- [ ] **Description:**
- pass the device_map into model_kwargs
- removing the unused device_map variable in the hf_pipeline function
call
- [ ] **Issue:** issue #13128
When using the from_model_id function to load a Hugging Face model for
text generation across multiple GPUs, the model defaults to loading on
the CPU despite multiple GPUs being available using the expected format
``` python
llm = HuggingFacePipeline.from_model_id(
model_id="model-id",
task="text-generation",
device_map="auto",
)
```
Currently, to enable multiple GPU , we have to pass in variable in this
format instead
``` python
llm = HuggingFacePipeline.from_model_id(
model_id="model-id",
task="text-generation",
device=None,
model_kwargs={
"device_map": "auto",
}
)
```
This issue arises due to improper handling of the device and device_map
parameters.
- [ ] **Explanation:**
1. In from_model_id, the model is created using model_kwargs and passed
as the model variable of the pipeline function. So at this moment, to
load the model with multiple GPUs, "device_map" needs to be set to
"auto" within model_kwargs. Otherwise, the model defaults to loading on
the CPU.
2. The device_map variable in from_model_id is not utilized correctly.
In the pipeline function's source code of tnansformer:
- The device_map variable is stored in the model_kwargs dictionary
(lines 867-878 of transformers/src/transformers/pipelines/\__init__.py).
```python
if device_map is not None:
......
model_kwargs["device_map"] = device_map
```
- The model is constructed with model_kwargs containing the device_map
value ONLY IF it is a string (lines 893-903 of
transformers/src/transformers/pipelines/\__init__.py).
```python
if isinstance(model, str) or framework is None:
model_classes = {"tf": targeted_task["tf"], "pt": targeted_task["pt"]}
framework, model = infer_framework_load_model( ... , **model_kwargs, )
```
- Consequently, since a model object is already passed to the pipeline
function, the device_map variable from from_model_id is never used.
3. The device_map variable in from_model_id not only appears unused but
also causes errors. Without explicitly setting device=None, attempting
to load the model on multiple GPUs may result in the following error:
```
Device has 2 GPUs available. Provide device={deviceId} to
`from_model_id` to use available GPUs for execution. deviceId is -1
(default) for CPU and can be a positive integer associated with CUDA
device id.
Traceback (most recent call last):
File "foo.py", line 15, in <module>
llm = HuggingFacePipeline.from_model_id(
File
"foo\site-packages\langchain_huggingface\llms\huggingface_pipeline.py",
line 217, in from_model_id
pipeline = hf_pipeline(
File "foo\lib\site-packages\transformers\pipelines\__init__.py", line
1108, in pipeline
return pipeline_class(model=model, framework=framework, task=task,
**kwargs)
File "foo\lib\site-packages\transformers\pipelines\text_generation.py",
line 96, in __init__
super().__init__(*args, **kwargs)
File "foo\lib\site-packages\transformers\pipelines\base.py", line 835,
in __init__
raise ValueError(
ValueError: The model has been loaded with `accelerate` and therefore
cannot be moved to a specific device. Please discard the `device`
argument when creating your pipeline object.
```
This error occurs because, in from_model_id, the default values in from_model_id for device and device_map are -1 and None, respectively. It would passes the statement (`device_map is not None and device < 0`) and keep the device as -1 so the pipeline function later raises an error when trying to move a GPU-loaded model back to the CPU.
19eb82e68b/libs/community/langchain_community/llms/huggingface_pipeline.py (L204-L213)
If no one reviews your PR within a few days, please @-mention one of baskaryan, efriis, eyurtsev, ccurme, vbarda, hwchase17.
---------
Co-authored-by: William FH <13333726+hinthornw@users.noreply.github.com>
Co-authored-by: Erick Friis <erick@langchain.dev>
Co-authored-by: vbarda <vadym@langchain.dev>
This PR introduces a new `azure_ad_async_token_provider` attribute to
the `AzureOpenAI` and `AzureChatOpenAI` classes in `partners/openai` and
`community` packages, given it's currently supported on `openai` package
as
[AsyncAzureADTokenProvider](https://github.com/openai/openai-python/blob/main/src/openai/lib/azure.py#L33)
type.
The reason for creating a new attribute is to avoid breaking changes.
Let's say you have an existing code that uses a `AzureOpenAI` or
`AzureChatOpenAI` instance to perform both sync and async operations.
The `azure_ad_token_provider` will work exactly as it is today, while
`azure_ad_async_token_provider` will override it for async requests.
If no one reviews your PR within a few days, please @-mention one of
baskaryan, efriis, eyurtsev, ccurme, vbarda, hwchase17.
