Default Branch

main 95bafdcccf · [docs] NVIDIA/Intel/CPU (#49650) · Updated 2026-10-28 08:29:05 +08:00

ZIP  TAR.GZ

Branches

tarekziade-mlinter-main-landing 23aa09b98a · Pin transformers-mlinter to main @ 5a85da7 · Updated 2026-10-28 16:44:27 +08:00	0 121		ZIP  TAR.GZ
jev-1-core c1b168ce2b · Tests: one decider per prompt/attention mode, revert unrelated colqwen2 line · Updated 2026-10-28 14:13:06 +08:00	156 16		ZIP  TAR.GZ
jev-2-kev 63ed92c687 · Merge branch 'jev-1-core' into jev-2-kev · Updated 2026-10-28 14:13:06 +08:00	156 29		ZIP  TAR.GZ
jev-3-von cab6f5acec · Merge branch 'jev-2-kev' into jev-3-von · Updated 2026-10-28 14:13:06 +08:00	156 44		ZIP  TAR.GZ
jev-4-clef 43f6539f9b · Merge branch 'jev-3-von' into jev-4-clef · Updated 2026-10-28 14:13:06 +08:00	156 64		ZIP  TAR.GZ
add-bolmo f6543fb633 · fix repo · Updated 2026-10-28 12:30:30 +08:00	5 11		ZIP  TAR.GZ
diffusion-sampler-protocol 627be246ed · Let a custom sampler own the temperature schedule · Updated 2026-10-28 11:28:34 +08:00	0 3		ZIP  TAR.GZ
gitattributes-linguist-generated e699b24971 · Mark generated files as linguist-generated, kept in sync by make fix-repo · Updated 2026-10-28 11:22:00 +08:00	0 1		ZIP  TAR.GZ
tp-mixin-persistent-ranks 58f08602d2 · Group the persistent ranks' code in one class · Updated 2026-10-28 10:06:15 +08:00	25 10		ZIP  TAR.GZ
fsdp-dtype-uniform-groups ca5e68fca8 · Make the bf16 training test work for models without a dtype plan · Updated 2026-10-28 10:04:30 +08:00	19 35		ZIP  TAR.GZ
llama_metaspace 73eba2dc3d · fix test · Updated 2026-10-28 08:32:48 +08:00	1 13		ZIP  TAR.GZ
modular_cleanup c4cf0e63c4 · moe sparseblock canonicals · Updated 2026-10-28 08:29:39 +08:00	39 20		ZIP  TAR.GZ
serge/fix/itf-0196584893e6-bd9a3f79 4f392a4b57 · [serge] Fix 2 integration tests for model seamless_m4t failing with other, raising TypeError (other (2)) · Updated 2026-10-28 08:03:06 +08:00	1 1		ZIP  TAR.GZ
lazy-package-distribution-mapping 7306cd0792 · Build the package distribution mapping lazily · Updated 2026-10-28 05:48:47 +08:00	1 1		ZIP  TAR.GZ
exporters-generation-split b6849092ad · Exporters: drop stage-4 leftovers from the decomposition and cache helpers · Updated 2026-10-28 05:46:13 +08:00	7 18		ZIP  TAR.GZ
run_amd_scheduled_ci_caller_testdeps cb827f6739 · [CI] MI300 test: important models on the rocm101-testdeps image (no Slack/report upload, mi355/mi250 not triggered) · Updated 2026-10-28 05:32:34 +08:00	3 3		ZIP  TAR.GZ
tarekziade-fix-bridgetower-processor-load dd94182127 · Resolve sub-processors of checkpoints whose config.json has no model_type · Updated 2026-10-28 02:11:53 +08:00	3 1		ZIP  TAR.GZ
build_ci_docker_image_amd_testdeps 03e8b582d3 · [CI] AMD docker test: build with onnxscript/bitsandbytes/peft, push to huggingface/transformers-pytorch-amd-gpu:rocm101-testdeps only · Updated 2026-10-28 02:03:01 +08:00	3 2		ZIP  TAR.GZ
tarekziade-fix-janus-str-text 0b437423aa · Fix JanusProcessor iterating a str text input character by character · Updated 2026-10-28 01:55:44 +08:00	3 1		ZIP  TAR.GZ
tarekziade-fix-luke-tied-lm-head d229e8eaa3 · Tie LUKE masked LM decoder weight to word embeddings · Updated 2026-10-28 01:50:06 +08:00	3 1		ZIP  TAR.GZ