Comments (3)
It depends a lot on the output of the model unfortunately! The Copilot plugin typically makes requests for 500 tokens, which can take a long time for FauxPilot (~24.8 seconds in a quick test I did just now with a single A6000 and the 16B model). But most of these requests tell the model to stop generating when it emits a newline, so it doesn't end up generating anywhere near 500 tokens and the model finishes much faster. With the 16B model on an A6000 can expect it to take ~50ms per token generated.
Occasionally, like when you are at the start of a function, Copilot will make a request that tries to generate a longer chunk of code (by stopping on sequences like \ndef
or \nif
), and that can be a bit slower (it's slower in the official Copilot as well).
If you have multiple GPUs you can generally expect the latency to decrease almost linearly (so two A6000s can generate almost twice as fast).
If you were trying to use this to serve many clients at once, you might be able to batch together multiple requests that come in at the same time, which would give you better throughput overall; I believe the underlying Triton inference server does support this.
In any case, improving inference latency is definitely something I'm interested in! But I think that will probably require looking at techniques like model compression and quantization.
from fauxpilot.
Also, if you happen to put together something that tries to benchmark this with some real-world code using the kinds of requests Copilot makes I'd definitely be curious to see the results!
from fauxpilot.
@moyix curious to know what happens to the concurrent requests to fauxpilot
. Those concurrent requests would be queued and would be served eventually? Is there a limit of how many concurrent requests could be made? Or the concurrent requests would be lost?
from fauxpilot.
Related Issues (20)
- Infinite time to (load?) and then it doen't even work?? HOT 4
- [bug] docker(version 20.10.21) version parse error in launch.sh HOT 1
- Support arm64 to minimize cost
- Maybe add windows/etc installer all-in-one in this project's 'releases'.
- 400 Bad Request when file has around 100 lines of code HOT 3
- C# support! HOT 2
- Hello all. The comments above have been very helpful in setting up the Copilot extension. I managed to get it to work with my instance and figured I would combine the steps I used (this is for Windows. Linux installation is similar, just different locations):
- It was working fine before... HOT 1
- Support for AMD GPUs HOT 1
- Triton doesnt exist anymore I think? HOT 3
- K8s deployment (via helm chart) HOT 2
- Caught signal 11 (Segmentation fault: address not mapped to object at address (nil)) HOT 1
- why my response are all !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! HOT 3
- Can I merge images of triton and client into one?eg fastertransformer_backend get content_fetch <fastertransformer&client>in CMakeLists ? HOT 1
- help me HOT 1
- What is the comparison of these model in huggingface? HOT 2
- Python Backend: "Expected all tensors to be on the same device, but found at least two devices, cpu and cuda:0" HOT 2
- [promptlib] proxy {"cause":{}} HOT 1
- ollama HOT 2
- Company Proxy 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 fauxpilot.