Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Bugfix][VLM] Make apply_fp8_linear work with >2D input #9812

Merged
merged 2 commits into from
Oct 30, 2024
Merged

[Bugfix][VLM] Make apply_fp8_linear work with >2D input #9812

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
1032cf3
Make apply_fp8_linear work with >2D input
mgoin Oct 29, 2024
ebd38af
Update w8a8_utils.py
mgoin Oct 29, 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
33 changes: 20 additions & 13 deletions vllm/model_executor/layers/quantization/utils/w8a8_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -96,21 +96,26 @@ def apply_fp8_linear(
# If dynamic, layer.input_scale is None and x_scale computed from x.
# If static, layer.input_scale is scalar and x_scale is input_scale.

# View input as 2D matrix for fp8 methods
input_2d = input.view(-1, input.shape[-1])
output_shape = [*input.shape[:-1], weight.shape[1]]

# cutlass_scaled_mm supports per tensor/channel W and per tensor/token A
if cutlass_fp8_supported:
qinput, x_scale = ops.scaled_fp8_quant(
input,
input_2d,
input_scale,
scale_ub=input_scale_ub,
use_per_token_if_dynamic=use_per_token_if_dynamic)

# Fused GEMM_DQ
return ops.cutlass_scaled_mm(qinput,
weight,
out_dtype=input.dtype,
scale_a=x_scale,
scale_b=weight_scale,
bias=bias)
output = ops.cutlass_scaled_mm(qinput,
weight,
out_dtype=input.dtype,
scale_a=x_scale,
scale_b=weight_scale,
bias=bias)
return output.view(*output_shape)

# torch.scaled_mm supports per tensor weights + activations only
# so fallback to naive if per channel or per token
Expand All @@ -119,7 +124,7 @@ def apply_fp8_linear(
# for matrices with batch dimension > 16.
# This could change in the future.
qinput, x_scale = ops.scaled_fp8_quant(
input,
input_2d,
input_scale,
num_token_padding=17,
use_per_token_if_dynamic=use_per_token_if_dynamic)
Expand All @@ -138,8 +143,10 @@ def apply_fp8_linear(
# A fix for discrepancy in scaled_mm which returns tuple
# for torch < 2.5 and a single value in torch >= 2.5
if type(output) is tuple and len(output) == 2:
return torch.narrow(output[0], 0, 0, input.shape[0])
return torch.narrow(output, 0, 0, input.shape[0])
output = output[0]

return torch.narrow(output, 0, 0,
input_2d.shape[0]).view(*output_shape)

else:
# Fallback for channelwise case, where we use unfused DQ
Expand Down Expand Up @@ -176,15 +183,15 @@ def apply_fp8_linear(
if type(output) is tuple and len(output) == 2:
output = output[0]
# Unpad (undo num_token_padding)
output = torch.narrow(output, 0, 0, input.shape[0])
x_scale = torch.narrow(x_scale, 0, 0, input.shape[0])
output = torch.narrow(output, 0, 0, input_2d.shape[0])
x_scale = torch.narrow(x_scale, 0, 0, input_2d.shape[0])

# DQ
# C = sw * sx * (X * W) + bias
output = output * x_scale * weight_scale.t()
if bias is not None:
output = output + bias
return output.to(dtype=input.dtype)
return output.to(dtype=input.dtype).view(*output_shape)


def apply_int8_linear(
Expand Down