使用numpy实现bert模型，使用hugging face 或pytorch训练模型，保存参数为numpy格式，然后使用numpy加载模型推理，可在树莓派上运行

之前分别用numpy实现了mlp，cnn，lstm，这次搞一个大一点的模型bert，纯numpy实现，最重要的是可在树莓派上或其他不能安装pytorch的板子上运行，推理数据

本次模型是随便在hugging face上找的一个新闻评论的模型，7分类

看这些模型参数，这并不重要，模型占硬盘空间都要400+M

bert.embeddings.word_embeddings.weight torch.Size([21128, 768])

bert.embeddings.position_embeddings.weight torch.Size([512, 768])

bert.embeddings.token_type_embeddings.weight torch.Size([2, 768])

bert.embeddings.LayerNorm.weight torch.Size([768])

bert.embeddings.LayerNorm.bias torch.Size([768])

bert.encoder.layer.0.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.0.attention.self.query.bias torch.Size([768])

bert.encoder.layer.0.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.0.attention.self.key.bias torch.Size([768])

bert.encoder.layer.0.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.0.attention.self.value.bias torch.Size([768])

bert.encoder.layer.0.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.0.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.0.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.0.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.0.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.0.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.0.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.0.output.dense.bias torch.Size([768])

bert.encoder.layer.0.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.0.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.1.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.1.attention.self.query.bias torch.Size([768])

bert.encoder.layer.1.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.1.attention.self.key.bias torch.Size([768])

bert.encoder.layer.1.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.1.attention.self.value.bias torch.Size([768])

bert.encoder.layer.1.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.1.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.1.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.1.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.1.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.1.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.1.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.1.output.dense.bias torch.Size([768])

bert.encoder.layer.1.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.1.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.2.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.2.attention.self.query.bias torch.Size([768])

bert.encoder.layer.2.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.2.attention.self.key.bias torch.Size([768])

bert.encoder.layer.2.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.2.attention.self.value.bias torch.Size([768])

bert.encoder.layer.2.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.2.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.2.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.2.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.2.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.2.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.2.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.2.output.dense.bias torch.Size([768])

bert.encoder.layer.2.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.2.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.3.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.3.attention.self.query.bias torch.Size([768])

bert.encoder.layer.3.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.3.attention.self.key.bias torch.Size([768])

bert.encoder.layer.3.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.3.attention.self.value.bias torch.Size([768])

bert.encoder.layer.3.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.3.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.3.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.3.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.3.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.3.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.3.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.3.output.dense.bias torch.Size([768])

bert.encoder.layer.3.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.3.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.4.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.4.attention.self.query.bias torch.Size([768])

bert.encoder.layer.4.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.4.attention.self.key.bias torch.Size([768])

bert.encoder.layer.4.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.4.attention.self.value.bias torch.Size([768])

bert.encoder.layer.4.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.4.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.4.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.4.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.4.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.4.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.4.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.4.output.dense.bias torch.Size([768])

bert.encoder.layer.4.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.4.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.5.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.5.attention.self.query.bias torch.Size([768])

bert.encoder.layer.5.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.5.attention.self.key.bias torch.Size([768])

bert.encoder.layer.5.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.5.attention.self.value.bias torch.Size([768])

bert.encoder.layer.5.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.5.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.5.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.5.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.5.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.5.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.5.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.5.output.dense.bias torch.Size([768])

bert.encoder.layer.5.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.5.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.6.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.6.attention.self.query.bias torch.Size([768])

bert.encoder.layer.6.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.6.attention.self.key.bias torch.Size([768])

bert.encoder.layer.6.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.6.attention.self.value.bias torch.Size([768])

bert.encoder.layer.6.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.6.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.6.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.6.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.6.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.6.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.6.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.6.output.dense.bias torch.Size([768])

bert.encoder.layer.6.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.6.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.7.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.7.attention.self.query.bias torch.Size([768])

bert.encoder.layer.7.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.7.attention.self.key.bias torch.Size([768])

bert.encoder.layer.7.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.7.attention.self.value.bias torch.Size([768])

bert.encoder.layer.7.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.7.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.7.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.7.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.7.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.7.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.7.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.7.output.dense.bias torch.Size([768])

bert.encoder.layer.7.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.7.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.8.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.8.attention.self.query.bias torch.Size([768])

bert.encoder.layer.8.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.8.attention.self.key.bias torch.Size([768])

bert.encoder.layer.8.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.8.attention.self.value.bias torch.Size([768])

bert.encoder.layer.8.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.8.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.8.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.8.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.8.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.8.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.8.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.8.output.dense.bias torch.Size([768])

