dylanyunlong
0 Followers · 0 Following
  • Joined on 2026-08-06
Repositories
10
 Projects Public Activity Starred Repositories
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-13 10:53:26 +08:00
a3c45d3b36 fix(build): match project_7 proven Dockerfile — remove ENV lines + .dockerignore
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-13 10:37:30 +08:00
5a05d4528c fix(stability): prevent OOM crash + disable garbled topk_softmax kernel
5696de5317 update sub 655
Compare 2 commits »
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-13 10:19:28 +08:00
544e255ec0 fix: copy_blocks use vllm_copy_cache
60e0b9da87 Revert "fix(precision): guard all corex .so outputs with nan_to_num + reduce max-model-len"
8acc47129b fix(precision): guard all corex .so outputs with nan_to_num + reduce max-model-len
8abc7cb0d7 diag: verify_so_loading.py — check all 13 prebuilt .so + base image dlopen chain
Compare 4 commits »
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-13 10:01:39 +08:00
275e117228 trigger rebuild
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-13 09:53:29 +08:00
01978e033a yaml: add enforce-eager only
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-13 09:47:01 +08:00
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-13 09:29:41 +08:00
5769737264 fix(build): restore ENV lines — python3/vllm/torch require PATH+PYTHONPATH+LD_LIBRARY_PATH
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-13 09:14:35 +08:00
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-13 09:03:17 +08:00
f457906c29 fix: copy_blocks + max_completion_tokens (no new .so)
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-13 08:57:26 +08:00
cf4c16bb23 fix: copy_blocks use ixf_F.vllm_copy_cache directly
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-13 00:13:40 +08:00
42e078eafe feat: 13th .so topk_softmax + fp32 fix + max_completion_tokens
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-13 00:06:25 +08:00
d1cc7d4e89 fix: yaml params only — enforce_eager + dtype half + comp168 settings
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-13 00:03:16 +08:00
ec23c03ebc fix: restore original Dockerfile comments
dylanyunlong pushed to main at dylanyunlong/project_7 2026-08-12 23:54:11 +08:00
31d22b66d7 feat: topk_softmax .so + fp32 router + enforce_eager + comp168 params
f1a408c75f fix: cap default_max_tokens at 8192 — prevent OOM kill on unlimited generation
dbc1ce9a71 fix: max-num-seqs=2 — t2_n_2测试需要n=2多路生成
Compare 3 commits »
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-12 23:44:32 +08:00
2ac877cee4 fix: remove ENV lines — match project_7 proven Dockerfile
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-12 19:14:44 +08:00
07e8681e2e fix: topk_softmax .so + fp32 router + enforce_eager + comp168 params
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-12 19:01:57 +08:00
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-12 18:35:16 +08:00
651cf440aa update sub 655
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-12 18:28:21 +08:00
dylanyunlong pushed to main at dylanyunlong/project_6 2026-08-12 18:09:27 +08:00
7969874cde fix: cast router_logits to float32 before topk kernel