chore(inference): add docs about the containerized inference stack 📝

Fixes #913

README.md

@@ -51,6 +51,10 @@ docker compose up --build
 Then, navigate to `http://localhost:3000` (It may take some time to boot up) and
 interact with the website.
 
+> **Note:** If an issue occurs with the build, please head to the
+> [FAQ](https://projects.laion.ai/Open-Assistant/docs/faq) and check out the
+> entries about Docker.
+
 > **Note:** When logging in via email, navigate to `http://localhost:1080` to
 > get the magic email login link.
 

docs/docs/faq/faq.md

@@ -9,6 +9,33 @@ For more details and information check out
 [this SO thread](https://stackoverflow.com/questions/66514436/difference-between-docker-compose-and-docker-compose)
 that explains it all in detail.
 
+### Enable Docker's BuildKit Backend
+
+[BuildKit](https://docs.docker.com/build/buildkit/) is Docker's new and improved
+builder backend. In addition to being faster and more efficient, it supports
+many new features, among which is the ability to provide a persistent cache,
+which outlives builds, to compilers and package managers. This is very useful to
+speed up consecutive builds, and is used by some container images of
+OpenAssistant's stack.
+
+The BuildKit backend is used by
+[default by Compose V2](https://www.docker.com/blog/announcing-compose-v2-general-availability/)
+(see above). <br/> But if you want to build an image with `docker build` instead
+of `docker compose build`, you might need to enable BuildKit.
+
+To do so, just add `DOCKER_BUILDKIT=1` to your environment.
+
+For instance:
+
+```shell
+export DOCKER_BUILDKIT=1
+```
+
+You could also, more conveniently,
+[enable BuildKit by default](https://docs.docker.com/build/buildkit/#:~:text=To%20enable%20docker%20BuildKit%20by%20default),
+or use
+[Docker Buildx](https://docs.docker.com/build/#:~:text=The%20new%20client%20Docker%20Buildx).
+
 ### Pre-commit
 
 We are using pre-commit to ensure the quality of the code as well as the same

inference/README.md

@@ -1,14 +1,64 @@
-# OpenAssitant Inference
+# OpenAssistant Inference
 
 Preliminary implementation of the inference engine for OpenAssistant.
 
-## Development Variant 1 (you'll need tmux)
+## Development Variant 1 (docker compose)
+
+The services of the inference stack are prefixed with "inference-" in the
+[unified compose descriptor](../docker-compose.yaml). <br/> Prior to building
+those, please ensure that you have Docker's new
+[BuildKit](https://docs.docker.com/build/buildkit/) backend enabled. See the
+[FAQ](https://projects.laion.ai/Open-Assistant/docs/faq#enable-dockers-buildkit-backend)
+for more info.
+
+To build the services, run:
+
+```shell
+docker compose --profile inference build
+```
+
+Spin up the stack:
+
+```shell
+docker compose --profile inference up -d
+```
+
+Tail the logs:
+
+```shell
+docker compose logs -f    \
+    inference-server      \
+    inference-worker      \
+    inference-text-client \
+    inference-text-generation-server
+```
+
+Attach to the text-client, and start chatting:
+
+```shell
+docker attach open-assistant-inference-text-client-1
+```
+
+> **Note:** In the last step, `open-assistant-inference-text-client-1` refers to
+> the name of the `text-client` container started in step 2.
+
+> **Note:** The compose file contains the bind mounts enabling you to develop on
+> the modules of the inference stack, and the `oasst-shared` package, without
+> rebuilding.
+
+> **Note:** You can spin up any number of workers by adjusting the number of
+> replicas of the `inference-worker` service to your liking.
+
+> **Note:** Please wait for the `inference-text-generation-server` service to
+> output `{"message":"Connected"}` before starting to chat.
+
+## Development Variant 2 (you'll need tmux)
 
 Run `./full-dev-setup.sh` to start the full development setup. Make sure to wait
 until the 2nd terminal is ready and says `{"message":"Connected"}` before
 entering input into the last terminal.
 
-## Development Variant 2 (you'll need multiple terminals)
+## Development Variant 3 (you'll need multiple terminals)
 
 Run a redis container (or use the one of the general docker compose file):