bert.encoder.layer.8.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.8.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.9.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.9.attention.self.query.bias torch.Size([768])

bert.encoder.layer.9.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.9.attention.self.key.bias torch.Size([768])

bert.encoder.layer.9.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.9.attention.self.value.bias torch.Size([768])

bert.encoder.layer.9.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.9.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.9.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.9.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.9.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.9.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.9.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.9.output.dense.bias torch.Size([768])

bert.encoder.layer.9.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.9.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.10.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.10.attention.self.query.bias torch.Size([768])

bert.encoder.layer.10.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.10.attention.self.key.bias torch.Size([768])

bert.encoder.layer.10.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.10.attention.self.value.bias torch.Size([768])

bert.encoder.layer.10.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.10.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.10.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.10.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.10.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.10.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.10.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.10.output.dense.bias torch.Size([768])

bert.encoder.layer.10.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.10.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.11.attention.self.query.weight torch.Size([768, 768])

bert.encoder.layer.11.attention.self.query.bias torch.Size([768])

bert.encoder.layer.11.attention.self.key.weight torch.Size([768, 768])

bert.encoder.layer.11.attention.self.key.bias torch.Size([768])

bert.encoder.layer.11.attention.self.value.weight torch.Size([768, 768])

bert.encoder.layer.11.attention.self.value.bias torch.Size([768])

bert.encoder.layer.11.attention.output.dense.weight torch.Size([768, 768])

bert.encoder.layer.11.attention.output.dense.bias torch.Size([768])

bert.encoder.layer.11.attention.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.11.attention.output.LayerNorm.bias torch.Size([768])

bert.encoder.layer.11.intermediate.dense.weight torch.Size([3072, 768])

bert.encoder.layer.11.intermediate.dense.bias torch.Size([3072])

bert.encoder.layer.11.output.dense.weight torch.Size([768, 3072])

bert.encoder.layer.11.output.dense.bias torch.Size([768])

bert.encoder.layer.11.output.LayerNorm.weight torch.Size([768])

bert.encoder.layer.11.output.LayerNorm.bias torch.Size([768])

bert.pooler.dense.weight torch.Size([768, 768])

bert.pooler.dense.bias torch.Size([768])

classifier.weight torch.Size([7, 768])

classifier.bias torch.Size([7])

为了实现numpy的bert模型，踩了两天的坑，一步步对比huggingface源码实现的，真的太难了～～～

这是使用numpy实现的bert代码，分数上和huggingface有稍微的一点点区别，可能是模型太大，保存的模型参数误差累计造成的！

看下面的代码真的有利于直接了解bert模型结构，各种细节简单又到位，自己都服自己，研究这个东西～～～

import numpy as np

def word_embedding(input_ids, word_embeddings):

    return word_embeddings[input_ids]

def position_embedding(position_ids, position_embeddings):

    return position_embeddings[position_ids]

def token_type_embedding(token_type_ids, token_type_embeddings):

    return token_type_embeddings[token_type_ids]

def softmax(x, axis=None):

    # e_x = np.exp(x).astype(np.float32) #

    e_x = np.exp(x - np.max(x, axis=axis, keepdims=True))

    sum_ex = np.sum(e_x, axis=axis,keepdims=True).astype(np.float32)

    return e_x / sum_ex

def scaled_dot_product_attention(Q, K, V, mask=None):

    d_k = Q.shape[-1]

    scores = np.matmul(Q, K.transpose(0, 2, 1)) / np.sqrt(d_k)

    if mask is not None:

        scores = np.where(mask, scores, np.full_like(scores, -np.inf))

    attention_weights = softmax(scores, axis=-1)

    # print(attention_weights)

    # print(np.sum(attention_weights,axis=-1))

    output = np.matmul(attention_weights, V)

    return output, attention_weights

def multihead_attention(input, num_heads,W_Q,B_Q,W_K,B_K,W_V,B_V,W_O,B_O):

    q = np.matmul(input, W_Q.T)+B_Q

    k = np.matmul(input, W_K.T)+B_K

    v = np.matmul(input, W_V.T)+B_V

    # 分割输入为多个头

    q = np.split(q, num_heads, axis=-1)

    k = np.split(k, num_heads, axis=-1)

    v = np.split(v, num_heads, axis=-1)

    outputs = []

    for q_,k_,v_ in zip(q,k,v):

        output, attention_weights = scaled_dot_product_attention(q_, k_, v_)

        outputs.append(output)

    outputs = np.concatenate(outputs, axis=-1)

    outputs = np.matmul(outputs, W_O.T)+B_O

    return outputs

def layer_normalization(x, weight, bias, eps=1e-12):

