423A35C7

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-11-06 19:02:17 +08:00

bd24350cbf [v1] add pair data converter (#9360)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-11-06 02:42:15 +08:00

bd30c0003b [train] fix denominator of ga in ksft loss (#9409)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-11-05 18:32:14 +08:00

8edd2622ce [docker] update npu dockerfile (#9407)

eaf963f67f [model] update kt code (#9406)

Compare 2 commits »

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-11-05 02:12:14 +08:00

56f45e826f [train] fix MPO re-weight (#9405)

14abb75126 [model] enable using FA in npu (#9397)

5a9939050e [model] add deepstack_merger_list to Qwen3-VL vision_model_keys (#9399)

Compare 3 commits »

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-11-04 18:02:19 +08:00

934b3084ee [train] KTransformers SFT as backend engine for LLaMA-Factory (#9400)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-11-04 01:42:14 +08:00

3ae15da9c0 [misc] lint code (#9395)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-11-03 17:32:15 +08:00

215580c77d [data] fix mm pluigin for qwen omni video training (#9388)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-10-30 23:48:09 +08:00

767b344fb4 [model] remove npu sdpa patch (#9368)

423A35C7 synced commits to main at 423A35C7/pytorch3d from mirror 2025-10-30 23:18:10 +08:00

2d4d345b6f Improve ball_query() runtime for large-scale cases (#2006)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-10-27 22:18:10 +08:00

3057db15c3 [readme] upd mcore readme (#9352)

423A35C7 synced commits to main at 423A35C7/pytorch3d from mirror 2025-10-27 21:48:10 +08:00

45df20e9e2 clang-format | Format fbsource with clang-format 21.

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-10-26 21:48:11 +08:00

13170577b2 [feat] support megatron-LM training by mcore_adapter (#9237)

129e918106 [data] Fix Qwen3VL plugin (#9297)

Compare 2 commits »

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-10-24 20:48:10 +08:00

9c0d033a15 [model] add qwen3vl 2b & 32b (#9343)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-10-23 03:58:10 +08:00

2a822178de [deps] fix yanked packages (#9333)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-10-21 19:18:11 +08:00

b842457ef4 [ci] revert mac os ci setup (#9316)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-10-19 01:58:10 +08:00

2c6aded5d4 [v1] kernel plugin (#9274)

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-10-18 01:28:11 +08:00

d9d67ba62d [misc] fix import error (#9299)

423A35C7 synced and deleted reference refs/tags/hiyouga/fix_deps at 423A35C7/LLaMA-Factory from mirror 2025-10-18 01:28:11 +08:00

423A35C7 synced commits to main at 423A35C7/LLaMA-Factory from mirror 2025-10-17 17:18:10 +08:00

a442fa90ad [misc] fix import error (#9296)

8c341cbaae [model] support hunyuan-mt model (#9284)

Compare 2 commits »

423A35C7 synced new reference hiyouga/fix_deps to 423A35C7/LLaMA-Factory from mirror 2025-10-17 17:18:10 +08:00