issues
search
hiyouga
/
LLaMA-Factory
Unify Efficient Fine-Tuning of 100+ LLMs
Apache License 2.0
25.36k
stars
3.14k
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Qwen2 lora微调后用llamafactory-cli export命令合并模型 推理结果有"assstant: "前缀
#4639
jfzleo
closed
2 minutes ago
1
预测推理特别慢,跑完GPU利用率为0了一直卡在那里好像是构建generation
#4638
Harryjun
closed
1 minute ago
1
能不能把eval loss曲线加到训练过程中?
#4637
goodmaney
closed
54 minutes ago
0
Web UI can be used to configure the port through os env variable.
#4636
IdleIdiot
closed
2 hours ago
1
Qwen2 debug 发现 labels全为-100
#4635
xjtulien
closed
2 hours ago
1
PiSSA训练和推理的疑问?
#4634
ConniePK
closed
2 hours ago
1
如何在yaml中配置环境变量中tensorboard的路径呢
#4633
xaiocaibi
closed
4 hours ago
1
910b qwen2 lora生成的模型如何合并权重
#4632
wphtrying
closed
4 hours ago
1
Qwen dpo训练卡住
#4631
yxk9810
closed
4 hours ago
1
请问可以微调没有lm head模型吗
#4630
jzzzf
closed
4 hours ago
1
deepspeed ds_z3_offload_config单卡全量微调训练glm4出现exits with return code = -9。出现该问题时,CPU内存(252G)占满,想问一下这个问题该如何解决?
#4629
ldknight
closed
5 hours ago
1
Yi-1.5-9B推理gpu利用率为0
#4628
Jack-mi
opened
6 hours ago
1
大模型微调分类任务,但是预测结果是不固定的
#4627
nvliajia
closed
10 hours ago
2
4张M40 配置,使用accelerate启动训练,出现TypeError: unsupported operand type(s) for *: 'NoneType' and 'int'
#4626
Micla-SHL
closed
17 hours ago
2
windows上start直接Fail并出现llamafactory-cli乱码
#4625
rizi960
closed
19 hours ago
2
训练34B-reward,Assertion `srcIndex < srcSelectDimSize` failed
#4624
xaiocaibi
closed
1 day ago
2
ModuleNotFoundError: No module named 'vllm.lora'
#4623
xaiocaibi
closed
1 day ago
2
RuntimeError: "addmm_impl_cpu_" not implemented for 'Half' 华为910 命令行推理报错
#4622
apachemycat
opened
1 day ago
1
使用A10对qwen-14b-chat进行Lora微调,2机2卡训练比1机2卡慢了10倍
#4620
WangxuP
closed
2 days ago
3
什么时候支持基于Ray的分布式Lora微调呢?
#4619
WangxuP
closed
2 days ago
0
关于对话模板作用以及其在lm-evaluation-harness仓库下对评测效果影响的问题
#4618
marvelcell
opened
2 days ago
0
DPO 训练时,prompt 与 answer 拼接问题,导致cutoff_length这一超参数无法对数据进行有效截断。
#4617
THZdyjy
closed
19 hours ago
2
qwen2-72b DPO 训练爆显存,OOM 问题;
#4616
THZdyjy
closed
2 days ago
1
🚨FAQs | 常见问题🚨
#4614
hiyouga
opened
2 days ago
0
ValueError: Output directory already exists and is not empty. Please set overwrite_output_dir.
#4612
teddy911405
closed
2 days ago
1
How to use a fine-tuned model to evaluate on a testset and save the output of the model?
#4611
bingkunyao
closed
2 days ago
1
华为NPU训练不了,用的例子里的训练脚本,镜像也是官方镜像
#4610
apachemycat
closed
1 day ago
4
ppo合并失败
#4609
luowei0701
opened
3 days ago
1
fsdp + DPO + fullyfintune会报错
#4608
qy1026
opened
3 days ago
1
大佬,fp8会考虑支持吗?
#4607
chengcheng8632
closed
3 days ago
0
[PPU]大佬有对ppu环境进行过测试么
#4606
willionZS
opened
3 days ago
0
Gemma2
#4605
OKC13
closed
3 days ago
1
能加入matmulfreellm吗?
#4604
quida01
closed
3 days ago
0
你好,请问 KTO是否支持history?
#4603
ldknight
closed
3 days ago
1
是否支持01-ai/Yi-VL-6B
#4602
LegendSun0
closed
2 days ago
2
怎样使用accelerate库进行微调呢?
#4601
shenxiaochenn
closed
3 days ago
1
kto训练完如何预测
#4600
tcxia
closed
3 days ago
3
更新了最新代码,为什么webui里的模型选择看不到Qwen2?
#4599
luchenwei9266
closed
3 days ago
1
glm系列模型做eval时应该将template参数设为什么
#4598
DaozeZhang
closed
3 days ago
5
8卡A800全参数预训练GLM4-9B-base,使用bf16,loss在暴涨后突然消失
#4597
lclcjj
opened
3 days ago
0
请问支持模型并行吗?如果我想要在48GB*8显卡上全量微调llama3-8b,怎么设置呢?
#4596
qy1026
closed
3 days ago
1
全参pt微调Qwen2-7B-Instruct模型,中断后继续训练,修改了lr但没生效
#4595
bigcash
closed
2 days ago
1
npu支持GPTQ量化导出吗
#4594
murray-z
closed
3 days ago
1
[bug] unsloth坏了
#4593
letterk
closed
3 days ago
1
执行命令 报错 flash_attn未安装, 安装后报错ImportError. 使用docker compose ,docker同样的问题
#4592
goodmaney
closed
3 days ago
6
模型没有加载完,gpu利用率已经是100%了
#4591
ceyun1
closed
4 days ago
1
Exit the process with the subprocess's return code when utilizing the CLI
#4590
injet-zhou
closed
4 days ago
0
Exit the process with the subprocess's return code when utilizing the CLI
#4589
injet-zhou
closed
4 days ago
0
有离线安装方案嘛
#4588
heimaojinzhangyz
closed
4 days ago
0
PISSA模式下进行qLORA训练,指定了lora_rank=8,但是训练出来的adaptor_config中是lora_rank=16
#4586
xiningnlp
closed
4 days ago
1
Next