PR #7853 Add EXAONE 4.0 model support for Inference V2

Add EXAONE 4.0 model support for Inference V2 #7853

tohtana merged 5 commits into deepspeedai:master from Bias92:add-exaone4-inference-v2

Bias92 requested a review from

hwchen2017 38 days ago

Bias92 requested a review from

tohtana 38 days ago

Bias92 force pushed from bd52e9d0 to 400d05a3 38 days ago

chatgpt-codex-connector commented on 2026-02-14

tohtana commented on 2026-02-15

PKUWZP requested a review from

PKUWZP 36 days ago

Add EXAONE 4.0 model support for Inference V2

2c312690

Fix QK-norm to use local head counts for TP compatibility

81f96e2b

Apply yapf formatting

66a7344a

Use n_heads_q_local and n_heads_kv_local for GQA compatibility

8ece3c17

Bias92 force pushed from fced31a9 to 8ece3c17 36 days ago

tohtana approved these changes on 2026-02-17

tohtana enabled auto-merge (squash) 35 days ago

Merge branch 'master' into add-exaone4-inference-v2

14dc60b3

tohtana merged f3a9819c into master 35 days ago

Reviewers

tohtana

hwchen2017

PKUWZP

Assignees

No one assigned

Labels

None yet

Milestone

No milestone