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

2.93k stars 450 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

RuntimeError: "nll_loss_out_frame" not implemented for 'Half'

#387 Li-Jicheng opened 4 days ago
2
增量预训练PT与有监督微调SFT的疑问

#386 VirgilG72 opened 4 days ago
1
notebook报错

#385 cheun726 opened 5 days ago
1
支持GLM4微调

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

#383 Faded1022 opened 6 days ago
1
大佬，DPO训练报错

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

#381 pangpang-xuan closed 1 week ago
2
Change Llama tokenizer from LlamaTokenizer to AutoTokenizer

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

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

#378 tszslovewanpu opened 1 month ago
1
医学大模型全流程体验

#377 YoshuaBengio opened 1 month ago
2
运行pretraining.py时报错：RuntimeError: CUDA error: device-side assert triggered

#376 Wenting1227 opened 1 month ago
4
DPO训练，报错：“IndexError: Invalid key: 0 is out of bounds for size 0”

#375 dage0127 closed 1 month ago
2
ppo训练时出现问题：UserWarning: KL divergence is starting to become negative: -233.50

#374 user2311717757 opened 1 month ago
2
vocab扩展后的模型合并问题

#373 sungatetop opened 1 month ago
1
有没有人能分享下自己微调后的模型id，我懒得弄，只想吃现成的

#372 aqpmzngldh closed 1 month ago
1
AMD 执行 run_pt.sh失败

#371 liuyang6055 opened 1 month ago
1
增加中文数据集汇总，本项目支持格式

#370 ZhuangXialie closed 2 months ago
0
dpo_training.py eal存在空的情况

#369 14686039 closed 2 months ago
2
关于提前结束训练

#368 tszslovewanpu closed 2 months ago
4
add max_length and max_prompt_length

#367 ZhuangXialie closed 2 months ago
0
对chat模型进行二次预训练后，自问自答

#366 wsl1014 opened 2 months ago
1
几步的训练怎么都是独立的，rm都没用sft的adapter

#365 cqray1990 closed 2 months ago
1
训练reward_modeling.py

#364 cqray1990 closed 2 months ago
1
orpo脚本NoneType问题

#363 songyao199681 closed 5 days ago
6
Typo

#362 ker2xu closed 2 months ago
0
reward_modeling咨询

#361 tuqingwen opened 2 months ago
1
Updates for readme and demo ipynb and a small update for deprecated function

#360 ker2xu closed 2 months ago
0
UserWarning: None of the inputs have requires_grad=True. Gradients will be None

#359 cove1011 closed 2 months ago
2
Regarding RLHF and DPO training data

#358 Aniketto16 opened 3 months ago
2
使用deepspeed 全参数sft后，inference 回答的都为空，有解决办法吗

#357 Yian320 opened 3 months ago
2
ValueError: operands could not be broadcast together with remapped shapes [original->remapped]: (3,2) and requested shape (1,2)

#356 Riapy opened 3 months ago
1
lora模型合并

#355 sevenandseven opened 3 months ago
2
提交重构后的代码

#354 youbingchenyoubing closed 3 months ago
1
运行inference.py文件，报AttributeError: property 'eos_token' of 'ChatGLMTokenizer' object has no setter

#353 liulint closed 3 months ago
1
扩充词表后能否直接进行SFT呢？

#352 HaotianLiu123 opened 3 months ago
0
预训练后模型出现自问自答、输出未知序列、重复口吃现象

#351 Peter-of-Astora opened 3 months ago
6
assert tokenzier_vocab_size > model_vocab_size

#350 sevenandseven closed 3 months ago
5
增量预训练效果评估

#349 qibao77 opened 3 months ago
1
关于Chatglm3的增量预训练

#348 XueMoonLit closed 3 months ago
1
llama进行rm训练的时候，出现问题ValueError: weight is on the meta device, we need a `value` to put in on cpu.

#347 cove1011 opened 3 months ago
1
使用qwen进行pretrain的时候出现了问题：Cannot copy out of meta tensor; no data!

#346 cove1011 opened 3 months ago
1
ValueError: The model does not have a language model head, please use a model that has one.

#345 cove1011 closed 3 months ago
1
chatglm3训练在rm之后，进行lora模型权重合并到base model，出现问题：ValueError: chatglm does not support sequence classification

#344 cove1011 closed 3 months ago
2
TypeError: ChatGLMForSequenceClassification.forward() got an unexpected keyword argument 'output_attentions'

#343 cove1011 closed 3 months ago
0
dpo训练出错

#342 cove1011 closed 3 months ago
5
ChatGLMForSequenceClassification rm步骤出错

#341 cove1011 closed 3 months ago
1
dpo_training训练chatglm3-6b模型报错。

#340 xiaochaich closed 3 months ago
1
chatglm2合并sft_qlora后，推理出现自动续答

#339 Lxhnnn closed 3 months ago
3
全量预训练baichuan-7b Out of memory

#338 FFFFFzx closed 3 months ago
3