Comments (4)
@hiyouga 后续不知道是否可以参考一下这个repo来改进下长序列微调?https://github.com/jzhang38/EasyContext
from llama-factory.
batchsize 是 1 吗,把 cutoff_len 降低一两倍看看能不能跑起来
from llama-factory.
batchsize 是 1 吗,把 cutoff_len 降低一两倍看看能不能跑起来
是的,per_device_train_batch_size=1,但是我就是需要packing到Qwen支持的最大长度32768呢,是显卡不够吗,贫穷的原因吗,还是说有什么更不占显存的配置呢~
这是我的zero3+offload的配置
{
"bf16": {
"enabled": true
},
"optimizer": {
"type": "AdamW",
"params": {
"lr": "auto",
"weight_decay": "auto",
"torch_adam": true,
"adam_w_mode": true
}
},
"zero_optimization": {
"stage": 3,
"offload_optimizer": {
"device": "cpu",
"pin_memory": true
},
"offload_param": {
"device": "cpu",
"pin_memory": true
},
"overlap_comm": true,
"contiguous_gradients": true,
"reduce_bucket_size": "auto",
"stage3_prefetch_bucket_size": "auto",
"stage3_param_persistence_threshold": "auto",
"sub_group_size": 1e9,
"stage3_max_live_parameters": 1e9,
"stage3_max_reuse_distance": 1e9,
"stage3_gather_16bit_weights_on_model_save": true
},
"gradient_accumulation_steps": "auto",
"gradient_clipping": "auto",
"train_batch_size": "auto",
"train_micro_batch_size_per_gpu": "auto",
"zero_allow_untested_optimizer": true
}
from llama-factory.
目前可能不太支持这么长序列的微调,后续会增加方法。建议先以 8k 长度训练
from llama-factory.
Related Issues (20)
- auto-awq kernels is needed to be installed to use .backward() #19 HOT 1
- Error in launch using a docker image HOT 1
- 支持glm-4-v-9b微调吗 HOT 1
- why not cast dtype in deepspeed zero3 ? HOT 6
- HelloGitHub Badge
- 如何在原有pre-training的100轮基础上再进行100轮呢? HOT 1
- Dropped invalid example Issue HOT 2
- 在webui中,选择evaluate&predict栏,出现的预览命令是llamafactory-cli train HOT 1
- Qwen2-7B lora训练后推理出错 HOT 2
- 关于 model/loader.py 里的 _get_init_kwargs 方法 HOT 2
- 8*A100 训练Qwen2-72B-Chat报显存不足 HOT 5
- 请问有大佬全量sft过qwen2-72B或者qwen1.5-72B的模型吗?有参数推荐吗
- 可以支持一下Skywork-MoE-Base的微调不
- Safetensors HOT 1
- 在llama3-8B-Instruct上进行全参数ICL微调后,模型的ICL效果变的很差,求助 HOT 2
- 全量微调Qwen2 7B的问题 HOT 4
- 命令行如何中断训练及其恢复训练 HOT 5
- Distributed training HOT 1
- accelerate lunch run work fine, but llamafactory-cli train is not working (OOM) HOT 5
- 采用最新代码,运行vllm(0.4.3)报错:undefined symbol: _ZN2at4_ops5zeros4ca... HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from llama-factory.