Simplify Tensor Parallel implementation with PyTorch TP #34184
ArthurZucker
merged
26 commits into
huggingface:main
from
kwen2501:tp_llama
Simplify Tensor Parallel implementation with PyTorch TP
e60fb879
ArthurZucker
commented
on 2024-10-21
Move tp_plan to config
fd7f7c72
ArthurZucker
commented
on 2024-10-24
kmehant
commented
on 2024-10-25
kmehant
commented
on 2024-10-25
Merge remote-tracking branch 'origin/main' into tp_llama
9224cabd
Lint
79cc524e
Format and warning
a2934b33
Disable copy-from check
a8fc418c
ArthurZucker
commented
on 2024-10-30
Conditionally get attr from config
e84a3889
make fix-copies
396d158c
Move base_model_tp_plan to PretrainedConfig
7b346b55
Merge remote-tracking branch 'origin/main' into tp_llama
d60679b5
Merge remote-tracking branch 'origin/main' into tp_llama
dda058af
kmehant
commented
on 2024-11-04
Move TP into from_pretrained
12fbbe70
Add device context for load
02c8c393
Do not serialize
073c521d
Move _tp_plan setting to post_init
db6e5eeb
Add has_tp_plan
5bb294ec
Add test_tp
290a7f18
Add 'Multi-gpu inference' doc
bd2e89c1
Merge remote-tracking branch 'origin/main' into tp_llama
4892ceff
Add backward support for device type identification
9648f316
Auto-detect accelerator
93ba2835
supports_tp_plan
73524c90
copyright year
f312e551
kwen2501
commented
on 2024-11-16
Merge remote-tracking branch 'origin/main' into tp_llama
ca93bdb9
Merge branch 'main' into tp_llama
dc2672f8
Fix copy
1e27d6f9
Login to write a write a comment.
Login via GitHub