Repository navigation
[docker] Upgrade SGLang to 0.5.21 and align the CUDA 13 runtime - #2476
Merged
Merged
Conversation
Preserve legacy CLI, YAML, model imports and weight-update APIs while adapting to SGLang's new configuration and weight-update sessions. Align Docker and Conda with straw 0.1.3, the batch-invariant DeepGEMM 0.2.0 fork, and the TMS host-memory fallback. Retain DeepEP v1 alignment and pin Mooncake to the version validated after KV memory remapping. Refresh the GLM deterministic patches and freeze the v0.5.21 patch stack. Enable CUDA compatibility libraries automatically at container startup. Validated GLM deterministic and zero-difference layerwise checks, targeted streaming/recovery/checkpoint GPU CI, and new/legacy SGLang CPU contracts.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Preserve legacy CLI, YAML, model imports and weight-update APIs while adapting to SGLang's new configuration and weight-update sessions.
Align Docker and Conda with straw 0.1.3, the batch-invariant DeepGEMM 0.2.0 fork, and the TMS host-memory fallback. Retain DeepEP v1 alignment and pin Mooncake to the version validated after KV memory remapping.
Refresh the GLM deterministic patches and freeze the v0.5.21 patch stack. Enable CUDA compatibility libraries automatically at container startup.
Validated GLM deterministic and zero-difference layerwise checks, targeted streaming/recovery/checkpoint GPU CI, and new/legacy SGLang CPU contracts.