shibing624 MedicalGPT issues

shibing624 / MedicalGPT

MedicalGPT: Training Your Own Medical GPT Model with ChatGPT Training Pipeline. 训练医疗大模型，实现了包括增量预训练(PT)、有监督微调(SFT)、RLHF、DPO、ORPO。

Apache License 2.0

3.37k stars 507 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

从头开始预训练 Loss下降缓慢

#426 dage0127 closed 2 weeks ago
3
复现医疗大模型与训练数据加载问题

#425 veresse closed 1 month ago
1
Qwen-7B领域大模型继续预训练

#424 veresse closed 1 month ago
6
shibing624/medical 数据集能不能传一份到魔塔社区那边？

#423 hecheng64 opened 2 months ago
2
麻烦改一下requirenments.txt，有一个包少一个等号

#422 immmor closed 2 months ago
0
耗时过长

#421 ucaslei opened 2 months ago
0
多gpu 的时候运行ppo_training.py报错，

#420 cqray1990 opened 2 months ago
1
请问，我用Qwen1.8b的模型微调，但报错TypeError: argument 'tokens': 'NoneType' object cannot be interpreted as an integer

#419 wangxinwwang opened 2 months ago
0
请教下数据处理部分，tokenizer分词后的position_ids字段是怎么生成的

#418 XiaozhuLove opened 2 months ago
4
使用Llama-2-13b-hf训练奖励模型报错

#417 cqray1990 closed 2 months ago
1
perplexity 微调之后变大了？需要继续微调？

#416 cqray1990 opened 2 months ago
1
sharegpt_gpt4的看了数据跟医疗没有关系，为什么也可以用于医疗多轮对话的微调呢？

#415 cqray1990 opened 2 months ago
1
请问 sft之后的模型效果和base 模型比较，大佬用的什么指标比较呢？可以提供仓库链接？

#414 cqray1990 closed 2 months ago
1
请问支持llama 3.1 微调？

#413 cqray1990 opened 2 months ago
2
词表扩充程序能否应用在Qwen2中

#412 LarryLong45 closed 2 months ago
1
qwen1.5-0.5b-chat按教程SFT后推理无结果

#411 LarryLong45 closed 2 months ago
8
Windows四卡3090平台跑baichuan2-13b时，感觉模型好像没有分布到各个显卡上，显存一下就满了oom了。怎么解决？

#410 Ruiruiz30 opened 3 months ago
3
关于奖励模型训练数据的构成

#409 Eren139 opened 3 months ago
6
run_rm.sh双卡运行全量参数，报RuntimeError: Expected to mark a variable ready only once. This error is caused by one of the following reasons:

#408 XiaozhuLove opened 3 months ago
1
Update README.md

#407 LIE624 closed 3 months ago
0
增量预训练

#406 cqray1990 closed 3 months ago
1
Create validate_jsonl.py

#405 LIE624 closed 3 months ago
0
Test the perplexity

#404 LIE624 closed 3 months ago
0
rm阶段，loss降到0，并且图看起来很奇怪

#403 zhengshi119 closed 1 month ago
9
Update README_EN.md

#402 LIE624 closed 3 months ago
0
Update model_quant.py

#401 LIE624 closed 3 months ago
0
Complete quantification

#400 LIE624 closed 3 months ago
0
Create Multi-GPUs-deployment.sh

#399 LIE624 closed 3 months ago
0
预测时，提示Attention Mask 未设置和Attention Mask 未设置

#398 huangrs494 closed 3 months ago
3
Is possible multilingual English and Spanish

#397 johnfelipe closed 3 months ago
1
大量数据加载问题

#396 dage0127 closed 3 months ago
2
internlm2-1.8b微调报错

#395 deep-practice closed 4 months ago
2
add full_train.py and run_full_train.sh

#394 LIE624 closed 4 months ago
0
训练数据集切分一次，多次重复使用的问题

#393 dage0127 closed 4 months ago
2
请问是否支持最新的InternLM 2.5？

#392 hao203 opened 4 months ago
1
运行sh ./run_ppo.sh时遇到错误ValueError: Target modules q_proj,v_proj not found in the base model. Please check the target modules and try again错误复现过程

#391 iomgaa-ycz closed 3 weeks ago
2
从头开始训练

#390 dage0127 closed 4 months ago
2
增量预训练，这样的input_ids的格式是不是有问题，帮忙看看

#389 minxiansheng closed 4 months ago
1
关于本地训练问题

#388 Ruiruiz30 closed 3 months ago
1
RuntimeError: "nll_loss_out_frame" not implemented for 'Half'

#387 Li-Jicheng opened 5 months ago
2
增量预训练PT与有监督微调SFT的疑问

#386 VirgilG72 opened 5 months ago
1
notebook报错

#385 cheun726 closed 3 months ago
1
支持GLM4微调

#384 turkeymz opened 5 months ago
1
大佬，DPO可以改成inputIds和attention_mask 输入吗

#383 Faded1022 opened 5 months ago
1
大佬，DPO训练报错

#382 cheun726 closed 5 months ago
4
PPO和SFT阶段数据集

#381 pangpang-xuan closed 5 months ago
2
Change Llama tokenizer from LlamaTokenizer to AutoTokenizer

#380 princepride closed 5 months ago
1
ValueError: Please specify target_modules in peft_config

#379 lyj-newbie closed 5 months ago
1
关于llama3的权重转换

#378 tszslovewanpu closed 3 months ago
1
医学大模型全流程体验

#377 YoshuaBengio closed 3 months ago
2