Comments (6)
在 output-CoDeX/checkpoint-80000/trainer_state.json
中,最后几行内容是:
{
[
{
"epoch": 2.92,
"learning_rate": 8.248803594695632e-06,
"loss": 0.0629,
"step": 79990
},
{
"epoch": 2.92,
"learning_rate": 8.212272135020273e-06,
"loss": 0.0627,
"step": 80000
}
],
"max_steps": 82221,
"num_train_epochs": 3,
"total_flos": 0.0,
"trial_name": null,
"trial_params": null
}
from kopa.
这个有可能是peft版本的问题,你可以检查一下保存下来的lora权重里面是不是一个空文件。peft库和transformers版本不匹配的时候就会出现这种问题。
from kopa.
一般来说只要用数据训练过,模型肯定是不会出现zero-shot水平的性能的,很有可能就是peft库版本的问题,可以参考peft库的这个issue
huggingface/peft#286
from kopa.
谢谢,我去试一下。
from kopa.
麻烦问下,这个问题有解决吗。我查看了保存的lora权重确实是空的有问题
我的peft库是requirement.txt里的0.3.0
尝试了注释
old_state_dict = model.state_dict
# model.state_dict = (
# lambda self, *_, **__: get_peft_model_state_dict(
# self, old_state_dict()
# )
# ).__get__(model, type(model))
但lora权重保存又太大
是需要升级0.4.0吗
from kopa.
麻烦问下,这个问题有解决吗。我查看了保存的lora权重确实是空的有问题 我的peft库是requirement.txt里的0.3.0 尝试了注释
old_state_dict = model.state_dict
# model.state_dict = ( # lambda self, *_, **__: get_peft_model_state_dict( # self, old_state_dict() # ) # ).__get__(model, type(model))
但lora权重保存又太大 是需要升级0.4.0吗
这个是peft库的问题,可以参考上面的链接里的方案尝试解决
from kopa.
Related Issues (20)
- 关于模型输出问题 HOT 2
- 您好,请问一下inductive设定是如何实现的 HOT 2
- 您好,请问UMLS的数据集被删除了吗,上传的数据中没找到UMLS的数据集 HOT 1
- Model issues HOT 1
- 关于inductive的一点疑问 HOT 1
- 给定query, 实现头/尾实体预测 HOT 1
- 结果为0.5,0.0,0.0,0.0 HOT 3
- requirements.txt上的git地址没了 HOT 1
- 训练相关问题 HOT 1
- 微调后推理的结果比论文中低很多FB15K-237N数据集+llama7B HOT 10
- wandb: 未找到命令 HOT 3
- 关于推理的问题 HOT 1
- 进行分布式训练时报错 HOT 5
- ValueError: Attempting to unscale FP16 gradients. HOT 2
- safetensors_rust.SafetensorError: Error while deserializing header: InvalidHeaderDeserialization HOT 4
- 寻求文章 HOT 1
- ImportError: cannot import name 'get_env_vars' from 'triton._C.libtriton.triton' (unknown location) HOT 7
- 关于论文表1中的P、R、F1指标的计算。 HOT 4
- data processing HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from kopa.