Fix `text-generation-server quantize` (#2103)
The subcommand did not work due to some broken imports.
This commit is contained in:
parent
bcb3faa1c2
commit
197c47a302
|
@ -316,7 +316,7 @@ def quantize(
|
|||
logger_level=logger_level,
|
||||
json_output=json_output,
|
||||
)
|
||||
from text_generation_server.utils.gptq.quantize import quantize
|
||||
from text_generation_server.layers.gptq.quantize import quantize
|
||||
|
||||
quantize(
|
||||
model_id=model_id,
|
||||
|
|
|
@ -12,7 +12,7 @@ from huggingface_hub import HfApi
|
|||
from accelerate import init_empty_weights
|
||||
from text_generation_server.utils import initialize_torch_distributed, Weights
|
||||
from text_generation_server.utils.hub import weight_files
|
||||
from text_generation_server.utils.gptq.quant_linear import QuantLinear
|
||||
from text_generation_server.layers.gptq.quant_linear import QuantLinear
|
||||
from loguru import logger
|
||||
from typing import Optional
|
||||
|
||||
|
|
Loading…
Reference in New Issue