Giters
yanqiangmiffy
/
InstructGLM
ChatGLM-6B 指令学习|指令数据|Instruct
Geek Repo:
Geek Repo
Github PK Tool:
Github PK Tool
Stargazers:
655
Watchers:
11
Issues:
33
Forks:
51
yanqiangmiffy/InstructGLM Issues
Problems in train_deepspeed.py with ZeRO stage 1|2|3
Updated
a year ago
使用lora训练,使用web_demo加载lora权重后,结果跟原生chatglm结果一样,lora权重没生效,这个是什么原因呢
Updated
a year ago
Comments count
1
请问下train_deepspeed.py 怎么引入lora.pt
Updated
a year ago
Comments count
1
ValueError: ChatGLMForConditionalGeneration does not support gradient checkpointing.
Updated
a year ago
Comments count
7
调教后的逻辑能力如何?
Updated
a year ago
Comments count
1
RuntimeError: torch.cat(): expected a non-empty list of Tensors
Closed
a year ago
Comments count
7
torch.distributed.elastic.multiprocessing.errors.ChildFailedError
Closed
a year ago
Comments count
1
用tokenizer_dataset_rows.py转换自己的数据报错datasets.builder.datasetgeneraationerror
Updated
a year ago
用BelleGroup/train_1M_CN训练后,为什么用数据集里的问题测,回答不一样
Updated
a year ago
train_lora最低需要多大显存GPU可以训练?除了batch size 还有别的参数可以降低显存使用吗?
Updated
a year ago
Comments count
2
训练python train_lora.py的时候显示 ModuleNotFoundError: No module named 'configuration_chatglm'
Updated
a year ago
Comments count
3
用peft加载lora后,generate时报错ValueError: 130000 is not in list,加载lora之前推理是正常的
Updated
a year ago
Lora+DeepSpeed多机多卡的问题
Closed
a year ago
测试数据打不开https://huggingface.co/datasets/BelleGroup/generated_train_0.5M_CN
Updated
a year ago
Comments count
2
预测时,torch.set_default_tensor_type(torch.cuda.HalfTensor)的问题
Closed
a year ago
datasets.builder.InvalidConfigName: Bad characters from black list '<>:/\|?*' found in 'data/belle_data.json'. They could create issues when creating a directory for this config on Windows filesystem.
Updated
a year ago
Comments count
1
运行web_demo_alpaca_lora.py报错,是单纯的显存不够嘛
Updated
a year ago
Comments count
1
4张32G的可以吗,作者可以用你写的其他开源数据集finetune看看效果吗,再放出转换和训练代码
Updated
a year ago
Comments count
2
4张 12G的 3060能训练吗
Updated
a year ago
Comments count
2
微调2:BELLE中文指令数据的问题
Updated
a year ago
Comments count
1
ValueError: 150000 is not in list
Updated
a year ago
Comments count
5
web_demo_belle生成结果时有大段重复的问题
Updated
a year ago
Comments count
10
怎么能把lora参数merge回原始模型呢?
Closed
a year ago
Comments count
1
ValueError: Please specify `target_modules` in `peft_config`
Closed
a year ago
Comments count
3
RuntimeError: expected scalar type Half but found Float
Closed
a year ago
Comments count
1
24G显存的3090可以训练吗?
Closed
a year ago
Comments count
1
请问有训练好的权重可以下载吗?
Closed
a year ago
Comments count
1
基于原始chatglm-6b训练效果好还是基于alpaca的lora继续微调效果好呢?
Closed
a year ago
Comments count
4
请问支持多卡吗,怎么改造?
Closed
a year ago
Comments count
2
最新update的代码中,web_demo推理时报错
Updated
a year ago
Comments count
2
运行 finetune.py 遇到问题:OSError: /data/pretrained-chatglm-6b/ does not appear to have a file named config.json
Updated
a year ago
Comments count
1
关于训练完成后,生成的答案总是带一些莫名奇妙的Q,A数据,真的不造是哪里出了问题,还望大佬赐教!谢谢!
Updated
2 years ago
Comments count
1
关于多轮对话的疑问
Updated
2 years ago