owen/worker-vllm
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
519 Commits 1 Branch 69 Tags
3d4af5df9bb91389374ce94ef56eab842a92db11
Commit Graph
49 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
velaraptor-runpod 3d4af5df9b chore: update readme vllm version 2026-04-30 20:25:35 -05:00
velaraptor-runpod 4f8a16df5d fix: update transformers to >=5 2026-04-30 19:41:09 -05:00
velaraptor-runpodandClaude Sonnet 4.6 fa42ecd79a fix: resolve lowercase HF cache paths when MODEL_NAME uses original casing 
Fixes FDE-174. Some model stores (e.g. RunPod pre-cached network volumes)
normalize repo IDs to lowercase. HuggingFace Hub caches using the original
casing, so MODEL_NAME=Qwen/Qwen2.5-Coder-32B-Instruct-AWQ would miss a
cache stored as models--qwen--qwen2.5-coder-32b-instruct-awq/ and attempt
a redundant download that fails on limited container storage.

If the exact-case HF cache directory is absent but a lowercase variant
exists, the latest snapshot path is returned directly so vLLM loads from
disk. Absolute paths and models with no lowercase cache are unchanged.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-30 17:54:14 -05:00
velaraptor-runpodandClaude Sonnet 4.6 178c72238e fix: surface LORA_MODULES parse failures instead of silently loading zero adapters 
Fixes FDE-194. Previously a malformed LORA_MODULES value was swallowed at
info level and the engine would start with no LoRA adapters, causing 500s
on any request using an adapter model name (e.g. npc-sim-*).

Changes:
- Log at error level when LORA_MODULES cannot be parsed as JSON
- Log at error level when individual adapter dicts fail LoRAModulePath validation
- Log a final error when all adapters fail to load so the cause is obvious
- Accept a single adapter dict (not just an array) for convenience
- Return early when LORA_MODULES is unset to skip unnecessary parsing

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-30 17:49:09 -05:00
velaraptor-runpodandClaude Sonnet 4.6 e6950bdebd feat: upgrade vLLM to 0.19.1 
- Bump vllm[flashinfer] to 0.19.1 in Dockerfile
- Add OpenAIServingRender (new required dependency in 0.19.x serving layer)
- Pass openai_serving_render to all four serving class constructors
- Remove log_error_stack param (removed upstream in 0.19.x)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-30 17:41:11 -05:00
velaraptor-runpod a774cefe85 fix: clean workspace file 2026-04-30 17:29:00 -05:00
velaraptor-runpod 9de17d49b7 feat: update vllm to 0.18.1 2026-04-30 17:25:23 -05:00
velaraptor-runpod 296556a6f7 feat: update vllm to 0.17.1 2026-04-30 17:21:03 -05:00
chrisvelaandGitHub a1544ea70d Merge pull request #277 from runpod-workers/feat/lmcache 
feat: uv installer, LMCache support, add /v1/responses and /v1/messages endpoints
 2026-04-28 19:44:00 -03:00
velaraptor-runpod 1ed25eea20 update readme with responses and messages routes 2026-04-16 15:30:30 -05:00
velaraptor-runpod dc4ad7ddeb fix: kv_transfer_config is dataclass not json, fix for env var 2026-04-09 14:33:11 -05:00
velaraptor-runpod 9035b0e07f fix: lmcache version 2026-04-09 14:32:19 -05:00
velaraptor-runpod c979f0020f update readmes on TRANSFORMERS_VERSION 2026-04-06 16:54:08 -05:00
velaraptor-runpod 6fbd480a26 fix: allow for transformers_version 2026-04-06 16:49:32 -05:00
velaraptor-runpod 3ef1fb8e7b requested changes 2026-04-03 15:53:40 -05:00
velaraptor-runpod 30e8514d63 Runpod not RunPod 2026-03-17 20:59:52 -05:00
velaraptor-runpod d9808815ee feat: update requirements.txt 2026-03-17 19:56:39 -05:00
velaraptor-runpod d8ed3b5353 feat: update 0.16.0, add lmcache 2026-03-17 19:54:11 -05:00
velaraptor-runpod 4c4e039565 feat: add messages route for anthropic/claude 2026-03-17 19:51:50 -05:00
chrisvelaandGitHub 9d1686960d Merge pull request #273 from runpod-workers/bug/hf-overides-rope-scaling 
bug: fix rope scaling to be forward compatible from hf_overrides
 2026-03-10 11:21:44 -05:00
velaraptor-runpod 45d1eeee47 bug: fix rope scaling to be forward compatible from hf_overrides 2026-03-06 15:34:11 -06:00
chrisvelaandGitHub 17efb0e7d0 Merge pull request #272 from runpod-workers/feat/vllm-0.16.0
Release / release (push) Waiting to run 
feat: Update to 0.16.0
 2026-03-05 13:06:45 -06:00
velaraptor-runpod 2b5f07df63 feat: Update to 0.16.0, remove NUM_GPU_BLOCKS_OVERRIDE in hub default since 0 will break 2026-03-04 16:38:40 -06:00
chrisvelaandGitHub 13fa71878e Merge pull request #269 from runpod-workers/feat/allow-engine-args-env
Release / release (push) Waiting to run 
feat: allow all AsyncEngineArgs as env vars
 2026-02-27 15:34:23 -06:00
velaraptor-runpod 8a9365bed4 remove DEFAULT_ARGS that are none, fix MAX_CONTEXT_LEN_TO_CAPTURE 2026-02-27 14:04:15 -06:00
velaraptor-runpod cd485a1af1 update readme 2026-02-25 22:57:17 -06:00
velaraptor-runpod b9043639e9 requested changes/refactor 2026-02-25 16:07:38 -06:00
chrisvelaandGitHub 407dbd7773 Merge pull request #270 from runpod-workers/feat/update-vllm-v0.15.1 
feat: update vllm to 0.15.1
 2026-02-25 15:35:58 -06:00
velaraptor-runpod f103c142c1 feat: update vllm to 0.15.1 2026-02-24 17:44:49 -06:00
velaraptor-runpod efb093e198 add as VLLM_RUNPOD prefix and update readme 2026-02-24 17:37:57 -06:00
velaraptor-runpod 42443f735e feat: allow engine args through VLLM_ and checks the engine args 2026-02-24 16:05:18 -06:00
chrisvelaandGitHub b7c6d4f9a2 feat: update dockerfile to 12.9.1 (#267)
Release / release (push) Waiting to run 
* feat: update dockerfile to 12.9.1

