Move OpenAPI export as an env var within the main function. This
allows for easy export by running main.
In addition, an env variable provides global and explicit state to
disable conditional wheel imports (ex. Exl2 and torch) which caused
errors at first.
Signed-off-by: kingbri <bdashore3@proton.me>
If flash attention is already turned off by exllamaV2 itself, don't
try creating a paged generator. Also condense all the redundant
logic into one if statement.
Also check arch_compat_overrides to see if flash attention should
be disabled for a model arch (ex. Gemma 2)
Signed-off-by: kingbri <bdashore3@proton.me>
Previously, the parameters under the "model" block in config.yml only
handled the loading of a model on startup. This meant that any subsequent
API request required each parameter to be filled out or use a sane default
(usually defaults to the model's config.json).
However, there are cases where admins may want an argument from the
config to apply if the parameter isn't provided in the request body.
To help alleviate this, add a mechanism that works like sampler overrides
where users can specify a flag that acts as a fallback.
Therefore, this change both preserves the source of truth of what
parameters the admin is loading and adds some convenience for users
that want customizable defaults for their requests.
This behavior may change in the future, but I think it solves the
issue for now.
Signed-off-by: kingbri <bdashore3@proton.me>
Move the large import errors into the check functions themselves.
This helps reduce the difficulty in interpreting where errors are
coming from.
Signed-off-by: kingbri <bdashore3@proton.me>
* Model: Clean up paged attention checks
* Model: Move cache_size checks after paged attn checks
Cache size is only relevant in paged mode
* Model: Fix no_flash_attention
* Model: Remove no_flash_attention
Ability to use flash attention is auto-detected, so this flag is unneeded. Uninstall flash attention to disable it on supported hardware.
* Model: More extensive checks for paged attention
Previously, TabbyAPI only checked for whether the user's hardware supports flash attention before deciding whether to enabled paged mode.
This adds checks for whether no_flash_attention is set, whether flash-attn is installed, and whether the installed version supports paged attention.
* Tree: Format
* Tree: Lint
* Model: Check GPU architecture first
Check GPU arch prior to checking whether flash attention 2 is installed
Use a queue-based system to get choices independently and send them
in the overall streaming payload. This method allows for unordered
streaming of generations.
The system is a bit redundant, so maybe make the code more optimized
in the future.
Signed-off-by: kingbri <bdashore3@proton.me>
For multiple generations in the same request, nested arrays kept their
original reference, resulting in duplications. This will occur with
any collection type.
For optimization purposes, a deepcopy isn't run for the first iteration
since original references are created.
This is not the most elegant solution, but it works for the described
cases.
Signed-off-by: kingbri <bdashore3@proton.me>
This adds the ability to add multiple choices to a generation. This
is only available for non-streaming gens for now, it requires some
more work to port over to streaming.
Signed-off-by: kingbri <bdashore3@proton.me>
While TabbyAPI doesn't need a config.yml to run, new users can get
confused by the task of copying config_sample.yml to config.yml.
Therefore, automatically do this in the start script to immediately
expose options to the user.
Signed-off-by: kingbri <bdashore3@proton.me>
Cache size must be a multiple of 256 to work properly in ExllamaV2.
Take the config value and set the cache size to one multiple above
the remainder of the cache size divided by 256.
This is because cache size can never be lower than max_seq_len.
If max_seq_len isn't a multiple of 256, this method will never
yield a number that's lower than max_seq_len since it's no longer
a source of truth.
Signed-off-by: kingbri <bdashore3@proton.me>
Waiting for request disconnect takes some extra time and allows
generation chunks to pile up, resulting in large payloads being sent
at once not making up a smooth stream.
Use the polling method in non-streaming requests by creating a background
task and then check if the task is done, signifying that the request
has been disconnected.
Signed-off-by: kingbri <bdashore3@proton.me>
Depending on the day of the week, Starlette can work with a CancelledError
or using await request.is_disconnected(). Run the same behavior for both
cases and allow cancellation.
Streaming requests now set an event to cancel the batched job and break
out of the generation loop.
Signed-off-by: kingbri <bdashore3@proton.me>
If a user is using GPU split, check compute capability on only those
GPUs. Autosplit assumes that all GPUs will be used.
Signed-off-by: kingbri <bdashore3@proton.me>
List comprehensions are the more "pythonic" way to approach mapping
values to a list. They're also more flexible across different collection
types rather than the inbuilt map method. It's best to keep one convention
rather than splitting down two.
Signed-off-by: kingbri <bdashore3@proton.me>
