LlamaFamily Llama-Chinese issues

LlamaFamily / Llama-Chinese

Llama中文社区，Llama3在线体验和微调模型已开放，实时汇总最新Llama3学习资料，已将所有代码更新适配Llama3，构建最好的中文Llama大模型，完全开源可商用

https://llama.family

14k stars 1.26k forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

pretrain.sh 中预训练样例数据未提供

#312 liuyongjie985 opened 7 months ago
3
SFT数据格式问题

#311 lovelyturtle opened 7 months ago
1
这个需要什么配置合适？用一张A100 显卡跑的7B模型，80G显存用了10G，回答case中的怎么去北京要60秒才返回结果

#310 MetaRunning opened 7 months ago
2
AMD 的显卡可以用起来吗

#309 jichangxiu opened 8 months ago
0
求大佬帮忙看看，为什么社区的微调代码刚执行到保存了一个checkpoint就报错

#308 bapu3212 opened 8 months ago
1
RuntimeError: FlashAttention only supports Ampere GPUs or newer.

#307 540627735 opened 8 months ago
1
pretrain中的pretrain.sh并不是从头开始训练吧，是增量训练吧

#306 zc-Chao opened 8 months ago
1
提交到slurm集群导致的端口冲突

#305 aozaki-touko opened 8 months ago
0
各位大神，为什么 pip install -r requirements.txt 时，里面依赖的版本有些找不到呢，请指教一下。

#304 githup-cuikang closed 8 months ago
1
中文社区提供的微调代码运行报错，好像是pytest有问题我也不太清楚，有没有大佬帮忙看一下

#303 bapu3212 opened 8 months ago
2
Error while deserializing header: HeaderTooLarge

#302 jack161641 opened 8 months ago
1
请问有人知道这个问题该怎么解决吗TypeError:object of Type Tensor is not JSON serializable.

#301 Ko24Bewjw opened 8 months ago
1
请问llama大模型实践指南纸质版中第一章第18页文献[1]从哪里看？

#300 skaterqiang opened 8 months ago
0
llama-2-13b多卡推理报错 RuntimeError: CUDA error: device-side assert triggered Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

#299 huling-2 opened 8 months ago
5
ollama上run本地部署的atom-7b-chat模型报错"error loading model"

#298 hubblebubblepig opened 8 months ago
2
readme中的数据准备案例最好完全对应上，这会让用户少犹豫两秒

#297 jiangKerman closed 6 months ago
0
Add files via upload

#296 HansJin213 closed 8 months ago
0
关于atom-7b-chat长文本微调应如何进行？

#295 hbj52 opened 8 months ago
5
TypeError: Object of type Tensor is not JSON serializable

#294 Alexhk01 opened 8 months ago
1
如何创建对话的template?

#293 UserName-wang opened 9 months ago
0
Vocab size mismatch causing model convert failure

#292 seuwins opened 9 months ago
0
LLAMA 2 HF tokenizer len is 32001, 迅雷7B model异常需更新。

#291 seuwins closed 6 months ago
1
运行sh脚本后，提示：ds: error: the following arguments are required: user_script, user_args

#290 UCASAlbert opened 9 months ago
0
训练损失从1.4下降到了0.5，训练5个epoch了，从开始到现在验证集ACCURACY一直是64%

#289 kunzeng-ch opened 9 months ago
1
requirements.txt里面标注版本号吧

#288 yazheng0307 closed 6 months ago
3
ValueError: We were not able to get the tokenizer using `AutoTokenizer.from_pretrained` with the string that you have passed /data/mlops/Qwen-7B-Chat. If you have a custom tokenizer, you can pass it as input. For now, we only support quantization for text model. Support for vision, speech and multimodel will come later.

#287 linzm1007 opened 9 months ago
0
Atom-7B-32K模型是不是有问题？

#286 YanZiBuGuiCHunShiWan opened 9 months ago
0
微调过程中学习率的问题

#285 wangdan7477 opened 10 months ago
1
https://llama.family/chat#/ 选择 Atom-7B 是否使用 https://huggingface.co/FlagAlpha/Llama2-Chinese-7b-Chat/模型

#284 ydh10002023 opened 10 months ago
3
AttributeError: 'NoneType' object has no attribute 'to'

#283 LiKe-rm closed 10 months ago
1
微调后"vocab_size": 32001

#282 chris-rong opened 10 months ago
1
预训练和全量参数微调该如何选择？

#281 buaaxiejun opened 10 months ago
0
微调的训练数据怎么准备？

#280 zhangjh closed 9 months ago
2
git clone Llama2-Chinese-13b时报错

#279 494599131 opened 10 months ago
1
70B模型sft和lora需要多大显存和内存

#278 Dagoli opened 10 months ago
1
微调的输出目录，所输出微调记录内容有问题。

#277 312shan opened 10 months ago
1
pretrain的数据为啥用的是sft的数据。

#276 Liufeiran123 opened 10 months ago
1
Docker训练，出现异常

#275 PanJinquan opened 10 months ago
0
如何拿到微调后的embedding向量呢

#274 wanghao19970205 opened 10 months ago
1
如何设置系统提词？

#273 tigermask1978 opened 11 months ago
1
使用vllm优化llama

#272 Lvjinhong opened 11 months ago
1
关于atom-7b-chat微调LORA，Attention mask should be of size (1, 1, 104, 208), but is torch.Size([1, 1, 104, 104])，不知道怎么解决

#271 mesdaq opened 11 months ago
1
微调过程

#270 Z-Zili opened 11 months ago
0
AttributeError: 'LlamaForCausalLM' object has no attribute 'chat'

#269 1801ZDL opened 11 months ago
2
中文乱码

#268 xiaolibuzai-ovo opened 11 months ago
1
完全按照原始代码但是有这个错误

#267 xiaomingtx6 opened 11 months ago
0
完全按照原始代码但是有这个错误

#266 xiaomingtx6 opened 11 months ago
1
api调用后，不输出预期的结果

#265 ghkl98 opened 11 months ago
1
增量预训练和全量参数微调有什么区别呢？

#264 chris-rong opened 11 months ago
0
> 我也和你的问题刚好相反，我一个4090的显卡运行出这样的错误。以下是报错信息： RuntimeError: CUDA error: out of memory CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

#263 Xiaozl11 closed 12 months ago
0

Previous Next