fix: use custom_tokenizer to workaround the trtllm + glm5 tokenizer loading issue#13
Closed
richardhuo-nv wants to merge 4 commits intoNVIDIA:mainfrom
Closed
fix: use custom_tokenizer to workaround the trtllm + glm5 tokenizer loading issue#13richardhuo-nv wants to merge 4 commits intoNVIDIA:mainfrom
richardhuo-nv wants to merge 4 commits intoNVIDIA:mainfrom
Conversation
Auto-detect container type at runtime: if /sgl-workspace exists (SGLang), use original install path unchanged; otherwise use portable /tmp build path with conditional dependency installation for non-SGLang containers.
* Add Kimi-K2.5 vLLM recipes and fix NIXL side channel host - Add kimi-k2.5 1k1k and 8k1k disagg GB200 recipes (from NVIDIA#7) - Fix vLLM NIXL handshake failures: set VLLM_NIXL_SIDE_CHANNEL_HOST to node's routable IP in get_process_environment() instead of leaving it as 0.0.0.0/localhost which caused transfer handshake failures - Update test_vllm_get_process_environment to cover NIXL host env var Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * ci: run checks on PRs targeting sa-submission-q2-2026 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
fix
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #13 +/- ##
=======================================
Coverage ? 60.15%
=======================================
Files ? 48
Lines ? 4081
Branches ? 0
=======================================
Hits ? 2455
Misses ? 1626
Partials ? 0 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
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
3 participants
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.
TRT-LLM is still on Transformers v4, while the GLM-5 model was built with Transformers v5. As a result, the GLM-5 tokenizer cannot be loaded directly with AutoTokenizer in Transformers v4.
Our current workaround is adapted from TensorRT-LLM’s glm_moe_dsa tokenizer implementation:
https://github.com/NVIDIA/TensorRT-LLM/blob/main/tensorrt_llm/tokenizer/glm_moe_dsa/tokenizer.py
This workaround uses the Rust tokenizer library to load tokenizer.json, and then initializes a Transformers v4 AutoTokenizer with appropriately translated settings from tokenizer_config.json.
At the moment, this workaround does not support chat_template, so we need to disable chat templating for now.