Skip to content

Eora #1302

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
Merged
merged 381 commits into from
Feb 21, 2025
Merged

Eora #1302

merged 381 commits into from
Feb 21, 2025

Conversation

@Qubitium
Copy link
Collaborator

@Qubitium Qubitium commented Feb 19, 2025 •
edited
Loading

Task TODO post merge:

  • Make sure Lora output has same format adapter_model.safetnsors + config that is compatible with HF/vLLM/SGLang lora code

Qubitium and others added 30 commits February 13, 2025 13:35
@Qubitium
fix override
dba585e
@Qubitium
simplify
38880f4
@Qubitium
fix missing modules item
437c939
@Qubitium
breaking: fix module.state update
9321b5b
@Qubitium
fix state should contain both W and WQ
5556f87
@CSY-ModelCloud
fix no super() for class obj
879b464
@CSY-ModelCloud
remove get attr
47840e4
@ZX-ModelCloud
call LoopProcessor.post_process()
89bf739 
Signed-off-by: ZX-ModelCloud <[email protected]>
@ZX-ModelCloud
Merge remote-tracking branch 'origin/eora' into eora
0930b52
@Qubitium
call processor.finalize
d01b6fb
@ZX-ModelCloud
Correctly call methods from self.gptq_model
e8ede3a 
Signed-off-by: ZX-ModelCloud <[email protected]>
@ZX-ModelCloud
Merge remote-tracking branch 'origin/eora' into eora
07f2918
@Qubitium
rename to calibration_data
ed7496d
@Qubitium
cleanup pack()..no need to clone weights..use T instead of t()
503b753
@ZX-ModelCloud
LoopProcessor add model_finalize()
238b2d3 
Signed-off-by: ZX-ModelCloud <[email protected]>
@ZX-ModelCloud
Merge remote-tracking branch 'origin/eora' into eora
51a4945
@Qubitium
cleanup pack()..rename var for clarity
aa59e4b
@Qubitium
pop wq from state
c322b95
@Qubitium
clean code..de-indent logic
74fd176
@Qubitium
add safety code to store original in/out features of W in NamedModule…
cf2fef1 
… state since the weight will be heavily changed during quant
@Qubitium
add stats() api and stats fields to processor
9d0273c
@Qubitium
ruff
e38c9ed
@ZX-ModelCloud
Fix circular import
fb42630 
Signed-off-by: ZX-ModelCloud <[email protected]>
@Qubitium
add license
17ee762
@ZX-ModelCloud
Merge remote-tracking branch 'origin/eora' into eora
b9d2f63
@ZX-ModelCloud
Merge remote-tracking branch 'origin/eora' into eora
f4221c1
@CSY-ModelCloud
add clearml back
8bbdf47
@ZX-ModelCloud
fix NamedModule.__getattr__() error
4d98b3b 
Signed-off-by: ZX-ModelCloud <[email protected]>
@Qubitium
add require_fwd property to processor
9872e7f
@Qubitium
simplify
5db8f02
CSY-ModelCloud and others added 28 commits February 20, 2025 17:33
@CSY-ModelCloud
fix checking
f90eb14
@Qubitium
fix env must be before torch imports
ecb9c53 
Signed-off-by: Qubitium <[email protected]>
@CSY-ModelCloud
move PYTORCH_ENABLE_MPS_FALLBACK to top
55ce173
@ZX-ModelCloud
ovis model require transformers<=4.48.3
a048815
@CSY-ModelCloud
print expected value
d04a9a3
@CSY-ModelCloud
[CI] fix names
b470f9a
@CSY-ModelCloud
[CI] fix xpu env reinstalled torch
36d4a13
@Qubitium
torch kernel will enable compile optimizations by default for torch 2…
b5e4820 
….6.0

Signed-off-by: Qubitium <[email protected]>
@Qubitium
fix transformers compat
fc0c518 
Signed-off-by: Qubitium <[email protected]>
@Qubitium
disable exllama kernel from quantization (remove from packable)
d709924 
Signed-off-by: Qubitium <[email protected]>
@CSY-ModelCloud
fix evalplus try toString a Decoder
96ca366
@CSY-ModelCloud
replace subprocess run by raising an error
ac7596e
@Qubitium
fix ci test_dynamic scores
f5ec991 
Signed-off-by: Qubitium <[email protected]>
@Qubitium
cleanup eora test
d27422b 
Signed-off-by: Qubitium <[email protected]>
@CSY-ModelCloud
fix sglang' transformers error
59eeca5
@ZX-ModelCloud
OVIS is compatible with transformers v4.49.0
65969b3
@CSY-ModelCloud
move ipex to new test files
9a3b6fc
@Qubitium
Update ovis.py
13d7f43
@CSY-ModelCloud
decrease batch to 16
6f4e35d
@Qubitium
Merge branch 'main' into eora
d00412f
@Qubitium
format
94ff1b7 
Signed-off-by: Qubitium <[email protected]>
@Qubitium
logs
83ba0ca 
Signed-off-by: Qubitium <[email protected]>
@Qubitium
fix ci lora config test
762cf4e 
Signed-off-by: Qubitium <[email protected]>
@Qubitium
fix ci: dynamic
e52c356 
Signed-off-by: Qubitium <[email protected]>
@Qubitium
fix ci: opt expects exllama when triton is used for quant
2b30708 
Signed-off-by: Qubitium <[email protected]>
@Qubitium
fix ci: transformers test oom
d36a645 
Signed-off-by: Qubitium <[email protected]>
@nbasyl
Add some comments to eora.py
5d2e5c0
@nbasyl
add comments to eora.py
406037c
@Qubitium Qubitium merged commit 7939d1a into main Feb 21, 2025
3 of 5 checks passed
@Qubitium Qubitium deleted the eora branch March 4, 2025 12:19
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@Qubitium @CSY-ModelCloud @ZX-ModelCloud @nbasyl