,9 @@ class PandasDataframeDataset(BaseDataset):
if self.df is not None:
return
# TODO: more robust support w/ data url
if self.dataset_def.url.uri.endswith(".csv"):
df = pandas.read_csv(self.dataset_def.url.uri)
elif self.dataset_def.url.uri.endswith(".xlsx"):
df = pandas.read_excel(self.dataset_def.url.uri)
elif self.dataset_def.url.uri.startswith("data:"):
parts = parse_data_url(self.dataset_def.url.uri)
data = parts["data"]
if parts["is_base64"]:
data = base64.b64decode(data)
else:
data = unquote(data)
encoding = parts["encoding"] or "utf-8"
data = data.encode(encoding)
mime_type = parts["mimetype"]
mime_category = mime_type.split("/")[0]
data_bytes = io.BytesIO(data)
if mime_category == "text":
df = pandas.read_csv(data_bytes)
else:
df = pandas.read_excel(data_bytes)
else:
raise ValueError(f"Unsupported file type: {self.dataset_def.url}")
df = get_dataframe_from_url(self.dataset_def.url)
if df is None:
raise ValueError(f"Failed to load dataset from {self.dataset_def.url}")
self.df = self._validate_dataset_schema(df)