v0.11.0

What's Changed

• release: Release 0.10.0post1 by @rebel-seinpark in #488
• Hotfix/deprecated logics by @rebel-thkim in #489
• other: bugfix in release pytest by @rebel-kblee in #491
• dependency: update torch version (2.9.0) by @rebel-kblee in #442
• dependency: transformers to 4.57.6 by @rebel-thkim in #476
• model: hotfix for gpt-oss bug by @rebel-thkim in #492
• model: gpt-oss hotfix (back merge) by @rebel-thkim in #493
• other: remove package with cpu dependency by @rebel-kblee in #494
• dependency: Update dependency rebel-compiler to 0.10.1.dev61+g264b62bf.prod by @rebel-thkim in #466
• dependency: Update dependency rebel-compiler to 0.10.1.dev77+gb96afccc.prod by @rebel-thkim in #498
• other: optimize package install in pytest by @rebel-thkim in #497
• model: fix mixtral & detr by @rebel-kblee in #478
• dependency: Update dependency rebel-compiler to 0.10.1.dev88+gdc4b08c4.prod by @rebel-thkim in #499
• dependency: Update dependency rebel-compiler to 0.10.1.dev106+gf49ca21a.prod by @rebel-thkim in #501
• dependency: Update dependency rebel-compiler to 0.10.1.dev131+g313992e8.prod by @rebel-thkim in #502
• dependency: Update dependency rebel-compiler to 0.10.1.dev144+g001fcb47.prod by @rebel-thkim in #503
• dependency: Update dependency rebel-compiler to 0.10.1.dev157+g4f65e832.prod by @rebel-thkim in #504
• dependency: Update dependency rebel-compiler to 0.10.1.dev209+ga5927402.prod by @rebel-thkim in #505
• dependency: Update dependency rebel-compiler to 0.10.1.dev216+gd0041304.prod by @rebel-thkim in #506
• model: gpt-oss not support bf16 yet by @rebel-thkim in #508
• dependency: Update dependency rebel-compiler to 0.10.1.dev225+gef394c3f.prod by @rebel-thkim in #507
• release: resolve conflicts between dev & main by @rebel-thkim in #511
• release: resolve conflicts between dev & main (final) by @rebel-thkim in #513
• hotfix: sync with dev by @rebel-thkim in #516
• release: v0.10.1 by @rebel-thkim in #518
• release: release v0.10.1 by @rebel-thkim in #514
• Revert "Revert "model: gpt-oss not support bf16 yet (#508)"" by @rebel-thkim in #520
• release: add missed commit in dev by @rebel-thkim in #523
• release: missed commit in v0.10.1 by @rebel-thkim in #521
• Dev by @rebel-thkim in #522
• dependency: Update dependency rebel-compiler to 0.10.2.dev16+ge826d8d5.prod by @rebel-thkim in #509
• model: Explicit quantized layer by @rebel-jongho in #268
• other(cli): Support for hf model kwargs to optimum_rbln_cli by @rebel-jongho in #426
• refactor: error log typo by @rebel-eunji in #527
• dependency: diffusers to 0.37.0 by @rebel-thkim in #526
• model: support qwen3vl vlm by @rebel-kblee in #500
• dependency: update torch version and target soc by @rebel-kblee in #525
• other: remove compiler check in ci by @rebel-thkim in #528
• model: additional features in qwen3vl by @rebel-kblee in #529
• other(pytest): fix to avoid sharing config by @rebel-kblee in #531
• model: gpt_oss dtype fix by @rebel-thkim in #530
• dependency: Update dependency rebel-compiler to 0.10.2.dev167+ga8769bfb.prod by @rebel-thkim in #524
• dependency: Update dependency rebel-compiler to 0.10.2.dev243+g83583dcb.prod by @rebel-thkim in #532
• model: fix multi batch video input handling by @rebel-kblee in #535
• dependency: Update dependency rebel-compiler to 0.10.2.dev264+g1fb8dffb.prod by @rebel-thkim in #534
• release: v0.10.2 by @rebel-thkim in #541
• other: change ReadMe.md by @rebel-thkim in #542
• other: change ReadMe.md (#542) by @rebel-thkim in #543
• other: torch version 2.9.1 -> 2.10.0 by @rebel-seinpark in #545
• doc(t5): fix Optimum RBLN Overview link for MkDocs by @rebel-junamsong in #550
• model: update qwen3 vl for disaggregate encoder in vllm by @rebel-kblee in #551
• model: fix page table manager with bucketing by @rebel-kblee in #547
