lucifer

Results 7 issues of lucifer

### Is there an existing issue for this? - [X] I have searched the existing issues. ### Is your feature request related to a problem? Please describe. 想用一些中文CLIP模型,但towhee目前似乎不支持自定义模型哦 ### Describe...

kind/feature

### Is there an existing issue for this? - [X] I have searched the existing issues ### Current Behavior 作者您好~ADGEN 训练集里有近12w条数据,按照您提供的参数,step=3000,per_device_train_batch_size=16,一个epoch只能覆盖3000*16=48000条数据? 是我理解的有问题吗?还是有其他的手段保证数据集上的full training ### Expected Behavior _No response_ ### Steps...

### Is there an existing issue for this? - [X] I have searched the existing issues ### Current Behavior evaluate.sh内容: PRE_SEQ_LEN=128 CHECKPOINT=viewgen0421-chatglm-6b-pt-128-2e-2 STEP=5000 CUDA_VISIBLE_DEVICES=1 python3 main.py \ --do_predict \ --validation_file...

### Is there an existing issue for this? - [X] I have searched the existing issues and checked the recent builds/commits ### What happened? *** Error setting up CodeFormer Traceback...

bug-report

您好!您在处理数组下标的时候存在不一致的问题: 假设单字符作为aspect或opnion [processor.py](https://github.com/wireless911/span-aste/blob/e158fea27c0926944bf029eefa50b569eb9658f9/utils/processor.py#L140C3-L140C111) 这里通过a2 + 1,实际上的a_end_idx=a_start_idx+1,但 https://github.com/wireless911/span-aste/blob/e158fea27c0926944bf029eefa50b569eb9658f9/models/model.py#L60C3-L60C3 这里的逻辑里单字符的a_end_idx和 a_start_idx是相等的 @wireless911 @ZhihengHao 但似乎对训练不会造成大的影响,只是构造出来的window=1的span是无效的

Description 1. 修改了模型保存方式,只保存PrefixEncoder部分 2. 优化了训练过程的日志打印方式 3. 新增了预测和cli代码 在chatGLM2+PT场景下,可以复现三元组抽取任务, 不同epoch的模型f1在测试集上表现为: - Epoch 1 0.4001908179539758 - Epoch 5 0.48332597625972235 - Epoch 7 0.5810401474342651 - Epoch 10 0.5901889823940757 基本和作者的实验相符