    mean = np.mean(x, axis=-1, keepdims=True)

    variance = np.var(x, axis=-1, keepdims=True)

    std = np.sqrt(variance + eps)

    normalized_x = (x - mean) / std

    output = weight * normalized_x + bias

    return output

def feed_forward_layer(inputs, weight, bias, activation='relu'):

    linear_output = np.matmul(inputs,weight) + bias

    if activation == 'relu':

        activated_output = np.maximum(0, linear_output)  # ReLU激活函数

    elif activation == 'gelu':

        activated_output = 0.5 * linear_output * (1 + np.tanh(np.sqrt(2 / np.pi) * (linear_output + 0.044715 * np.power(linear_output, 3))))  # GELU激活函数

    elif activation == "tanh" :

        activated_output = np.tanh(linear_output)

    else:

        activated_output = linear_output  # 无激活函数

    return activated_output

def residual_connection(inputs, residual):

    # 残差连接

    residual_output = inputs + residual

    return residual_output

def tokenize_sentence(sentence, vocab_file = 'vocab.txt'):

    with open(vocab_file, 'r', encoding='utf-8') as f:

        vocab = f.readlines()

        vocab = [i.strip() for i in vocab]

        # print(len(vocab))

    tokenized_sentence = ['[CLS]'] + list(sentence) + ["[SEP]"] # 在句子开头添加[cls]

    token_ids = [vocab.index(token) for token in tokenized_sentence]

    return token_ids

# 加载保存的模型数据

model_data = np.load('bert_model_params.npz')

word_embeddings = model_data["bert.embeddings.word_embeddings.weight"]

position_embeddings = model_data["bert.embeddings.position_embeddings.weight"]

token_type_embeddings = model_data["bert.embeddings.token_type_embeddings.weight"]

def model_input(sentence):

    token_ids = tokenize_sentence(sentence)

    input_ids = np.array(token_ids)  # 输入的词汇id

    word_embedded = word_embedding(input_ids, word_embeddings)

    position_ids = np.array(range(len(input_ids)))  # 位置id

    # 位置嵌入矩阵，形状为 (max_position, embedding_size)

    position_embedded = position_embedding(position_ids, position_embeddings)

    token_type_ids = np.array([0]*len(input_ids))  # 片段类型id

    # 片段类型嵌入矩阵，形状为 (num_token_types, embedding_size)

    token_type_embedded = token_type_embedding(token_type_ids, token_type_embeddings)

    embedding_output = np.expand_dims(word_embedded + position_embedded + token_type_embedded, axis=0)

    return embedding_output

def bert(input,num_heads):

    ebd_LayerNorm_weight = model_data['bert.embeddings.LayerNorm.weight']

    ebd_LayerNorm_bias = model_data['bert.embeddings.LayerNorm.bias']

    input = layer_normalization(input,ebd_LayerNorm_weight,ebd_LayerNorm_bias)     #这里和模型输出一致

    for i in range(12):

        # 调用多头自注意力函数

        W_Q = model_data['bert.encoder.layer.{}.attention.self.query.weight'.format(i)]

        B_Q = model_data['bert.encoder.layer.{}.attention.self.query.bias'.format(i)]

        W_K = model_data['bert.encoder.layer.{}.attention.self.key.weight'.format(i)]

        B_K = model_data['bert.encoder.layer.{}.attention.self.key.bias'.format(i)]

        W_V = model_data['bert.encoder.layer.{}.attention.self.value.weight'.format(i)]

        B_V = model_data['bert.encoder.layer.{}.attention.self.value.bias'.format(i)]

        W_O = model_data['bert.encoder.layer.{}.attention.output.dense.weight'.format(i)]

        B_O = model_data['bert.encoder.layer.{}.attention.output.dense.bias'.format(i)]

        attention_output_LayerNorm_weight = model_data['bert.encoder.layer.{}.attention.output.LayerNorm.weight'.format(i)]

        attention_output_LayerNorm_bias = model_data['bert.encoder.layer.{}.attention.output.LayerNorm.bias'.format(i)]

        intermediate_weight = model_data['bert.encoder.layer.{}.intermediate.dense.weight'.format(i)]

        intermediate_bias = model_data['bert.encoder.layer.{}.intermediate.dense.bias'.format(i)]

        dense_weight = model_data['bert.encoder.layer.{}.output.dense.weight'.format(i)]

        dense_bias = model_data['bert.encoder.layer.{}.output.dense.bias'.format(i)]

        output_LayerNorm_weight = model_data['bert.encoder.layer.{}.output.LayerNorm.weight'.format(i)]

