mistral: comment batching param (#20868)

Addresses #20523
This commit is contained in:
Erick Friis 2024-04-24 17:38:21 -07:00 committed by GitHub
parent 7c5063ef60
commit 5da9dd1195
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -18,6 +18,10 @@ from tokenizers import Tokenizer # type: ignore
logger = logging.getLogger(__name__)
MAX_TOKENS = 16_000
"""A batching parameter for the Mistral API. This is NOT the maximum number of tokens
accepted by the embedding model for each document/chunk, but rather the maximum number
of tokens that can be sent in a single request to the Mistral API (across multiple
documents/chunks)"""
class DummyTokenizer: