massquantity
massquantity
Try calling `model.set_embeddings()` after `rebuild_model` and before `predict`.
Hi, unfortunately, this feature is not supported. We will consider adding it.
Is it possible that the EC2 server encounters some peak period from time to time, so the computing resources become limited?
That's what the AWS advertisement suggests. Based on your description, I couldn't think of any other reasons.
_**This is the ChatGPT answer to `why dedicated instance in ec2 becomes slow`**_: There are several reasons why a dedicated instance in Amazon EC2 might become slow at times. Even...
Yes, `DatasetPure.build_trainset` will encode the data. The `predict` function will convert the original user and item internally, so you do not need to map it by yourself.
看上去是数据类型的问题，你可以贴一下完整的报错信息，或者在这里打印 `data`， https://github.com/massquantity/DBRL/blob/master/dbrl/trainer/pretrain.py#L22 看是哪个数据变成了`torch.cuda.IntTensor`
feat 是特征，比如数据集里 age, brand 之类的
推荐的流程是 action 向量和所有的 item 向量计算相似度，选择相似度最大的 K 个 item。将两个向量 norm 后再点乘就是计算 cosine 相似度，这个在推荐系统里比较常用。这个仓库里的是对论文的一个实现，又无意严格按照论文的所有设定。
使用的数据是来自淘宝的，pv 是浏览，cart 是加入购物车，fav 是加入喜欢，buy 是购买。reward 是随意定的，RL 里的 reward 定义本身就是难点。