        output_LayerNorm_bias = model_data['bert.encoder.layer.{}.output.LayerNorm.bias'.format(i)]

        output = multihead_attention(input, num_heads,W_Q,B_Q,W_K,B_K,W_V,B_V,W_O,B_O)

        output = residual_connection(input,output)

        output1 = layer_normalization(output,attention_output_LayerNorm_weight,attention_output_LayerNorm_bias)    #这里和模型输出一致

        output = feed_forward_layer(output1, intermediate_weight.T, intermediate_bias, activation='gelu')

        output = feed_forward_layer(output, dense_weight.T, dense_bias, activation='')

        output = residual_connection(output1,output)

        output2 = layer_normalization(output,output_LayerNorm_weight,output_LayerNorm_bias)   #一致

        input = output2

    bert_pooler_dense_weight = model_data['bert.pooler.dense.weight']

    bert_pooler_dense_bias = model_data['bert.pooler.dense.bias']

    output = feed_forward_layer(output2, bert_pooler_dense_weight.T, bert_pooler_dense_bias, activation='tanh')    #一致

    return output

# for i in model_data:

#     # print(i)

#     print(i,model_data[i].shape)

id2label = {0: 'mainland China politics', 1: 'Hong Kong - Macau politics', 2: 'International news', 3: 'financial news', 4: 'culture', 5: 'entertainment', 6: 'sports'}

if __name__ == "__main__":

    # 示例用法

    sentence = '马拉松决赛'

    # print(model_input(sentence).shape)

    output = bert(model_input(sentence),num_heads=12)

    # print(output)

    classifier_weight = model_data['classifier.weight']

    classifier_bias = model_data['classifier.bias']

    output = feed_forward_layer(output[:,0,:], classifier_weight.T, classifier_bias, activation='')

    # print(output)

    output = softmax(output,axis=-1)

    label_id = np.argmax(output,axis=-1)

    label_score = output[0][label_id]

    print(id2label[label_id[0]],label_score)

这是hugging face上找的一个别人训练好的模型，roberta模型作新闻7分类，并且保存模型结构为numpy格式，为了上面的代码加载

import numpy as np

from transformers import AutoModelForSequenceClassification,AutoTokenizer,pipeline

model = AutoModelForSequenceClassification.from_pretrained('uer/roberta-base-finetuned-chinanews-chinese')

tokenizer = AutoTokenizer.from_pretrained('uer/roberta-base-finetuned-chinanews-chinese')

text_classification = pipeline('sentiment-analysis', model=model, tokenizer=tokenizer)

print(text_classification("马拉松决赛"))

# print(model)

# 打印BERT模型的权重维度

for name, param in model.named_parameters():

    print(name, param.data.shape)

# # # 保存模型参数为NumPy格式

model_params = {name: param.data.cpu().numpy() for name, param in model.named_parameters()}

np.savez('bert_model_params.npz', **model_params)

# model_params

对比两个结果：

hugging face：[{'label': 'sports', 'score': 0.9929242134094238}]

numpy：sports [0.9928773]

使用numpy实现bert模型，使用hugging face 或pytorch训练模型，保存参数为numpy格式，然后使用numpy加载模型推理，可在树莓派上运行的更多相关文章

