Support Qwen3 Embedding & Reranker on Gaudi for aice/v1.21.0 branch #1456

Open

gyou2021 wants to merge 19 commits intoHabanaAI:aice/v1.21.0from

gyou2021:gyou/aice/v1.21.0/qwen3

Commits on May 27, 2025

Optimized Qwen3 and Qwen3-MoE on Gaudi, supporting BF16 and inc based FP8 inference.
gyou2021
committed
Optimized Qwen3 and Qwen3-MoE on Gaudi, supporting BF16 and INC based FP8 inference.
gyou2021
committed

Commits on Jun 11, 2025

Commits on Jun 13, 2025

Merge remote-tracking branch 'upstream/aice/v1.21.0' into gyou/aice/v1.21.0/qwen3
gyou2021
committed
Merge remote-tracking branch 'upstream/aice/v1.21.0' into gyou/aice/v1.21.0/qwen3
gyou2021
committed
Supported EP of INC FP8 quant of Qwen3-235B-A22B.
gyou2021
committed

Commits on Jun 16, 2025

Commits on Jun 19, 2025

Commits on Jun 20, 2025

Commits on Jun 21, 2025

Fixed the bug in the qwen3 reranker score.
gyou2021
committed