github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-07-03 19:57:51 +00:00
Code Issues Packages Projects Releases Wiki Activity

llamacpp wrong default value passed for f16_kv (#3320)

Browse Source 
Fixes default f16_kv value in llamacpp; corrects incorrect parameter
passed.

See:
ba3959eafd/llama_cpp/llama.py (L33)

Fixes #3241
Fixes #3301
This commit is contained in:
Ivan Zatevakhin 2023-04-23 02:46:55 +01:00 committed by GitHub
parent 3a1bdce3f5
commit 77bb6c99f7
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
langchain/llms/llamacpp.py
View File

@ -31,13 +31,13 @@ class LlamaCpp(LLM):
"""Token context window."""
n_parts: int = Field(-1, alias="n_parts")
"""Number of parts to split the model into.
"""Number of parts to split the model into.
If -1, the number of parts is automatically determined."""
seed: int = Field(-1, alias="seed")
"""Seed. If -1, a random seed is used."""
f16_kv: bool = Field(False, alias="f16_kv")
f16_kv: bool = Field(True, alias="f16_kv")
"""Use half-precision for key/value cache."""
logits_all: bool = Field(False, alias="logits_all")
@ -50,7 +50,7 @@ class LlamaCpp(LLM):
"""Force system to keep model in RAM."""
n_threads: Optional[int] = Field(None, alias="n_threads")
"""Number of threads to use.
"""Number of threads to use.
If None, the number of threads is automatically determined."""
n_batch: Optional[int] = Field(8, alias="n_batch")