Comments (3)
Hey @hustxiayang, IIRC, applying the tree-attention results in significantly more compute than using argmax, so the throughput cost is too high to justify the value.
Don't quote me on this but IIRC without the tree the speedup is 2x, while with the tree there is a maximum of ~3x speedup but throughput significantly decreases (divided by 10-20x).
To confirm with @Narsil eventually, hope it sheds some light on the question in the meantime.
from text-generation-inference.
Hi, I suppose you observe speedups from one of:
text-generation-inference/gemma-7b-it-medusa
text-generation-inference/Mixtral-8x7B-Instruct-v0.1-medusa
text-generation-inference/Mistral-7B-Instruct-v0.2-medusa
If that is the case, which dataset(s) did you use?
Another question: do you plan to support typical sampling mentioned in their paper?
from text-generation-inference.
This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 5 days.
from text-generation-inference.
Related Issues (20)
- How to clean the TGI guidance cache? HOT 5
- TGI crashes on Multi GPUs HOT 1
- Tgi crash on multi GPUs HOT 10
- Description of Tools format within Documentation HOT 2
- OPENAI Format Completion endpoint does not include finish reason and logprobs HOT 1
- `bitsandbytes==0.43.0` not supported on MacOS HOT 1
- Support for no_repeat_ngram_size generation parameter HOT 3
- OpenAI supports `top_p = 0.0` and `top_p = 1.0` but TGI fails with a validation error with either of these values. HOT 4
- Gibberish generated with deepseek-ai/deepseek-coder-6.7b-base HOT 4
- CPU and Memory Utilization for TGI HOT 1
- Model offloading (preparing for Llama3 405b) HOT 3
- Nightly load test results HOT 2
- Regression: get_weights_col_packed_qkv() quantize parameter not included in calls so fails with error missing positional parfameter - merge error?
- Shard process was signaled to shutdown with signal 4 rank=0 Error: ShardCannotStart HOT 13
- Can I somehow change attention type from 'FlashAttention' in the text-server-launcher? HOT 1
- `tool_calls` gives sporadic EoF parsing errors HOT 4
- max_batch_size limit doesn't work well at queue.next_batch() HOT 2
- can't start server with small --max-total-tokens. But works fine with big stting HOT 4
- Add support for Mistral-Nemo HOT 4
- TGI fails with local LORA adapters 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 text-generation-inference.