AI

# integrates Baidu’s Wenxin Qianfan large model into WordPress and compares it with GPT

Last weekend, Baidu’s Wenxin Qianfan model was integrated into WordPress.
1. Basic process:
1. First apply to experience the Wenxin Qianfan large model on Baidu, which requires certification first.

2. After passing, activate the large model, because Baidu's large model is charged for use and is charged by token. You need to have a balance in your account to activate it.

3. Then, create an application so that it has appid, api key and secret key

4. Then, read the documentation and connect the corresponding interface.
Basically, you get the access token through the api key and secret key, and then submit the question and get the answer.

2. Key code
1. Obtain the key code of the access token

"`php
private function getAccessToken(){
$curl = curl_init();
curl_setopt_array($curl, array(
CURLOPT_URL => “https://aip.baidubce.com/oauth/2.0/token?client_id=”.$this->client_id.”&client_secret=”.$this->client_secret.”&grant_type=client_credentials”,
CURLOPT_TIMEOUT => 30,
CURLOPT_RETURNTRANSFER => true,
CURLOPT_CUSTOMREQUEST => 'POST',
CURLOPT_HTTPHEADER => array(
'Content-Type: application/json',
'Accept: application/json'
),

));
$response = curl_exec($curl);
curl_close($curl);
$rtn = json_decode($response);
return $rtn->access_token;
}
“`

2. Call the key code of Ernie Bot large model

"`php
public function runErnieBot($message) {
$curl = curl_init();
curl_setopt_array($curl, array(
CURLOPT_URL => “https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/completions?access_token={$this->getAccessToken()}”,
CURLOPT_TIMEOUT => 30,
CURLOPT_RETURNTRANSFER => true,
CURLOPT_CUSTOMREQUEST => 'POST',
CURLOPT_POSTFIELDS =>$message,
CURLOPT_HTTPHEADER => array(
'Content-Type: application/json'
),
));
$response = curl_exec($curl);
curl_close($curl);
return $response;
}
“`

3. Call the key code of Ernie Bot Turbo large model

"`php
public function runErnieBotTurbo($message) {
$curl = curl_init();
curl_setopt_array($curl, array(
CURLOPT_URL => “https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/eb-instant?access_token={$this->getAccessToken()}”,
CURLOPT_TIMEOUT => 30,
CURLOPT_RETURNTRANSFER => true,
CURLOPT_CUSTOMREQUEST => 'POST',
CURLOPT_POSTFIELDS =>$message,
CURLOPT_HTTPHEADER => array(
'Content-Type: application/json'
),
));
$response = curl_exec($curl);
curl_close($curl);
return $response;
}
“`

In the past few days of testing, the performance of Baidu Wenxin Qianfan's large model in Chinese is indeed much better than GPT.
The Chinese level of GPT is the level of "explanatory text".
The Chinese level of Baidu Wenxin Qianfan Large Model is at least better than the "explanatory text".