[torch.compile] Refactor Attention Quant Fusion Pass and Remove Boilerplate (#37373)

BadrBasowid committed 7 months ago

[CI/Build] Resolve a dependency deadlock when installing the test dependencies used in CI (#37766)

Run Yu committed 7 months ago

[4/n] Migrate FP4/W4A8 CUTLASS kernels to torch stable ABI (#37503)

mikaylagawarecki committed 7 months ago

[Quantization][Autoround][XPU] Add `W4A16` Support (#37986)

Yi Liu committed 7 months ago

[Bugfix] Fix FusedMoE weight loading with padded hidden dimensions (#37010)

SandishKumarHN committed 7 months ago

[Fix] handle PaddleOCR-VL image processor max_pixels across Transformers v4/v5 (#38629)

zhang-prog committed 7 months ago

[Bugfix] clamp dA_cumsum differences to prevent Inf in Mamba2 SSD kernels (#37501)

Jingu Kang committed 7 months ago

[Bugfix][Async] Fix async spec decoding with hybrid models (#38556)

Matthew Bonanni committed 7 months ago

[Docs] PD with Nixl compat matrix (#38628)

Nicolò Lucchesi committed 7 months ago

[Online Quant] [QeRL] Minor code cleanup (#38574)

Kyle Sayers committed 7 months ago

replace cuda_device_count_stateless() to current_platform.device_count() (#37841)

wliao2 committed 7 months ago

[CI] Avoid concurrent docker pull in intel XPU CI runners to prevent rate limit issues (#38594)

wenjun liu committed 7 months ago

[Bugfix][CI] Skip flaky `test_eagle` test (#38566)

Nicolò Lucchesi committed 7 months ago

Fix MLA runs when use_inductor_graph_partition=True (#38631)

ElizaWszola committed 7 months ago

[CI] fix LM Eval Qwen3.5 Models (B200) (#38632)

Jiangyun Zhu committed 7 months ago

[XPU]move testing dependencies from Dockerfile to xpu-test.in (#38596)

sihao_li committed 7 months ago

[EPLB] Add alternative communication for EPLB weight exchange (#33176)

Ilya Markov committed 7 months ago

[EPD] update EPD script arguments (#36742)

liuzhenwei committed 7 months ago

[CI Failure] pin colmodernvbert revision (#38612)

wang.yuqi committed 7 months ago

Fix document of torchrun_example.py (#31113)

Maosheng Liao committed 7 months ago