• model: qwen3vl vit encoder support with tp by @rebel-kblee in #537
• other(err msg): add message, fix class type by @rebel-jongho in #552
• other: add timeout in package install by @rebel-kblee in #553
• release: v0.10.3 by @rebel-thkim in #554
• other: fix circular import error by @rebel-seinpark in #556
• model: Moe custom op args by @rebel-thkim in #557
• other: skip some pytest which cannot fix in compiler by @rebel-kblee in #561
• model(whisper): allocate extra block for prefill padding (vllm) by @rebel-eunji in #560
• dependency: Update dependency rebel-compiler to 0.10.4.dev201+g84f77eb1.prod by @rebel-thkim in #539
• model: lower minimum flash attention seq len and partition length by @rebel-thkim in #569
• model: vlm's max seq len by @rebel-thkim in #565
• other(test): fix disallowed flash-attn tests after lowered minimums (#569) by @rebel-jongho in #571
• dependency: Update dependency rebel-compiler to 0.10.4.dev321+gc1b63c2d.prod by @rebel-thkim in #568
• other: attention_validation with variable by @rebel-thkim in #577
• release: v0.10.4 by @rebel-thkim in #578
• dependency: Update dependency rebel-compiler to 0.10.5.dev109+ga425966d.prod by @rebel-thkim in #573
• dependency: Update dependency rebel-compiler to 0.10.5.dev143+g4b9a219c.prod by @rebel-thkim in #580
• model: gemma3 minor fix by @rebel-thkim in #570
• dependency: update latest transformers (5.8.1) & torch (2.11.0) by @rebel-kblee in #549
• model: fix issue in bump up #549 by @rebel-kblee in #583
• model: fix bumpup remaining issue by @rebel-kblee in #585
• model: sync with ccl all2all signature by @rebel-kblee in #562
• model: fix gpt oss ccl padding by @rebel-kblee in #586
• model: dev fix dtype in moe signature by @rebel-jangys in #587
• model: fix failure models in ci 0612 by @rebel-kblee in #588
• model: fix failed case in 0615 (whisper with multi batch) by @rebel-kblee in #590
• dependency: Update dependency rebel-compiler to 0.10.5.dev320+ge2df577b.prod by @rebel-thkim in #581
• performance(decoderonly): per-chiplet KV cache block estimation to prevent CR03 OOM (SR-244) by @rebel-jongho in #584
• model: bump-up with cosmos-guardrail==0.3.1 by @rebel-seinpark in #591
• model: support exaone 4.5 generally via bump up by @rebel-kblee in #593
• dependency: diffusers to 0.38.0 by @rebel-thkim in #558
• model: add get_image_features() in vlm models by @rebel-kblee in #594
• model: sync with act_type custom kernel by @rebel-kblee in #592
• model(paligemma): tolerate the labels kwarg from the processor during generation by @rebel-junamsong in #598
• model(distilbert): accept and ignore token_type_ids in QA forward by @rebel-junamsong in #596
• refactor: rename tensor_parallel_size to num_devices by @rebel-jongho in #597
• model: support max_seq_len bucketing for transformer encoders by @rebel-thkim in #599
• dependency: Update dependency rebel-compiler to 0.10.5.dev415+gd86eeb83.prod by @rebel-thkim in #595
• performance: normalize weights to contiguous before compile to cut host peak memory by @rebel-jongho in #600
• model: support gemma4 by @rebel-dkhong in #563
• model: honor deprecated tensor_parallel_size in nested submodule config by @rebel-jongho in #601
• refactor: Gemma3 effcient prefill by @rebel-dkhong in #602
• refactor: resolve shape mismatch when extending cache_position on vLLM by @rebel-dkhong in #604
• refactor: remove default bucketing for image_prefill_chunk_size & max_soft_tokens by @rebel-dkhong in #605
• model(seq2seq): raise instead of silently truncating over-long encoder input by @rebel-jongho in #606
• other: Revert "model(seq2seq): raise instead of silently truncating over-lon… by @rebel-jongho in #607
• release: v0.11.0 by @rebel-kblee in #608
• release: v0.11.0 by @rebel-kblee in #611

Full Changelog: v0.10.0...v0.11.0