Skip to content
This repository was archived by the owner on Oct 25, 2024. It is now read-only.

Improve MPT series SQ #1640

Merged
merged 3 commits into from
Jun 28, 2024
Merged

Improve MPT series SQ #1640

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
4ce5cf7
improve SQ mpt
changwangss Jun 27, 2024
f1dc6b8
Update modeling_auto.py
changwangss Jun 27, 2024
c4450af
fix mpt achitectures
changwangss Jun 28, 2024
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
7 changes: 7 additions & 0 deletions intel_extension_for_transformers/transformers/modeling/modeling_auto.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -840,6 +840,12 @@ def forward(self, input: torch.Tensor) -> tuple[torch.Tensor, None]:
or device_map == torch.device("cpu")
) and model.config.model_type == "chatglm":
model = model.float()
if (
not torch.cuda.is_available()
or device_map == "cpu"
or device_map == torch.device("cpu")
) and model.config.model_type == "mpt":
model.config.architectures = ["MptForCausalLM"]
model.eval()
model_type = model.config.model_type.replace("_", "-")

Expand Down Expand Up @@ -1077,6 +1083,7 @@ def calib_func(model):
recipes=quantization_config.recipes,
example_inputs=example_inputs,
)

model = quantization.fit(
model,
conf,
Expand Down