Comments (4)
zero3 有通信成本
from llama-factory.
zero3 有通信成本
zero3 20分钟
zero2 17秒
这个通信成本好高,感觉好奇怪 @hiyouga
from llama-factory.
#!/bin/bash export NCCL_NET=IB export NCCL P2P DISABLE=1 deepspeed --num_gpus 3 ../../src/train_bash.py \ --deepspeed ../deepspeed/ds_z3_config.json \ --stage sft \ --do_train \ --model_name_or_path Meta-Llama-3-8B-Instruct \ --dataset qa \ --dataset_dir ../../data \ --template default \ --finetuning_type lora \ --lora_target q_proj,v_proj \ --output_dir ../../saves/LLaMA3-8B/lora_zero3/sft \ --overwrite_cache \ --overwrite_output_dir \ --cutoff_len 1024 \ --preprocessing_num_workers 16 \ --per_device_train_batch_size 1 \ --per_device_eval_batch_size 1 \ --gradient_accumulation_steps 2 \ --lr_scheduler_type cosine \ --logging_steps 1 \ --warmup_steps 20 \ --save_steps 100 \ --eval_steps 100 \ --evaluation_strategy steps \ --learning_rate 5e-5 \ --num_train_epochs 3.0 \ --max_samples 3000 \ --val_size 0.1 \ --ddp_timeout 180000000 \ --plot_loss \ --fp16
最新代码 sft llama3-8B zero2的速度是zero3的162倍,是不是有点不正常 @hiyouga @fenglui
from llama-factory.
from llama-factory.
Related Issues (20)
- Yi-34B模型使用双卡deepspeed zero2 训练加载模型时占用CPU 内存>200G 不足导致失败 HOT 1
- the number of output lines is more than the number of input lines when batch inference (qwen 1.5, single node, multiple GPUs)
- llama factory推理时,如何配置seed参数 HOT 3
- [rank2]: ImportError: Megatron is not installed. please build it from source. HOT 1
- eval运行mmlu时,results.json中的结果少了一项 HOT 1
- 部署自己模型的api HOT 2
- 【ascend】 RuntimeError: [ERROR] HCCL error in: torch_npu/csrc/distributed/ProcessGroupHCCL.cpp:64 HOT 1
- Ascend NPU训练成功但是推理报错 HOT 1
- lora微调是否支持deepspeed HOT 5
- 4X4090 ,llama_pro examples, Error: Expected attn_mask dtype to be bool or to match query dtype, but got attn_mask.dtype: c10::Half and query.dtype: float instead. HOT 3
- 如何自定义损失函数 HOT 1
- 出现乱码该如何解决
- 历史提交包含git lfs,导致clone后无法上传到自建git平台 HOT 2
- ceval评测结果都为0 HOT 1
- KTO训练报错 HOT 2
- Meta-Llama-3-8B-Instruct进行微调,合并权重后的模型正常。对模型进行int4量化后,量化模型推理出现问题
- RuntimeError: Internal: could not parse ModelProto from /mnt/data/legalexp/LLM_exp/MiniCPM/minicpm_finetune_baseline/MiniCPM-2B-sft-bf16/tokenizer.model HOT 1
- dpo训练后导出模型 HOT 1
- llama3-8B-base模型全量微调mmlu掉点 HOT 5
- 使用KTO进行多机训练过程中再进行验证,报错RuntimeError: still have inglight params [{id:388, "status":"AVALIBLE"}]
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.