yangzhuxinyzx 15c302e85d [Model] Integrate current main before GLM MTP configuration checks ...

Signed-off-by: yangzhuxinyzx <153831768+yangzhuxinyzx@users.noreply.github.com>

2026-10-09 08:59:55 +08:00

..

fixtures

…

gguf

[Memory] Preserve dense precision for tied GGUF vocabulary heads

2026-10-07 23:43:20 +08:00

glm5next

[Test] Format the repeated GLM MTP conversion regression

2026-10-09 08:48:57 +08:00

language

…

multimodal

…

quantization

…

qwen3_5

…

qwen4_exp

[Bugfix][Qwen4Exp] Read the PLE row-gather policy when the layer is built

2026-10-09 08:47:28 +08:00

__init__.py

…

registry.py

…

test_adapters.py

…

test_bailing_moe_v3.py

…

test_deepseek_v4_fp8_capability.py

…

test_deepseek_v4_mega_moe.py

…

test_deepseek_v4_pipeline.py

…

test_deepseek_v4_rocm_swa_ragged.py

…

test_deepseek_v4_sm70_routes.py

[Perf][SM70][DeepSeek-V4] Bucket speculative decode graphs by context so the sparse indexer can take cuBLAS

2026-10-05 02:13:12 +08:00

test_deepseek_v4_sm70_sparse_policy.py

[Bugfix] Retain sparse attention policy during layer construction

2026-10-04 01:30:42 +08:00

test_dflash_context_slots.py

[Bugfix] Select DFlash context slots by attention layer name

2026-10-09 08:40:17 +08:00

test_gguf_download.py

…

test_initialization.py

…

test_language_model_cache_is_weak.py

…

test_mtp_skip_checkpoint_weight.py

…

test_oot_registration.py

…

test_pixtral_rope_compat.py

[Bugfix] Preserve Pixtral vision loading with Transformers 5

2026-10-04 15:20:46 +08:00

test_ple_optional_cuda_import.py

[Bugfix] Defer PLE CUDA bindings until GPU transport is used

2026-10-09 08:40:16 +08:00

test_qwen4_ple_placeholder.py

[Test] Isolate PLE placeholder role and lazy environment state

2026-10-03 11:36:32 +08:00

test_registry.py

…

test_terratorch.py

…

test_transformers.py

…

test_utils.py

…

test_vision.py

…

utils.py

…