Skip to content

Issues: triton-inference-server/server

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.

Sign up for GitHub

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

640 Open 3,233 Closed
640 Open 3,233 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

How can I construct a pb_utils.Tensor without using numpy?
#8022 opened Feb 19, 2025 by fighterhit
Streaming support on Infer endpoint when DECOUPLED mode is true
#8021 opened Feb 19, 2025 by adityarap
1
Inconsistent HF token requirements for cached gated models: Triton vs vLLM deployments
#8020 opened Feb 19, 2025 by haka-qylis
"output tensor shape does not match size of output" when using python backend and providing a custom environment
#8019 opened Feb 19, 2025 by Isuxiz
1
why triton server used so many thread in same triton proc?
#8017 opened Feb 18, 2025 by soulseen
Performance Discrepancy Between NVIDIA Triton and Direct Faster-Whisper Inference
#8016 opened Feb 18, 2025 by YuBeomGon
Python Backend on Windows
#8012 opened Feb 13, 2025 by mhbassel
Unable to load model from S3 bucket
#8008 opened Feb 12, 2025 by jmlaubach
Got run time error 0 active drivers ([]). There should only be one. when using PipelineModule through ray and deepspeed
#8007 opened Feb 12, 2025 by consciousgaze
1
Python Backend support implicit state management for Sequence Inference
#8006 opened Feb 12, 2025 by zhuichao001
k8s-onprem Chart doesn't work with OpenShift's default security posture
#8004 opened Feb 11, 2025 by jharmison-redhat
1
1
ONNX Model IR Version 10 Support
#8001 opened Feb 11, 2025 by RohanAdwankar
Can't build r25.01 (r24.12 builds okay) on ubuntu-22.04 (unclear build errors). Also can't build r24.12 on ubuntu-24.04 (C++ errors)
#7997 opened Feb 8, 2025 by vadimkantorov
2
The system looks the same, but errors occur on some machines, but the reason is unknown
#7996 opened Feb 8, 2025 by coder-2014
[BUG] [GenAI-Perf] openai-fronted server with --endpoint-type completions openai OpenAI related
#7995 opened Feb 7, 2025 by jihyeonRyu
Batching module: backends Issues related to the backends python Python related, whether backend, in-process API, client, etc question Further information is requested
#7994 opened Feb 7, 2025 by riyajatar37003
build.py setting docker build args for secrets even when build-secret flag is not present build Issues pertaining to builds
#7992 opened Feb 6, 2025 by BenjaminBraunDev
@mc-nv
2
libtriton_fil.so missing on Arm64 containers 24.12 and 25.01 module: backends Issues related to the backends module: platforms Issues related to platforms, hardware, and support matrix
#7991 opened Feb 5, 2025 by dagardner-nv 25.02
@mc-nv
1
Performance issue - High queue times in perf_analyzer performance A possible performance tune-up question Further information is requested
#7986 opened Feb 4, 2025 by asaff1
4
Something like "model instance index" inside python backend enhancement New feature or request module: backends Issues related to the backends python Python related, whether backend, in-process API, client, etc
#7984 opened Feb 3, 2025 by vadimkantorov
1
Expected model dimensions when expected shape is not suitable to batch
#7981 opened Jan 31, 2025 by codeofdutyAI
[Question] triton-client numpy 2 support
#7979 opened Jan 30, 2025 by john-pixforce
Pytorch backend: Model is run in no_grad mode even with INFERENCE_MODE=false
#7974 opened Jan 28, 2025 by hakanardo
Method 'forward' is not defined error !
#7968 opened Jan 26, 2025 by MHmi1
1
vLLM backend Hugging Face feature branch model loading enhancement New feature or request
#7963 opened Jan 23, 2025 by knitzschke
1
ProTip! Adding no:label will show everything without a label.