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

Jump to bottom

Add support for Qwen #103

Merged
merged 12 commits into from
Dec 5, 2023
Merged

Add support for Qwen #103

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
6832881
WIP: qwen
tgaddair Dec 5, 2023
c51f9ee
MLP and attention
tgaddair Dec 5, 2023
3ac5ab4
Model
tgaddair Dec 5, 2023
aa79757
Generalize layer_weights
tgaddair Dec 5, 2023
52d973f
Fixed init
tgaddair Dec 5, 2023
1b1895b
Constants
tgaddair Dec 5, 2023
1c8d28c
Generalized
tgaddair Dec 5, 2023
ceb95c0
Updated deps
tgaddair Dec 5, 2023
13eaaf9
Updated poetry lock
tgaddair Dec 5, 2023
285fbc1
Revert
tgaddair Dec 5, 2023
3939482
Updated license
tgaddair Dec 5, 2023
36f76e7
Fixed adapter loading
tgaddair Dec 5, 2023
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
Next Next commit
WIP: qwen
  • Loading branch information
@tgaddair
tgaddair committed Dec 5, 2023
commit 68328814fce079e23b87da9196991839077f02be
15 changes: 15 additions & 0 deletions server/lorax_server/models/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,6 +52,7 @@
from lorax_server.models.flash_llama import (
FlashLlama,
)
from lorax_server.models.flash_qwen import FlashQwen
from lorax_server.models.flash_santacoder import (
FlashSantacoderSharded,
)
Expand All @@ -65,6 +66,7 @@
__all__.append(FlashRWSharded)
__all__.append(FlashSantacoderSharded)
__all__.append(FlashLlama)
__all__.append(FlashQwen)

MISTRAL = True
try:
Expand Down Expand Up @@ -281,6 +283,19 @@ def get_model(
trust_remote_code=trust_remote_code,
)
raise NotImplementedError("Mistral model requires flash attention v2")

if model_type == "qwen":
if MISTRAL:
return FlashQwen(
model_id,
adapter_id,
adapter_source,
revision,
quantize=quantize,
dtype=dtype,
trust_remote_code=trust_remote_code,
)
raise NotImplementedError("Qwen model requires flash attention v2")

if model_type == "opt":
return OPTSharded(
Expand Down
Loading