-
Notifications
You must be signed in to change notification settings - Fork 175
[Handoff to @Oseltamivir Claude /loop] [Klaud Cold] Update dsv4-fp4-b300-sglang (+mtp) SGLang image to v0.5.12-cu130 #1455
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Open
functionstackx
wants to merge
1
commit into
main
Choose a base branch
from
update-dsv4-fp4-b300-sglang-v0.5.12
base: main
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
Open
Changes from all commits
Commits
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
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
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
Oops, something went wrong.
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Outdated suggestions cannot be applied.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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
🔴 Bumping
dsv4-fp4-b300-sglang(line 1986) anddsv4-fp4-b300-sglang-mtp(line 2027) from the SHA-pinnedlmsysorg/sglang:deepseek-v4-b300@sha256:...custom image to the genericlmsysorg/sglang:v0.5.12-cu130strips the patched transformers that registersmodel_type: "deepseek_v4", soAutoTokenizer.from_pretrained("deepseek-ai/DeepSeek-V4-Pro")will crash withKeyError: '\''deepseek_v4'\''before the server is even probed. Either hold this change until upstream sglang ships transformers withdeepseek_v4support, or have the recipepip installa patched transformers inside the container before invoking the bench client. The PR author has already acknowledged this in the timeline.Extended reasoning...
What the bug is
Both modified entries (
dsv4-fp4-b300-sglangat line 1986 anddsv4-fp4-b300-sglang-mtpat line 2027) swap out a SHA-pinned custom image for the genericlmsysorg/sglang:v0.5.12-cu130image. The customdeepseek-v4-b300@sha256:...builds bundle a patchedtransformersthat registers a model type fordeepseek_v4(theconfig.jsonofdeepseek-ai/DeepSeek-V4-Prodeclaresmodel_type: "deepseek_v4"). The genericv0.5.12-cu130image bundles the upstreamtransformersrelease, which has nodeepseek_v4entry in its model-type registry.Code path that triggers the failure
image: lmsysorg/sglang:v0.5.12-cu130.AutoTokenizer.from_pretrained("deepseek-ai/DeepSeek-V4-Pro").transformersdownloads the model repo'sconfig.json, readsmodel_type: "deepseek_v4", then attempts to look it up inCONFIG_MAPPING.transformersinv0.5.12-cu130does not havedeepseek_v4registered →KeyError: 'deepseek_v4'is raised before the SGLang server is ever probed.Why existing code doesn't prevent it
The recipe scripts for these two configs only change the image tag; nothing in the recipe pipes in a
pip install transformers ...upgrade to bring in deepseek-v4 support. The sister entrydsv4-fp4-b200-sglangat line 1699 is still pinned tolmsysorg/sglang:deepseek-v4-blackwell@sha256:df18bfc...for exactly this reason, and other DSV4 entries (e.g. trtllm variants at lines 1781/1802/3016/3039) all use specifically-taggedtrtllm-deepseek-v4:feat-deepseek_v4-9aa3715images. Every DSV4 config in this file requires a special image withdeepseek_v4support — the b300 sglang variants are no exception.Author confirmation
The PR author (functionstackx) acknowledged this directly in this PR's timeline on 2026-05-18T07:45:18Z: "the generic
v0.5.12-cu130image bundles a transformers that doesn'''t recognisemodel_type: "deepseek_v4", so the bench client crashes inAutoTokenizer.from_pretrainedwithKeyError: '\''deepseek_v4'\''. ... the generic-image bump is NOT viable until sglang ships transformers withdeepseek_v4support." They closed the PR as not viable, then reopened it with sweep labels intentionally disabled to avoid auto-triggering failing runs while they debug.Step-by-step proof
.github/configs/nvidia-master.yamlat line 1986 — image is nowlmsysorg/sglang:v0.5.12-cu130, model isdeepseek-ai/DeepSeek-V4-Pro.v0.5.12-cu130image:docker pull lmsysorg/sglang:v0.5.12-cu130.python -c "from transformers import AutoTokenizer; AutoTokenizer.from_pretrained('\''deepseek-ai/DeepSeek-V4-Pro'\'')".KeyError: '\''deepseek_v4'\''raised fromCONFIG_MAPPING.__getitem__because upstream transformers in this image has no entry fordeepseek_v4.dsv4-fp4-b300-sglang-mtp) — same image, same model, identical failure.Impact
Both⚠️ Note: the deepseek-v4-b300 tag is a custom DSV4 build; the generic v0.5.12-cu130 may or may not retain DSV4-specific features."
dsv4-fp4-b300-sglanganddsv4-fp4-b300-sglang-mtpsweep runs will fail at tokenizer load 100% of the time. No benchmarks will be produced. The PR description itself acknowledges this risk: "How to fix
Either (a) revert the image to the SHA-pinned custom
deepseek-v4-b300builds and wait for upstream sglang to ship a transformers release withdeepseek_v4registered, or (b) keep the generic image bump but have the recipepip installa transformers build containingdeepseek_v4support inside the container before invoking the bench client. Option (a) is the safer choice and matches what is already done for the b200 sister entry at line 1699.