Comments (4)
fp16 weight requires 16G, and each concurrency requires 1GB to save kv_cache. As the concurrency increases, a total of 30+ GB is normal.
Try int8 quantization to save memory.
from lmdeploy.
https://github.com/InternLM/lmdeploy/blob/main/docs/en/quantization.md
from lmdeploy.
This issue is marked as stale because it has been marked as invalid or awaiting response for 7 days without any further response. It will be closed in 5 days if the stale label is not removed or if there is no further response.
from lmdeploy.
This issue is closed because it has been stale for 5 days. Please open a new issue if you have similar issues or you have any new updates now.
from lmdeploy.
Related Issues (20)
- [Feature] have plan to support MiniCPM-V? HOT 1
- [Bug] AttributeError: 'InternVLChatConfig' object has no attribute 'hidden_size' HOT 20
- [Feature] Support for THUDM/glm-4v-9b HOT 1
- about InternVL−Chat−V1.5 8 bit quantization HOT 2
- High GPU memory for running InternVL-Chat-V1-5-AWQ HOT 4
- [Bug] CUDA OOM during calibration even with 5x 4090s? Falling back to `--device cpu` also fails (with different error) HOT 3
- [Bug] 部署cogvlm2运行时,接受的多个并发之间存在干扰,后面的请求使用前面请求传的图像 HOT 1
- [Feature Request] OpenAI-compatible `stop` param HOT 7
- [Bug] failed to set temperature 1.2 HOT 1
- [Bug] 量化模型时无输出 HOT 4
- [Docs] Where is prefix cache data stored? HOT 6
- [Feature] Speculative Decoding HOT 2
- [Bug] `detokenize_incrementally`: OverflowError: out of range integral type conversion attempted HOT 4
- [Docs] Guidance on setting `num_tokens_per_iter` and `max_prefill_iters` to optimal values HOT 3
- logits输出有问题[Bug] HOT 2
- [Bug] Space is incorrectly removed from start of generated text for `/v1/completion` endpoint HOT 4
- [Bug] Many concurrent requests with `--enable-prefix-caching` AND `--quant-policy 8` crashes with: `CUDA runtime error: an illegal memory access was encountered /opt/lmdeploy/src/turbomind/utils/allocator.h:231` HOT 7
- [Feature] `min_p` sampling parameter HOT 1
- [Feature] Qwen 2 Support HOT 2
- [Bug] xcomposer 4khd lora weight error in lmdeploy HOT 4
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 lmdeploy.