Search results

[Bug]: ImportError: cannot import name 'SamplingParams' from 'vllm' in Docker image build

bugSomething isn't workingSomething isn't working

Status: Open.

#2584 In vllm-project/tpu-inference;

· karan opened on May 11, 2026

[Usage]: Cannot run Gemma-4-26B-A4B-IT on TPU v5e

bugSomething isn't workingSomething isn't working

contribution-welcome

Status: Open.

#2453 In vllm-project/tpu-inference;

· IKACE opened on Apr 30, 2026

[Bug]: gemma4 lm-eval is too low (flexible-extract 0.1683)

bugSomething isn't workingSomething isn't working

Status: Open.

#2332 In vllm-project/tpu-inference;

· lk-chen opened on Apr 20, 2026

[Bug]: gpt-oss's lm-eval "strict-match" drops to 0.0 after {lm-eval, transformers} version upgrade

bugSomething isn't workingSomething isn't working

Status: Open.

#2318 In vllm-project/tpu-inference;

· lk-chen opened on Apr 18, 2026

[Bug]: KeyError during lm_eval (MMLU task)

bugSomething isn't workingSomething isn't working

Status: Open.

#2072 In vllm-project/tpu-inference;

· lk-chen opened on Mar 27, 2026

[Bug]: TPU to CPU model loading performance degradation for GPT-OSS 120B TPU inference run

bugSomething isn't workingSomething isn't working

Status: Open.

#1707 In vllm-project/tpu-inference;

· lepan-google opened on Feb 12, 2026

[Bug][FP8][Jax] jax-path cannot load lm_head properly for model "Qwen/Qwen3-32B-FP8"

bugSomething isn't workingSomething isn't working

Status: Open.

Bug

#1684 In vllm-project/tpu-inference;

· lk-chen opened on Feb 10, 2026

[Bug] Qwen/Qwen3-4B-Instruct-2507-FP8 has different latency/performance on jax-path vs. torchax-path

bugSomething isn't workingSomething isn't working

Status: Open.

#1661 In vllm-project/tpu-inference;

· lk-chen opened on Feb 9, 2026

[Bug]: test_qwen3.py:test_model_loading failure from transformer pylib

bugSomething isn't workingSomething isn't working

Status: Open.

Bug

#1604 In vllm-project/tpu-inference;

· lk-chen opened on Feb 3, 2026

[Bug]:LLGuidance structured decoding breaks under concurrent requests

bugSomething isn't workingSomething isn't working

Status: Open.

#1531 In vllm-project/tpu-inference;

· XMaster96 opened on Jan 26, 2026

[Bug]: Mosaic failed to compile TPU kernel for gpt-oss on v6e

bugSomething isn't workingSomething isn't working

Status: Open.

#1425 In vllm-project/tpu-inference;

· piotrfrankowski opened on Jan 8, 2026

[Bug]: Not able to run FP8 model on v5e

bugSomething isn't workingSomething isn't working

Status: Open.

#1262 In vllm-project/tpu-inference;

· depksingh opened on Dec 8, 2025