transformers
Feat: save_pretrained for tensor parallel (and other parallelisms) models
#37919
Merged
Overview
Commits
13
Changes
View On GitHub

Feat: save_pretrained for tensor parallel (and other parallelisms) models #37919

S1ro1 merged 13 commits into main from save-pretrained-dtensor
S1ro1
S1ro1 tmp: initial save pretrained with dtensors
c9412513
HuggingFaceDocBuilderDev
S1ro1 Feat: add correctness tests
fde7277b
S1ro1 S1ro1 requested a review from ArthurZucker ArthurZucker 284 days ago
S1ro1 Refactor: version checks
61c9f248
S1ro1 S1ro1 marked this pull request as ready for review 284 days ago
S1ro1 S1ro1 changed the title tmp: initial save pretrained with dtensors Feat: save_pretrained for model sharded with DTensors 284 days ago
ArthurZucker
ArthurZucker commented on 2025-05-05
S1ro1 Temp: 1:1 checkpoint llama4
5829b2e4
S1ro1 refactor
98f3f3fe
S1ro1 Tests
0b1ac766
S1ro1 S1ro1 changed the title Feat: save_pretrained for model sharded with DTensors Feat: save_pretrained for tensor parallel (and other parallelisms) models 276 days ago
S1ro1 Feat: works
8b316319
S1ro1 Style
ddbe4194
S1ro1
S1ro1 commented on 2025-05-13
ArthurZucker
ArthurZucker commented on 2025-05-15
S1ro1 Feat: version checks + minor fixes
34fa7f85
S1ro1 Style
7f84aeff
S1ro1 Fix: version checks in tests
f3a441d5
ArthurZucker
ArthurZucker approved these changes on 2025-05-19
S1ro1 Feat: move more stuff into tensor_parallel.py
3eb798c1
S1ro1 S1ro1 enabled auto-merge (squash) 266 days ago
S1ro1 Merge branch 'main' into save-pretrained-dtensor
dfed1ab6
S1ro1 S1ro1 merged 46a4b7c9 into main 266 days ago
S1ro1 S1ro1 deleted the save-pretrained-dtensor branch 266 days ago
BenjaminBossan
ydshieh

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Terms Of Use Privacy Policy Refund Policy Impressum