This allows for users to use nccl or native depending on the GPU setup.
NCCL is only available with Linux built wheels.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
use_as_default was not being properly applied into model overrides.
For compartmentalization's sake, apply all overrides in a single function
to avoid clutter.
In addition, fix where the traditional /v1/model/load endpoint checks
for draft options. These can be applied via an inline config, so let
any failures fallthrough.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
These added extra complexity and should be removed and replaced
with a single parameter.
Changes:
- /v1/model/load must use model_name and draft_model_name
- /v1/model/embedding/load must use embedding_model_name
- /v1/template/switch must use prompt_template_name
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
The model card is a unified structure for sharing model params.
Rather than kwargs, use this instead.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
This shouldn't even be an exposed option since changing it always
breaks inference with the model. Let the model's config.json handle
it.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
For the TP loader, GPU split cannot be an empty array. However,
defaulting the parameter to an empty array makes it easier to calculate
the device list. Therefore, cast an empty array to None using
falsy comparisons at load time.
Also add draft_gpu_split to the load request.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
The props endpoint is a standard used by llamacpp APIs which returns
various properties of a model to a server. It's still recommended to
use /v1/model to get all the parameters a TabbyAPI model has.
Also include the contents of a prompt template when fetching the current
model.
Signed-off-by: kingbri <8082010+bdashore3@users.noreply.github.com>
Previously, the flow for parsing chat completion messages and rendering
from the prompt template was disconnected between endpoints. Now, create
a common function to render and handle everything appropriately afterwards.
Signed-off-by: kingbri <bdashore3@proton.me>
* More robust checks for OAI chat completion message lists on /v1/encode endpoint
* Added TODO to support other aspects of chat completions
* Fix oversight where embeddings was not defined in advance on /v1/chat/completions endpoint
If an API key sends a dummy model, it shouldn't error as the server
is catering to clients that expect specific OAI model names. This
is a problem with inline model loading since these names would error
by default. Therefore, add an exception if the provided name is in the
dummy model names (which also doubles as inline strict exceptions).
However, the dummy model names weren't configurable, so add a new
option to specify exception names, otherwise the default is gpt-3.5-turbo.
Signed-off-by: kingbri <bdashore3@proton.me>
Adds the ability to load vision parts of text + image models. Requires
an explicit flag in config because there isn't a way to automatically
determine whether the vision tower should be used.
Signed-off-by: kingbri <bdashore3@proton.me>
* Model: Fix inline loading and draft key
There was a lack of foresight between the new config.yml and how
it was structured. The "draft" key became "draft_model" without updating
both the API request and inline loading keys.
For the API requests, still support "draft" as legacy, but the "draft_model"
key is preferred.
Signed-off-by: kingbri <bdashore3@proton.me>
* OAI: Add draft model dir to inline load
Was not pushed before and caused errors of the kwargs being None.
Signed-off-by: kingbri <bdashore3@proton.me>
* Model: Fix draft args application
Draft model args weren't applying since there was a reset due to how
the old override behavior worked.
Signed-off-by: kingbri <bdashore3@proton.me>
* OAI: Change embedding model load params
Use embedding_model_name to be inline with the config.
Signed-off-by: kingbri <bdashore3@proton.me>
* API: Fix parameter for draft model load
Alias name to draft_model_name.
Signed-off-by: kingbri <bdashore3@proton.me>
* API: Fix parameter for template switch
Add prompt_template_name to be more descriptive.
Signed-off-by: kingbri <bdashore3@proton.me>
* API: Fix parameter for model load
Alias name to model_name for config parity.
Signed-off-by: kingbri <bdashore3@proton.me>
* API: Add alias documentation
Signed-off-by: kingbri <bdashore3@proton.me>
---------
Signed-off-by: kingbri <bdashore3@proton.me>
There are two ways to load a model:
1. Via the load endpoint
2. Inline with a completion
The defaults were not applying on the inline load, so rewrite to fix
that. However, while doing this, set up a defaults dictionary rather
than comparing it at runtime and remove the pydantic default lambda
on all the model load fields.
This makes the code cleaner and establishes a clear config tree for
loading models.
Signed-off-by: kingbri <bdashore3@proton.me>
Using aiofiles, there's no longer a possiblity of blocking file operations
that can hang up the event loop. In addition, partially migrate
classes to use asynchronous init instead of the normal python magic method.
The only exception is config, since that's handled in the synchonous
init before the event loop starts.
Signed-off-by: kingbri <bdashore3@proton.me>
- add models for config options
- add function to regenerate config.yml
- replace references to config with pydantic compatible references
- remove unnecessary unwrap() statements
TODO:
- auto generate env vars
- auto generate argparse
- test loading a model
The config categories can have defined separation, but preserve
the dynamic nature of adding new config options by making all the
internal class vars as dictionaries.
This was necessary since storing global callbacks stored a state
of the previous global_config var that wasn't populated.
Signed-off-by: kingbri <bdashore3@proton.me>
Using "auto" for rope alpha removes ambiguity on how to explicitly
enable automatic rope calculation. The same behavior of None -> auto
calculate still exists, but can be overwritten if a model's tabby_config.yml
includes `rope_alpha`.
Signed-off-by: kingbri <bdashore3@proton.me>
It's best to pass them down the config stack.
API/User config.yml -> model config.yml -> model config.json -> fallback.
Doing this allows for seamless flow and yielding control to each
member in the stack.
Signed-off-by: kingbri <bdashore3@proton.me>
Storing a pathlib type makes it easier to manipulate the model
directory path in the long run without constantly fetching it
from the config.
Signed-off-by: kingbri <bdashore3@proton.me>
* Add healthcheck
- localhost only /healthcheck endpoint
- cURL healthcheck in docker compose file
* Update Healthcheck Response
- change endpoint to /health
- remove localhost restriction
- add docstring
* move healthcheck definition to top of the file
- make the healthcheck show up first in the openAPI spec
* Tree: Format
Use the tensor parallel loader when the flag is enabled. The new loader
has its own autosplit implementation, so gpu_split_auto isn't valid
here.
Also make it easier to determine which cache type to use rather than
multiple if/else statements.
Signed-off-by: kingbri <bdashore3@proton.me>
Embedding models are managed on a separate backend, but are run
in parallel with the model itself. Therefore, manage this in a separate
container with separate routes.
Signed-off-by: kingbri <bdashore3@proton.me>
Add an API parameter to set the timeout in seconds. Keep it to None
by default for uninterrupted downloads.
Signed-off-by: kingbri <bdashore3@proton.me>
Place OAI specific routes in the appropriate folder. This is in
preperation for adding new API servers that can be optionally enabled.
Signed-off-by: kingbri <bdashore3@proton.me>
