The rope alpha calculation caused an error if max seq len isn't
provided. This is because the model's max sequence length was not
stored as the target for alpha calculation.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
Jobs should be started and immediately cleaned up when calling the
generation stream. Expose a stream_generate function and append
this to the base class since it's more idiomatic than generate_gen.
The exl2 container's generate_gen function is now internal.
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>
If a model is being unloaded, that means its being shut down and
no requests should be accepted from then on.
Also, remove model_is_loaded since we simply check if the container
is None now.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
Maybe move this out of the class entirely, but for now, it makes
sense to encapsulate this logic.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
Apparently the "mirostat" parameter has been updated by frontends
to pass a number. ExllamaV2 expects a boolean, but most pass a number
anyway, so just alias mirostat_mode and mirostat together.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
kwargs is pretty ugly when figuring out which arguments to use. The
base requests falls back to defaults anyways, so pass in the params
object as is.
However, since Python's typing isn't like TypeScript where types
can be transformed, the type hinting has a possiblity of None showing
up despite there always being a value for some params.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
Log the parameters passed into the generate gen function rather than
the generation settings to reduce complexity.
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>
Torch - 2.6.0
ExllamaV2 - 0.2.8
Flash-attn - 2.7.4.post1
Cuda wheels are now 12.4 instead of 12.1, feature names need to be
migrated over.
Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
The previous code overrode the existing gpu split and device idx
values. This now sets an independent draft_gpu_split value and
adjusts the gpu_devices check only if the draft_gpu_split array
is larger than the gpu_split array.
Draft gpu split is not Tensor Parallel, and defaults to gpu_split_auto
if a split is not provided.
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>
* Ensure that length of positive/negative prompt + max_tokens does not exceed max_seq_len
* Ensure that total required pages for CFG request does not exceed allocated cache_size
The vision module from the ExllamaV2 backend is used in files outside
the backends contained folder. Therefore, import ExllamaV2 as an
optional dependency here.
Signed-off-by: kingbri <bdashore3@proton.me>
If vision is enabled and the model doesn't support it, send an
error asking the user to reload. Also, add a method to unload the
vision tower.
Signed-off-by: kingbri <bdashore3@proton.me>
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
* Support image_url inputs containing URLs or base64 strings following OAI vision spec
* Use async lru cache for image embeddings
* Add generic wrapper class for multimodal embeddings
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>
