Skip to content

Commit

Permalink
Address final comments
Browse files Browse the repository at this point in the history
  • Loading branch information
pablomlago committed Jan 21, 2025
1 parent 1b90234 commit 06e5c1f
Show file tree
Hide file tree
Showing 4 changed files with 4 additions and 3 deletions.
1 change: 0 additions & 1 deletion src/brevitas/graph/equalize.py
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,6 @@
import torch.nn as nn
import torch.nn.utils.parametrize as parametrize

from brevitas import config
from brevitas import torch_version
from brevitas.fx import GraphModule
from brevitas.fx import Node
Expand Down
4 changes: 2 additions & 2 deletions src/brevitas_examples/llm/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -51,8 +51,8 @@ usage: main.py [-h] [--config CONFIG] [--model MODEL] [--seed SEED]
[--scaling-min-val SCALING_MIN_VAL] [--quant-sdpa]
[--functional-sdpa-quant] [--replace-mha]
[--weight-equalization] [--rotation {fx,layerwise,fused_no_fx}]
[--rotation-mode {had,ort}] [--rotation-orphan-sink]
[--rotation-sdpa-regions]
[--rotation-mode {had,ort}] [--optimize-rotations]
[--rotation-orphan-sink] [--rotation-sdpa-regions]
[--act-equalization {None,layerwise,fx}]
[--act-equalization-alpha ACT_EQUALIZATION_ALPHA]
[--load-awq LOAD_AWQ]
Expand Down
1 change: 1 addition & 0 deletions src/brevitas_examples/llm/config/default_template.yml
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,7 @@ model: facebook/opt-125m
no_float16: false
no_quantize: false
nsamples: 128
optimize_rotations: false
quant_sdpa: false
quantize_input_zero_point: false
quantize_last_layer: false
Expand Down
1 change: 1 addition & 0 deletions tests/brevitas_examples/llm_test_template.yml
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,7 @@ model: facebook/opt-125m
no_float16: false
no_quantize: false
nsamples: 128
optimize_rotations: false
quant_sdpa: false
quantize_input_zero_point: false
quantize_last_layer: false
Expand Down

0 comments on commit 06e5c1f

Please sign in to comment.