init

app.py

@@ -0,0 +1,32 @@
+import gradio as gr
+import torch
+from transformers import AutoTokenizer, RobertaForQuestionAnswering
+
+tokenizer = AutoTokenizer.from_pretrained("tsmatz/roberta_qa_japanese")
+model = RobertaForQuestionAnswering.from_pretrained("tsmatz/roberta_qa_japanese")
+
+def answer(text, question):
+  inputs = tokenizer(question, text, add_special_tokens=True, return_tensors="pt")
+  input_ids = inputs["input_ids"].tolist()[0]
+
+  outputs = model(**inputs)
+  answer_start_scores = outputs.start_logits
+  answer_end_scores = outputs.end_logits
+
+  answer_start = torch.argmax(answer_start_scores)
+  answer_end = torch.argmax(answer_end_scores) + 1
+
+  answer = tokenizer.convert_tokens_to_string(
+      tokenizer.convert_ids_to_tokens(input_ids[answer_start:answer_end])
+  )
+
+  return answer
+
+gr.Interface(
+    answer,
+    [
+        gr.Textbox(label="Text:", placeholder="Text...", lines=5),
+        gr.Textbox(label="Question:", placeholder="Question...", lines=1)
+    ],
+    "text",
+).launch()

requirements.txt

@@ -0,0 +1,10 @@
+torch
+datasets
+evaluate
+transformers[sentencepiece]
+transformers
+sentencepiece
+fugashi
+fugashi[unidic]
+fugashi[unidic-lite]
+ipadic