hf_text-generation-inference/router
OlivierDehaene a6a0c97ed9
feat: prefill chunking (#2600)
* wip

* rollback

* refactor to use prefix/postfix namming + fix all_input_ids_tensor

* maybe patching vlms?

* fix filter and concat

* wip, no filter, no concat

* current

* add prepare_for_prefill

* working

* load tested

* re-create slots

* re-create slots

* fix slot_filtering_indices

* feedback loop

* remove log

* fix benchmarker

* fix vlm and seq2seq

* rename to cache and input lengths

* fix prefill logprobs

* fix launcher

* fix logprobs?

* idk at this point

* max input length

* omfg

* remove debugging lines

* fix tests

* fix mllama

* fix cargo tests

* remove support chunking for paged

* Fixing non blocked attentions

* Fixing dtype + AMD, Ipex targets.

* lint fix.

* rename

* Fix prefix_caching variable, remove defaults in server (confusing a lot
of the times).

* Add simple resolution when user specifies ATTENTION=paged.

* Put back non default simple tests.

* Fix env name

---------

Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
2024-10-16 12:49:33 +02:00
..
src feat: prefill chunking (#2600) 2024-10-16 12:49:33 +02:00
Cargo.toml Improve support for GPUs with capability < 8 (#2575) 2024-09-27 16:19:42 +02:00
README.md chore: add pre-commit (#1569) 2024-02-16 11:58:58 +01:00
build.rs chore(github): add templates (#264) 2023-05-02 15:43:19 +02:00

README.md

Router

Also named webserver throughout the docs.

This router is handling most of the logic to handle the "batches" tell when to pass new prefill requests and pausing decode requests, which ones etc...

It uses gRPC to communicate with the shards which can therefore be kept much simpler and focus on having the most efficient forward passes as possible.

Continuous batching

One important feature of text-generation-inference is enabled by this router.

Continuous batching is the act of regularly running queries in the same forward step of the LLM (a "batch") and also removing them when they are finished.

In order for continuous batching to be useful, you need to have more compute available with respect to the memory requirements of your model. This is essentially true for LLMs and the larger the model, the truer it gets (since you have to pool multiple GPUs to load the model, you effectively have a lot of compute power at your hands).

Static batching is the act of doing several queries at the same time, but usually this is controlled by the client, and therefore the amount of batching is decided beforehand.

For text-generation, and LLMs which are memory bound we can try to be much more efficient with the available compute, by having client sending us single queries, and let the router mix&match queries into or out of batches to make the use the compute the most efficiently. This is possible because for LLMs the total compute for running the model is much bigger than doing mix&match of the batches themselves.

Simple continuous batching

text-generation works by feeding a prompt to a model, and iteratively calling forward on the model to produce new text, 1 token at a time.

The first idea is simple, when a query arrives, we start working on it directly. When new queries arrive, we simply wait for the current forward to be finished then batch the current running prompt with the new query, and call forward.

Whenever either query is finished: either the model produce EOS (end of sentence) token or the query reached the allowed limit. We simply drop it from the batch, remove all the allocated memory and we can continue with the rest until nothing is left.

This simple idea generalizes very well and we could potentially stack many requests in the same batch.

One thing to note, is that queries can be potentially run with different parameters meaning different way to choose the next token (sampling, not sampling, temperature, top_k etc..). This is not problematic for the proposed approach we just need to do the sampling independantly on each member of the batch.

Prefill, decode and past key values

In order to make LLMs and text-generation efficient, there's actually a very powerful trick that can be used, which is the "caching" of some attention matrices. More on that in the first part of this blog

What this means, is that the first "pass" of a prompt is different from the subsequent "forward" passes. Since for the first one we have to compute the entire attention matrix, whereas in the follow-ups only require to compute the new token attention. The first pass is called prefill throughout this codebase where as the follow-ups are called decode.

Since prefill is much more expensive than decode we don't want to do it all the time, but a currently running query is probably doing decode. If we want to do the continuous batching as explained previously we need to run prefill at some point in order to create the attention matrix required to be able to join the decode group.

text-generation-inference uses a bunch of different strategies and parameters in order to enable you to find the sweet spot between exploiting the hardware and perceived latency.

With no continuous batching at all, latency is going to be super good, but throughput (meaning the total number of requests allowed in a given timeframe) is going to be super bad (since it's essentially 1).

With static batching, you can probably reach the maximum throughput (by using the maximum total batch size applicable to your hardware), but the latency is super bad since in order to have maximum throughput you need to wait for requests to come in before processing.

With continuous batching you can find a sweet spot. In general latency is the most critical parameter users care about. But a 2x latency slowdown for 10x more users on the same hardware is an acceptable tradeoff.

Token streaming

This is a very important aspect of client UX. As mentionned above, latency is the most critical perceived quality of an LLM API.

With token streaming, the server can start answering after the first prefill pass directly, without waiting for all the generation to be done. For extremely long queries this means clients can start to see something happening orders of magnitude before the work is done. Seeing something in progress allows them to cut short if it's not what's wanted but also it "feels" better.