Comments (10)
Hi, have you solved it, I come across the same problem
from driving-with-llms.
Hi @Phoebe-ovo @xzebin775 thanks for reporting this issue! We are not able to reproduce this error on the GPUs we have. Could you please let me know what GPUs were you using?
Also, can you try setting load_in_8bit
to False to see if this issue can be solved?
from driving-with-llms.
The GPU I used is V100, what GPUs were you using?
from driving-with-llms.
It is GTX 1080 Ti.
from driving-with-llms.
Thanks for confirming. Can you try if setting load_in_8bit to False in here solves the problem?
from driving-with-llms.
I set load_in_8bit to False, but I get the error below. It seems I cann't load the model to GPU
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 86.00 MiB (GPU 0; 10.92 GiB total capacity; 10.44 GiB already allocated; 22.62 MiB free; 10.45 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
from driving-with-llms.
Upon thorough investigation, we are not able to reproduce the error on the GPUs we have (NVIDIA A100 and 3090), but it might related to other issues. I suggest you try these:
- Try clean up the virtualenv (rm env -rf), pull the latest main and setup the environment following the Setup section in the README
- Try using Python 3.9+
Also, we noticed the base model we used decapoda-research/llama-7b-hf
was removed by the author in the huggingface model repo and we are testing the workarounds.
from driving-with-llms.
I meet the same problem, and I set "do_sample" = False, then it worked. Don't know what impact this will have. (same with GPUV100)
from driving-with-llms.
Related Issues (11)
- Visualization of the results HOT 1
- Unspecified signals HOT 1
- About the prompt HOT 2
- The environment
- The code and training command of the stage-1: vector representation pre-training stage HOT 1
- 401 Client Error: Unauthorized for url HOT 2
- You can optimkze frequent routes like soc designs
- Dataset and LLM
- How to Use Multi-GPU Training HOT 2
- Pre trained model HOT 2
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 driving-with-llms.