**Description:** Deprecated version of Chroma >=0.5.5 <0.5.12 due to a
serious correctness issue that caused some embeddings for deployments
with multiple collections to be lost (read more on the issue in Chroma
repo)
**Issue:** chroma-core/chroma#2922 (fixed by chroma-core/chroma##2923
and released in
[0.5.13](https://github.com/chroma-core/chroma/releases/tag/0.5.13))
**Dependencies:** N/A
**Twitter handle:** `@t_azarov`
Example updated for vectorstore ChromaDB.
If we want to apply multiple filters then ChromaDB supports filters like
this:
Reference: [ChromaDB
filters](https://cookbook.chromadb.dev/core/filters/)
Thank you.
Thank you for contributing to LangChain!
**Description:** Box AI can return responses, but it can also be
configured to return citations. This change allows the developer to
decide if they want the answer, the citations, or both. Regardless of
the combination, this is returned as a single List[Document] object.
**Dependencies:** Updated to the latest Box Python SDK, v1.5.1
**Twitter handle:** BoxPlatform
- [x] **Add tests and docs**: If you're adding a new integration, please
include
1. a test for the integration, preferably unit tests that do not rely on
network access,
2. an example notebook showing its use. It lives in
`docs/docs/integrations` directory.
- [x] **Lint and test**: Run `make format`, `make lint` and `make test`
from the root of the package(s) you've modified. See contribution
guidelines for more: https://python.langchain.com/docs/contributing/
Additional guidelines:
- Make sure optional dependencies are imported within a function.
- Please do not add dependencies to pyproject.toml files (even optional
ones) unless they are required for unit tests.
- Most PRs should not touch more than one package.
- Changes should be backwards compatible.
- If you are adding something to community, do not re-import it in
langchain.
If no one reviews your PR within a few days, please @-mention one of
baskaryan, efriis, eyurtsev, ccurme, vbarda, hwchase17.
Co-authored-by: Erick Friis <erick@langchain.dev>
Given the current erroring behavior, every time we've moved a kwarg from
model_kwargs and made it its own field that was a breaking change.
Updating this behavior to support the old instantiations /
serializations.
Assuming build_extra_kwargs was not something that itself is being used
externally and needs to be kept backwards compatible
Chunking of the input array controlled by `self.chunk_size` is being
ignored when `self.check_embedding_ctx_length` is disabled. Effectively,
the chunk size is assumed to be equal 1 in such a case. This is
suprising.
The PR takes into account `self.chunk_size` passed by the user.
---------
Co-authored-by: Erick Friis <erick@langchain.dev>
**Description:** Add support to delete documents automatically from the
caches & chat message history by adding a new optional parameter, `ttl`.
- [x] **Add tests and docs**: If you're adding a new integration, please
include
1. a test for the integration, preferably unit tests that do not rely on
network access,
2. an example notebook showing its use. It lives in
`docs/docs/integrations` directory.
- [x] **Lint and test**: Run `make format`, `make lint` and `make test`
from the root of the package(s) you've modified. See contribution
guidelines for more: https://python.langchain.com/docs/contributing/
---------
Co-authored-by: Nithish Raghunandanan <nithishr@users.noreply.github.com>
Co-authored-by: Erick Friis <erick@langchain.dev>
In the previous implementation, `skip_count` was counting all the
documents in the collection. Instead, we want to filter the documents by
`session_id` and calculate `skip_count` by subtracting `history_size`
from the filtered count.
---------
Co-authored-by: Chester Curme <chester.curme@gmail.com>
## Description
By default, `HuggingFaceEndpoint` instantiates both the
`InferenceClient` and the `AsyncInferenceClient` with the
`"server_kwargs"` passed as input. This is an issue as both clients
might not support exactly the same kwargs. This has been highlighted in
https://github.com/huggingface/huggingface_hub/issues/2522 by
@morgandiverrez with the `trust_env` parameter. In order to make
`langchain` integration future-proof, I do think it's wiser to forward
only the supported parameters to each client. Parameters that are not
supported are simply ignored with a warning to the user. From a
`huggingface_hub` maintenance perspective, this allows us much more
flexibility as we are not constrained to support the exact same kwargs
in both clients.
## Issue
https://github.com/huggingface/huggingface_hub/issues/2522
## Dependencies
None
## Twitter
https://x.com/Wauplin
---------
Co-authored-by: Erick Friis <erick@langchain.dev>
`unstructured.partition.auto.partition` supports a `url` kwarg, but
`url` in `UnstructuredLoader.__init__` is reserved for the server URL.
Here we add a `web_url` kwarg that is passed to the partition kwargs:
```python
self.unstructured_kwargs["url"] = web_url
```
