fanqie03
fanqie03
How to use machine learning? The result is a little strange. Did I use it wrong? 
Hello @KonstantinKorotaev Another question is if i use bicycle tag to auto rectangle. The front end will turn all labels into bicycles. But the back end gives different labels. I...
Why remove it?
哪个结果？？是哪一层结果吗？
现在出来的效果是不同的人，模型认为是同一个人的相似度都很高
我写了个和Tengine debug类似的脚本，可以输出mxnet每一层的结果 https://blog.csdn.net/night_mfc/article/details/108219908 希望对你有所帮助
I meet some problem ` 2018-09-15 13:57:42 [scrapy.core.engine] DEBUG: Crawled (201) (referer: None) 2018-09-15 13:57:42 [scrapy.pipelines.files] WARNING: File (code: 201): Error downloading file from referred in ` Browser can open...
my deal is use myself pipeline. ` from concurrent.futures import ThreadPoolExecutor class MyImagePipeline(): def __init__(self, file_path, pool_size=None): self.file_path=file_path self.executor = ThreadPoolExecutor(pool_size) @classmethod def from_crawler(cls,crawler): return cls(crawler.settings.get('IMAGES_STORE'), crawler.settings.get('POOL_SIZE')) def process_item(self, item,...
根据crnn的特点倒退出字符位置，有训练好的模型可以参考下倒推的逻辑实现 https://github.com/fanqie03/char-detection/issues/1#issuecomment-1553912204
从这里开始 [ result = converter.decode(preds, length, raw=True)](https://github.com/fanqie03/char-detection/blob/6f0f8e9f53ede6da690d7acd33585d2b5fe9ad0e/crnn/CRNN.py#LL189C30-L189C30) 基本原理是基于识别结果倒推一下单字位置