From 801d4c62067ea1284fc5c0f6477750bc2e28130c Mon Sep 17 00:00:00 2001
From: tianwei <liutianweidlut@gmail.com>
Date: Fri, 14 Jul 2023 14:24:21 +0800
Subject: [PATCH] hotfix(example): fix ag_news example evaluation typo (#2505)

---
 example/text_cls_AG_NEWS/tcan/evaluator.py    | 7 +++++--
 example/transformer/ag_news/code/evaluator.py | 9 ++++++---
 2 files changed, 11 insertions(+), 5 deletions(-)

diff --git a/example/text_cls_AG_NEWS/tcan/evaluator.py b/example/text_cls_AG_NEWS/tcan/evaluator.py
index 6470e1dcb5..2fe7a1dd9b 100644
--- a/example/text_cls_AG_NEWS/tcan/evaluator.py
+++ b/example/text_cls_AG_NEWS/tcan/evaluator.py
@@ -4,7 +4,7 @@
 import gradio
 from torchtext.data.utils import get_tokenizer, ngrams_iterator
 
-from starwhale import PipelineHandler, multi_classification
+from starwhale import Text, PipelineHandler, multi_classification
 from starwhale.api.service import api
 
 from .model import TextClassificationModel
@@ -24,7 +24,10 @@ def __init__(self) -> None:
 
     @torch.no_grad()
     def ppl(self, data: dict, **kw):
-        ngrams = list(ngrams_iterator(self.tokenizer(data["text"]), 2))
+        content = (
+            data["text"].content if isinstance(data["text"], Text) else data["text"]
+        )
+        ngrams = list(ngrams_iterator(self.tokenizer(content), 2))
         tensor = torch.tensor(self.vocab(ngrams)).to(self.device)
         output = self.model(tensor, torch.tensor([0]).to(self.device))
         pred_value = output.argmax(1).item()
diff --git a/example/transformer/ag_news/code/evaluator.py b/example/transformer/ag_news/code/evaluator.py
index f408c45d07..5be074f5d8 100644
--- a/example/transformer/ag_news/code/evaluator.py
+++ b/example/transformer/ag_news/code/evaluator.py
@@ -2,7 +2,7 @@
 
 from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassification
 
-from starwhale import PipelineHandler, multi_classification
+from starwhale import Text, PipelineHandler, multi_classification
 
 ROOTDIR = Path(__file__).parent.parent
 _LABEL_NAMES = ["LABEL_0", "LABEL_1", "LABEL_2", "LABEL_3"]
@@ -15,12 +15,15 @@ def __init__(self) -> None:
         model = AutoModelForSequenceClassification.from_pretrained(
             str(ROOTDIR / "models")
         )
-        self.mode = pipeline(
+        self.model = pipeline(
             task="text-classification", model=model, tokenizer=tokenizer
         )
 
     def ppl(self, data):
-        _r = self.mode(data["text"])
+        content = (
+            data["text"].content if isinstance(data["text"], Text) else data["text"]
+        )
+        _r = self.model(content)
         return _LABEL_NAMES.index(_r[0]["label"])
 
     @multi_classification(