Tapas Base Masklm

🚀 谷歌表格问答模型（tapas_masklm_base_reset）

本模型对应于原始仓库中的 tapas_masklm_base_reset，它可用于表格问答等相关任务，能帮助我们从表格数据中获取所需信息。

🚀 快速开始

以下是使用该模型的具体步骤和示例代码：

from transformers import TapasTokenizer, TapasForMaskedLM
import pandas as pd
import torch

tokenizer = TapasTokenizer.from_pretrained("google/tapas-base-masklm")
model = TapasForMaskedLM.from_pretrained("google/tapas-base-masklm")

data = {'Actors': ["Brad Pitt", "Leonardo Di Caprio", "George Clooney"],
        'Age': ["56", "45", "59"],
        'Number of movies': ["87", "53", "69"]
}
table = pd.DataFrame.from_dict(data)
query = "How many movies has Leonardo [MASK] Caprio played in?"

# prepare inputs
inputs = tokenizer(table=table, queries=query, padding="max_length", return_tensors="pt")

# forward pass
outputs = model(**inputs)

# return top 5 values and predictions
masked_index = torch.nonzero(inputs.input_ids.squeeze() == tokenizer.mask_token_id, as_tuple=False)
logits = outputs.logits[0, masked_index.item(), :]
probs = logits.softmax(dim=0)
values, predictions = probs.topk(5)

for value, pred in zip(values, predictions):
  print(f"{tokenizer.decode([pred])} with confidence {value}")

💻 使用示例

基础用法

# 此代码展示了如何使用Tapas模型进行表格问答任务
# 包括加载模型和分词器、准备表格数据和查询、进行前向传播以及获取预测结果
from transformers import TapasTokenizer, TapasForMaskedLM
import pandas as pd
import torch

tokenizer = TapasTokenizer.from_pretrained("google/tapas-base-masklm")
model = TapasForMaskedLM.from_pretrained("google/tapas-base-masklm")

data = {'Actors': ["Brad Pitt", "Leonardo Di Caprio", "George Clooney"],
        'Age': ["56", "45", "59"],
        'Number of movies': ["87", "53", "69"]
}
table = pd.DataFrame.from_dict(data)
query = "How many movies has Leonardo [MASK] Caprio played in?"

# prepare inputs
inputs = tokenizer(table=table, queries=query, padding="max_length", return_tensors="pt")

# forward pass
outputs = model(**inputs)

# return top 5 values and predictions
masked_index = torch.nonzero(inputs.input_ids.squeeze() == tokenizer.mask_token_id, as_tuple=False)
logits = outputs.logits[0, masked_index.item(), :]
probs = logits.softmax(dim=0)
values, predictions = probs.topk(5)

for value, pred in zip(values, predictions):
  print(f"{tokenizer.decode([pred])} with confidence {value}")