Commit Graph
119 Commits
Author SHA1 Message Date
jialin 4bbae673dc fix: draft model local path do not search 2025-11-03 14:22:52 +08:00
jialin e66ee5d3f1 fix: catalog kv cache 2025-11-03 11:00:42 +08:00
jialin 20872100d1 fix: kv cache default value 2025-11-03 10:38:41 +08:00
jialin 06f33ded10 fix: algorithm options do not display 2025-11-03 10:29:32 +08:00
jialin adb5cd72b9 chore: show quant in catalog form 2025-11-03 10:14:30 +08:00
jialin 303429c52c chore: clear space onblur 2025-11-03 10:14:30 +08:00
jialin 1a92cc8543 feat: search draft models from remote 2025-11-03 10:14:30 +08:00
jialin 4af71b5b52 chore: speculative_config init in updating 2025-11-03 10:14:30 +08:00
jialin 2dab89363d chore: draft model downloading 2025-11-03 10:14:30 +08:00
jialin ab9f9a9679 chore: update reference to baseline 2025-11-03 10:14:30 +08:00
jialin 86c2f8eda1 fix: backend change 2025-11-03 10:14:30 +08:00
jialin 56eba0d4fc fix: default spec by mode 2025-11-03 10:14:30 +08:00
jialin 7936d0063d chore: update models locales 2025-11-03 10:14:30 +08:00
jialin 3613fb3f69 feat: add speculative decoding 2025-11-03 10:14:30 +08:00
jialin 608b2e31c7 fix: schdule type gpu_selctor issue 2025-10-31 17:14:38 +08:00
jialin 2006827968 chore: update extended_kv_cache config 2025-10-31 15:03:23 +08:00
jialin f6a3497b0c fix: do not trigger evaluation replica 0 2025-10-31 13:10:52 +08:00
jialin d7ef64746c fix: custom backend version suffix 2025-10-30 19:08:45 +08:00
jialin b733b3615d fix: remove input box for gpus_per_replica 2025-10-30 16:38:55 +08:00
jialin e9fa3e3aa3 fix: gpus_per_replica appears null 2025-10-30 10:52:48 +08:00
jialin b77f796661 feat: add sglang parameters 2025-10-29 18:56:41 +08:00
jialin c077db1129 fix: replicas trigger compatibility 2025-10-29 17:24:26 +08:00
jialin 646d235c65 fix: adjust gpus_per_replicas ux 2025-10-28 17:21:22 +08:00
jialin a86bed294e fix: gpu selector issues 2025-10-28 11:17:29 +08:00
jialin 697f1f9c25 fix: built-in backend versions endwith -custom 2025-10-27 16:09:09 +08:00
jialin 61065ee0b7 chore: scheduling group 2025-10-27 14:13:47 +08:00
jialin df65d3e2f6 fix: kv cache support in builtin backend 2025-10-27 14:13:47 +08:00
jialin 7e81871118 fix: miss cluster id in evaluation 2025-10-24 17:55:54 +08:00
jialin ba80ade643 fix: my models status 2025-10-24 16:11:12 +08:00
jialin a815e71bfe chore: show existing when add file 2025-10-23 11:52:49 +08:00
jialin cc0580301e fix: do not trigger evaluation after switching model 2025-10-22 14:28:18 +08:00
jialin 818c4fc9a2 style: form tabs header gap 2025-10-22 10:21:27 +08:00
jialin fc7350b7e2 style: scroll to field 2025-10-21 19:59:49 +08:00
jialin 1e4018350f chore: worker selector 2025-10-21 15:59:48 +08:00
jialin ccb6e72057 fix: textarea scroll bar, deactive count message 2025-10-21 13:24:21 +08:00
jialin 3a7a4ccf30 fix: check driver command for adding worker 2025-10-20 19:11:02 +08:00
jialin 9b843151fe feat: add gpus_per_replicas 2025-10-20 14:53:33 +08:00
jialin 005295ac97 style: gguf model ux 2025-10-17 10:15:55 +08:00
jialin 07243b206e feat: enable kv cache 2025-10-16 18:05:30 +08:00
jialin f732f5c390 fix: cluster issues 2025-10-16 15:05:21 +08:00
jialin 1da16a04fe fix: do not set default version 2025-10-16 15:05:21 +08:00
jialin ca9ceda2f7 fix: update backend parameters after changing backend 2025-10-15 13:30:58 +08:00
jialin 4196351e48 fix: backend edit format 2025-10-15 12:18:50 +08:00
jialin f02b972dc0 fix(styles): my models 2025-10-14 20:09:43 +08:00
jialin 8398ac59a1 style: update model modal type 2025-10-14 20:09:43 +08:00
jialin 83403a1a1a feat: add extended_kv_cache 2025-10-14 20:09:43 +08:00
jialin 620c113fae fix(style): cluster state do not display 2025-10-14 13:09:22 +08:00
jialin 903500dcfa chore: infinite scroller 2025-10-13 19:55:08 +08:00
jialin b5fa40047a chore: pluggable locale config 2025-10-13 14:40:26 +08:00
jialin eb509166ee style: backend list styles 2025-10-13 14:40:26 +08:00