github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-08-02 01:23:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

community: Cap AzureOpenAIEmbeddings chunk_size at 2048 instead of 16. (#25852)

Browse Source 
**Description:** Within AzureOpenAIEmbeddings there is a validation to
cap `chunk_size` at 16. The value of 16 is either an old limitation or
was erroneously chosen. I have checked all of the `preview` and `stable`
releases to ensure that the `embeddings` endpoint can handle 2048
entries
[Azure/azure-rest-api-specs](https://github.com/Azure/azure-rest-api-specs/tree/main/specification/cognitiveservices/data-plane/AzureOpenAI/inference).
I have also found many locations that confirm this limit should be 2048:
-
https://learn.microsoft.com/en-us/azure/ai-services/openai/reference#embeddings
-
https://learn.microsoft.com/en-us/azure/ai-services/openai/quotas-limits

**Issue:** fixes #25462
This commit is contained in:
Kyle Winkelman 2024-08-29 11:48:04 -05:00 committed by GitHub
parent 08c9c683a7
commit 201bdf7148
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
libs/community/langchain_community/embeddings/azure_openai.py
View File

@ -91,10 +91,10 @@ class AzureOpenAIEmbeddings(OpenAIEmbeddings):
values["azure_ad_token"] = values.get("azure_ad_token") or os.getenv(
"AZURE_OPENAI_AD_TOKEN"
)
# Azure OpenAI embedding models allow a maximum of 16 texts
# Azure OpenAI embedding models allow a maximum of 2048 texts
# at a time in each batch
# See: https://learn.microsoft.com/en-us/azure/ai-services/openai/reference#embeddings
values["chunk_size"] = min(values["chunk_size"], 16)
values["chunk_size"] = min(values["chunk_size"], 2048)
try:
import openai # noqa: F401
except ImportError:

6
libs/community/langchain_community/embeddings/openai.py
View File

@ -307,10 +307,10 @@ class OpenAIEmbeddings(BaseModel, Embeddings):
)
if values["openai_api_type"] in ("azure", "azure_ad", "azuread"):
default_api_version = "2023-05-15"
# Azure OpenAI embedding models allow a maximum of 16 texts
# at a time in each batch
# Azure OpenAI embedding models allow a maximum of 2048
# texts at a time in each batch
# See: https://learn.microsoft.com/en-us/azure/ai-services/openai/reference#embeddings
values["chunk_size"] = min(values["chunk_size"], 16)
values["chunk_size"] = min(values["chunk_size"], 2048)
else:
default_api_version = ""
values["openai_api_version"] = get_from_dict_or_env(