深度学习原理与框架-猫狗图像识别-卷积神经网络(代码) 1.cv2.resize(图片压缩) 2..get_shape()[1:4].num_elements(获得最后三维度之和) 3.saver.save(训练参数的保存) 4.tf.train.import_meta_graph(加载模型结构) 5.saver.restore(训练参数载入)
1.cv2.resize(image, (image_size, image_size), 0, 0, cv2.INTER_LINEAR) 参数说明:image表示输入图片,image_size表示变 ...
【4】TensorFlow光速入门-保存模型及加载模型并使用
本文地址:https://www.cnblogs.com/tujia/p/13862360.html 系列文章: [0]TensorFlow光速入门-序 [1]TensorFlow光速入门-tenso ...
MindSpore保存与加载模型
技术背景近几年在机器学习和传统搜索算法的结合中,逐渐发展出了一种Search To Optimization的思维,旨在通过构造一个特定的机器学习模型,来替代传统算法中的搜索过程,进而加速经典图论等 ...
NeHe OpenGL教程第三十一课：加载模型
转自[翻译]NeHe OpenGL 教程前言声明,此 NeHe OpenGL教程系列文章由51博客yarin翻译(2010-08-19),本博客为转载并稍加整理与修改.对NeHe的OpenGL管线 ...
tensorflow学习笔记2：c++程序静态链接tensorflow库加载模型文件
首先需要搞定tensorflow c++库,搜了一遍没有找到现成的包,于是下载tensorflow的源码开始编译: tensorflow的contrib中有一个makefile项目,极大的简化的接下来 ...
PyTorch保存模型与加载模型+Finetune预训练模型使用
Pytorch 保存模型与加载模型 PyTorch之保存加载模型参数初始化参数的初始化其实就是对参数赋值.而我们需要学习的参数其实都是Variable,它其实是对Tensor的封装,同时提供了da ...
[Pytorch]Pytorch 保存模型与加载模型(转)
转自:知乎目录: 保存模型与加载模型冻结一部分参数,训练另一部分参数采用不同的学习率进行训练 1.保存模型与加载简单的保存与加载方法: # 保存整个网络 torch.save(net, PAT ...
tensorflowjs下载源文件到本地不能加载模型解决方案
大多数情况(非源文件错误)下载源文件到本地不能加载模型,那么你可能需要搭建一个本地WEB服务器. 1.安装apache或ngnix,可以参照这个博客 2.强烈推荐一个Chrome插件Web Serve ...
WPF 3D动态加载模型文件
原文:WPF 3D动态加载模型文件这篇文章需要读者对WPF 3D有一个基本了解,至少看过官方的MSDN例子. 一般来说关于WPF使用3D的例子,都是下面的流程: 1.美工用3DMAX做好模型,生成一 ...
[译]Vulkan教程(31)加载模型
[译]Vulkan教程(31)加载模型 Loading models 加载模型 Introduction 入门 Your program is now ready to render textured ...

随机推荐

[Linux/Apache Http]Apache Http(d)服务访问时报: 403 Forbidden You don't have permission to access /cdh/ on this server.
1 问题描述 http错误代码403:403 Forbidden 资源不可用.服务器理解客户的请求,但拒绝处理它.通常由于服务器上文件或目录的权限设置导致. 2 解决思路胜利的果实: 确保关闭sel ...
[Linux]浅析"command > /dev/null 2>&1 &" 与 "command 1>/dev/null 2>&1 &"
1 问题描述 1.1 问题描述在一项目中查看CENTOS 服务器的定时任务crontab时查看到如下这段命令: 命令clearLog.sh > /dev/null 2>&1 &a ...
C# 通过一个控制台打开另一个控制台
现有个需求是通过一个主程序获取配置的线程数和进程数打开连一个控制台程序,将线程数和系统编码作为参数传给控制台程序. 下面附上Demo. 1 private static void Main(strin ...
Java并发(一)----进程、线程、并行、并发
一.进程与线程进程程序由指令和数据组成,但这些指令要运行,数据要读写,就必须将指令加载至 CPU,数据加载至内存.在指令运行过程中还需要用到磁盘.网络等设备.进程就是用来加载指令.管理内存.管理 ...
一文吃透Arthas常用命令！
Arthas 常用命令简介 Arthas 是Alibaba开源的Java诊断工具,动态跟踪Java代码:实时监控JVM状态,可以在不中断程序执行的情况下轻松完成JVM相关问题排查工作 .支持JDK ...
USART串口_第三课
串口发送与接收 1.阻塞式发送 1.1.练习receive() 和Transmit() 测试1:测试接收发送函数 receive() 和 Transmit() 在main()中写入测试代码:将stm3 ...
docker安装python+nginx
一个容器安装python和nginx dockerfile FROM centos:7.9.2009 USER root RUN yum install gcc openssl-devel bzip2 ...
prometheus node-exporter安装
目录 prometheus node-exporter安装包安装在linux服务器 Daemonset安装 Docker安装 helm安装 prometheus node-exporter安装安装 ...
一个.Net版本的ChatGPT SDK
ChatGPT大火,用它来写代码.写表白书.写文章.写对联.写报告.写周边...... 啥都会! 个人.小公司没有能力开发大模型,但基于开放平台,根据特定的场景开发应用,却是非常火热的. 为了避免重复 ...
产品质量管理利器，华为云发布CodeArts Defect缺陷管理服务
摘要:近日,华为云CodeArts Defect缺陷管理服务正式上线,提供结构化缺陷跟踪流程和标准化的质量度量模型. 本文分享自华为云社区<产品质量管理利器,华为云发布CodeArts Defe ...

使用numpy实现bert模型，使用hugging face 或pytorch训练模型，保存参数为numpy格式，然后使用numpy加载模型推理，可在树莓派上运行

使用numpy实现bert模型，使用hugging face 或pytorch训练模型，保存参数为numpy格式，然后使用numpy加载模型推理，可在树莓派上运行的更多相关文章

随机推荐

热门专题