No module named 'ray' in docker container build in 4090

**Describe the bug**
I build docker container in RTX 4090 machine. Run uv run python examples/run_grpo_math.py command(Just follow the instruction in README). I got this error keep pop out and no stop, and the process just block:
```
(_env_builder pid=2086) INFO:nemo_rl.utils.venvs:Using existing venv at /opt/ray_venvs/nemo_rl.models.generation.vllm.vllm_worker.VllmGenerationWorker
(raylet) Traceback (most recent call last):
(raylet) File "/opt/nemo_rl_venv/lib/python3.12/site-packages/ray/_private/workers/default_worker.py", line 8, in
(raylet) import ray
(raylet) ModuleNotFoundError: No module named 'ray'
(raylet) [2025-10-16 08:25:48,459 E 786 786] (raylet) worker_pool.cc:586: Some workers of the worker process(2958) have not registered within the timeout. The process is dead, probably it crashed during start.
(raylet) Traceback (most recent call last):
(raylet) File "/opt/nemo_rl_venv/lib/python3.12/site-packages/ray/_private/workers/default_worker.py", line 8, in
(raylet) import ray
(raylet) ModuleNotFoundError: No module named 'ray'
```
**Steps/Code to reproduce bug**

Running Machine: Linux cpu2 5.15.0-139-generic #149~20.04.1-Ubuntu SMP x86_64 GNU/Linux
CUDA Driver: 12.9
NVIDIA GeForce RTX 4090
```
docker buildx build -t nemo-rl:latest -f Dockerfile .
docker run -it --name nemo-rl --gpus all -d nemo-rl
uv run python examples/run_grpo_math.py
```

**Expected behavior**

Expect run training effectively.

**Additional context**

Add any other context about the problem here.