issues
search
Facico
/
Chinese-Vicuna
Chinese-Vicuna: A Chinese Instruction-following LLaMA-based Model —— 一个中文低资源的llama+lora方案，结构参考alpaca
https://github.com/Facico/Chinese-Vicuna
Apache License 2.0
4.14k
stars
421
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
python chat.py时报错
#56
googlebox007
closed
1 year ago
2
无法成功执行generate.sh以及interaction.sh
#55
alisyzhu
closed
1 year ago
2
测试时输出回答无法停止，直到256长度限制，loss很快收敛性，到0.82左右就不再下降
#54
rookiebird
closed
1 year ago
3
单机多卡一直报超时错误，请教下大佬有没有啥解决的办法啊
#53
xqmmy
closed
1 year ago
11
断点重训：如何设置resume_from_checkpoint
#52
simonqian
closed
1 year ago
4
文档提供的colab链接不能运行
#51
Linnest2020
closed
1 year ago
1
finetune时报错：KeyError: 'models.llama'
#50
simonqian
closed
1 year ago
17
请教下单机多卡训练的卡死问题
#49
zhoujx4
closed
1 year ago
8
请问下这个repo和 https://github.com/lm-sys/FastChat 有没有关系？
#48
fishfree
closed
1 year ago
5
13B finetune error: AssertionError: No inf checks were recorded for this optimizer.
#47
lyogavin
closed
1 year ago
3
generate_quant.py 脚本运行失败
#46
greatewei
closed
1 year ago
5
关于finetune_contine.sh报错与使用finetune后的模型输出错误
#45
BUPTccy
closed
1 year ago
4
V100显卡，完全follow说明跑的finetue，会出现乱码。 似乎是8bit的原因。
#44
jzsbioinfo
closed
1 year ago
3
环境问题,不太理解..
#43
bambooqj
closed
1 year ago
3
readme 的环境配置指导文件是否存在错误 ？
#42
robin-human
closed
1 year ago
9
3070ti训练时报错：cublasLt ran into an error!
#41
vegech1cken
closed
1 year ago
5
是否可以使输出 自定义格式？
#40
ZenXir
closed
1 year ago
3
为什么fine-tune过程中loss会忽大忽小呢？
#39
alisyzhu
closed
1 year ago
22
关于继续训练（continue finetuning）
#38
creamiracle
closed
1 year ago
12
fix typo in tools/readme.md
#37
eltociear
closed
1 year ago
0
关于语料
#36
ZenXir
closed
1 year ago
1
Vicuna 里的 finetune.py 和 alpaca-lora 里的finetune.py finetune后的模型可以互通吗？
#35
ZenXir
closed
1 year ago
2
执行finetune.sh时出现问题
#34
tianbuwei
closed
1 year ago
1
Vicuna.cpp 衍生自 Llama.cpp 还是 Alpaca.cpp？ 基于 Alpaca 和 Llama 中文embedding后 finetune出来的模型会通用吗？
#33
ZenXir
closed
1 year ago
6
基于13B的LLAMA模型fine-tune，loss特别大，而lr初始就是0，这是正常的吗？
#32
alisyzhu
closed
1 year ago
18
运行generate.py报错：huggingface_hub.utils._validators.HFValidationError
#31
Harpsichord1207
closed
1 year ago
3
关于语料
#30
ZenXir
closed
1 year ago
7
torch.cuda.OutOfMemoryError
#29
iamblue
closed
1 year ago
3
https://chat.lmsys.org/ 問到第三輪會跑超久
#28
iamblue
closed
1 year ago
1
Path Typo in `tools/merge_lora.py`
#27
amy17519
closed
1 year ago
0
关于语料问题
#26
ZenXir
closed
1 year ago
3
你好，请问上下文对话中你们是引入了langchain吗？
#25
haxx12113
closed
1 year ago
2
finetune的 MAX_STEPS = None 意义是什么？可以改成其他吗？
#24
ZenXir
closed
1 year ago
13
量化后的模型4bit，模型输出是乱码
#23
GUORUIWANG
closed
1 year ago
3
modify examples typo in README
#22
HUGHNew
closed
1 year ago
1
对中文理解上，加一层embedding 会不会更好？
#21
ZenXir
closed
1 year ago
4
llama_quant.py 用datasets进行评估的意义
#20
GUORUIWANG
closed
1 year ago
4
llama能加载多个lora模型的参数吗
#19
greatewei
opened
1 year ago
3
TypeError: dispatch_model() got an unexpected keyword argument 'offload_index'
#18
alisyzhu
closed
1 year ago
2
使用llama-13b-hf预训练模型，训练过程学习率变成0，想问下原因?
#17
chenzk1993
closed
1 year ago
9
fix cpp ctx-len
#16
LZY-the-boys
closed
1 year ago
0
undefined reference to `ggml_new_tensor_1d' `ggml_new_tensor_2d'
#15
yuxuan2015
closed
1 year ago
5
后续会考虑加入RLHF吗
#14
yuxuan2015
opened
1 year ago
2
运行 generate.py 提示找不到 config.json
#13
ToSev7en
closed
1 year ago
5
关于LLAMA词表的疑问
#12
uloveqian2021
opened
1 year ago
19
情景设定+多轮对话
#11
ZenXir
closed
1 year ago
5
Generation visualization reformat and clean
#10
LZY-the-boys
closed
1 year ago
0
add support for beam search streamly output
#9
LZY-the-boys
closed
1 year ago
0
merge 13B模型和finetune出来的13B结果进程直接被Killed 当时机器内存还有23G，显存未使用
#8
ZenXir
closed
1 year ago
3
多轮对话，语境，和回复循环的问题
#7
ZenXir
closed
1 year ago
11
Previous
Next