Skip to content
This repository has been archived by the owner on Aug 30, 2024. It is now read-only.

[GPTQ Enhance] Fix convert_quantized model.py for dtype #158

Merged
merged 1 commit into from
Mar 8, 2024
+3 −3
Merged

[GPTQ Enhance] Fix convert_quantized model.py for dtype #158

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 1 addition & 1 deletion neural_speed/convert/convert_quantized_llama.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,7 +71,7 @@ def convert_to_q4_bestla_tensor(src_name, dst_name, model, fout, q_config, n_hea
int_weight = (int_weight - 8) * 16
gptq_scales = gptq_scales / 16
gptq_zeros = (gptq_zeros - 8) * 16
weight_dtype == "int4"
weight_dtype = "int4"

dst = np.zeros((int_weight.shape[0], int_weight.shape[1] * 4), dtype=np.int8)
int_weight = np.ascontiguousarray(int_weight.numpy())
Expand Down
2 changes: 1 addition & 1 deletion neural_speed/convert/convert_quantized_mistral.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -74,7 +74,7 @@ def convert_to_q4_bestla_tensor(src_name, dst_name, model, fout, q_config, n_hea
int_weight = (int_weight - 8) * 16
gptq_scales = gptq_scales / 16
gptq_zeros = (gptq_zeros - 8) * 16
weight_dtype == "int4"
weight_dtype = "int4"

dst = np.zeros((int_weight.shape[0], int_weight.shape[1] * 4), dtype=np.int8)
int_weight = np.ascontiguousarray(int_weight.numpy())
Expand Down
2 changes: 1 addition & 1 deletion neural_speed/convert/convert_quantized_mixtral.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,7 +80,7 @@ def convert_to_q4_bestla_tensor(src_name, dst_name, model, fout, q_config, n_hea
int_weight = (int_weight - 8) * 16
gptq_scales = gptq_scales / 16
gptq_zeros = (gptq_zeros - 8) * 16
weight_dtype == "int4"
weight_dtype = "int4"

dst = np.zeros((int_weight.shape[0], int_weight.shape[1] * 4), dtype=np.int8)
int_weight = np.ascontiguousarray(int_weight.numpy())
Expand Down
Loading