* update readme on VLLM_NIGHTLY build arg
 2026-02-19 10:13:14 +01:00
chrisvelaandGitHub d69cc021e8 Merge pull request #268 from runpod-workers/fix/spec-config-0-to-none
Release / release (push) Waiting to run 
fix: spec config env vars should be none if zero
 2026-02-18 15:51:51 -06:00
velaraptor-runpod 61faa8f137 fix: spec config env vars should be none if zero 2026-02-18 15:41:19 -06:00
chrisvelaandGitHub 1606cff557 Merge pull request #265 from runpod-workers/fix/zero-max-model-num_batches
Release / release (push) Waiting to run 
fix: check for zero param and set to None
 2026-02-13 15:26:06 -06:00
velaraptor-runpod e705c9494b fix: check for zero param and set to None 2026-02-13 15:23:54 -06:00
chrisvelaandGitHub b749aa5718 Merge pull request #264 from runpod-workers/fix/max_num_batched_tokens
Release / release (push) Waiting to run 
fix: max num batched tokens
 2026-02-13 12:38:01 -06:00
velaraptor-runpod 4705ba8a7c fix: check max_num_batched_tokenz if max_model_len not set 2026-02-13 03:29:52 -06:00
velaraptor-runpod 767c66c301 make minimal changes 2026-02-13 03:23:44 -06:00
velaraptor-runpod fefdbe21a9 update changes 2026-02-13 03:16:43 -06:00
velaraptor-runpod ee961ad28d Update hub.json 2026-02-13 03:08:19 -06:00
velaraptor-runpod 2e8c251447 Merge branch 'main' into feat/update-vllm-v0.15.0 2026-02-13 03:01:05 -06:00
velaraptor-runpod c3cf43b228 Update hub.json 2026-02-13 00:22:16 -06:00
velaraptor-runpod 7ec10b98cd Update utils.py 2026-02-12 15:28:31 -06:00
velaraptor-runpod 340bc0b3c6 fix: served model name 2026-02-10 21:42:58 -06:00
velaraptor-runpod e1e9ef74ad add changes from pr 2026-02-06 18:10:09 -06:00
velaraptor-runpod 461f89cea6 add torch-c-dlpack-ext requirement 2026-02-06 17:03:39 -06:00
velaraptor-runpod 8eb55b90c1 add changes for v0.15.0 2026-02-05 17:24:16 -06:00
chrisvelaandGitHub 3851d53f93 add ENABLE_EXPERT_PARALLEL engine arg for MoE models (#239)
Release / release (push) Waiting to run 
* enable expert parallel arg for moe models

* add ENABLE_EXPERT_PARALLEL to hub config
 2025-11-17 19:25:19 +01:00