Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 2 additions & 3 deletions setup.py
Original file line number Diff line number Diff line change
Expand Up @@ -129,10 +129,9 @@ def localversion_func(version: ScmVersion) -> str:
),
("datasets>=4.0.0,<=4.6.0" if BUILD_TYPE == "release" else "datasets>=4.0.0"),
(
# auto-round 0.9.1 cannot work with accelerate <1.10.0
"auto-round>=0.9.6,<=0.10.2"
"auto-round>=0.10.2,<=0.10.2"
if BUILD_TYPE == "release"
else "auto-round>=0.9.6"
else "auto-round>=0.10.2"
),
(
"accelerate>=1.6.0,<=1.12.0"
Expand Down
4 changes: 2 additions & 2 deletions src/llmcompressor/modifiers/autoround/base.py
Original file line number Diff line number Diff line change
Expand Up @@ -262,7 +262,7 @@ def apply_autoround(self, state, subgraph):

# Build kwargs for AutoRound initialization
ar_quant_scheme = self._mapping_config_to_autoround()
fp_layers = self.get_unquantized_layer_names(decoding_layer)
ignore_layers = self.get_unquantized_layer_names(decoding_layer)
kwargs = {
"tokenizer": "", # A placeholder
"scheme": ar_quant_scheme,
Expand All @@ -271,7 +271,7 @@ def apply_autoround(self, state, subgraph):
"enable_torch_compile": self.enable_torch_compile,
"batch_size": self.batch_size,
"device_map": self.device_ids,
"fp_layers": ",".join(fp_layers) if fp_layers else "",
"ignore_layers": ",".join(ignore_layers) if ignore_layers else "",
}

llmc_registered_qparams = self._preprocess_qparams(decoding_layer)
Expand Down