--- title: "Pre-built images and their versions" description: "Each TIR pre-built image's versions, with the Python, CUDA, and Ubuntu each one installs, and whether it opens JupyterLab." content_type: "concept" product: "tir-instance" platform: "TIR" url: "/tir/tir-instance/concepts/pre-built-images/" --- # Pre-built images and their versions **Pre-built** images come from E2E with a machine learning framework already installed, and most of them open JupyterLab. The list changes over time. These are the images in Delhi as of Oct 14, 2026. | Image | Version | Python | CUDA | Ubuntu | SSH key required | JupyterLab | Image Tag | | --------------------------- | ------------------- | ---------- | ------ | ---------- | ---------------- | ---------- | --------------------------------------------------------------------------------- | | Transformers | v4.51.3 | 3.10 | 12.4 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch-transformers:24.03-2.3.0-4.51.3` | | Transformers | v4.45.2 | 3.10 | 12.4 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch-transformers:24.03-2.3.0-4.45.2` | | Transformers | v4.39.3 | 3.10 | 12.4 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch-transformers:24.03-2.3.0-4.39.3` | | Transformers | v4.31.0 | 3.10 | 12.1 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch-transformers:23.06-2.1.0-4.31.0` | | PyTorch | v2.8 | 3.12 | 12.9.1 | 24.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch:25.06-2.8` | | PyTorch | v2.3.0 | 3.10 | 12.4 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch:24.03-2.3.0` | | PyTorch | v2.1.0 (cuda-12.2) | 3.10 | 12.2 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch:23.10-2.1.0` | | PyTorch | v2.1.0 (cuda-12.1) | 3.10 | 12.1 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch:23.06-2.1.0` | | PyTorch | v1.14.0 | 3.8 | 12.0 | 20.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch:23.02-1.14.0` | | TensorFlow | v2.15.0 | 3.10 | 12.4 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-tensorflow:24.03-2.15.0` | | TensorFlow | v2.13.0 | 3.10 | 12.2 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-tensorflow:23.10-2.13.0` | | TensorFlow | v2.12.0 | 3.10 | 12.1 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-tensorflow:23.06-2.12.0` | | TensorFlow | v1.15.5 | 3.8 | 12.0 | 20.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-tensorflow:23.02-1.15.5` | | Diffusers | v0.27.2 | 3.10 | 12.4 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch-diffusers:24.03-2.3.0-0.27.2` | | Diffusers | v0.18.2 | 3.10 | 12.1 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch-diffusers:23.06-2.1.0-0.18.2` | | NVIDIA Nemo | 25.11 | 3.12 | 13.0 | 24.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:25.11` | | NVIDIA Nemo | 25.07 | 3.12 | 12.9 | 24.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:25.07` | | NVIDIA Nemo | 25.04 | 3.12 | 12.8 | 24.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:25.04` | | NVIDIA Nemo | 24.12 | 3.10 | 12.6 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:24.12` | | NVIDIA Nemo | 24.03 | 3.10 | 12.3 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:24.03` | | NVIDIA Nemo | 24.01.gemma | 3.10 | 12.3 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:24.01.gemma` | | NVIDIA Nemo | 24.01.starcoder2 | 3.10 | 12.3 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:24.01.starcoder2` | | NVIDIA Nemo | 24.01.speech | 3.10 | 12.3 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:24.01.speech` | | NVIDIA Nemo | 24.01.01.framework | 3.10 | 12.3 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:24.01.01.framework` | | NVIDIA Nemo | 23.08 | 3.10 | 12.2 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-nemo:23.08` | | NVIDIA RAPIDS | v23.06 | 3.10 | 12.1 | 20.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-rapids-cuda:23.06-11.8` | | NVIDIA RAPIDS | v24.02 | 3.10 | 12.0 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-rapids-cuda:24.02-12.0` | | Fast.AI | v2.7.14 | 3.10 | 12.1 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-pytorch-fastai:23.06-2.1.0-2.7.14` | | SciPy | v1.7.0 | 3.8 | 12.1 | 20.04 | No | Yes | `registry.e2enetworks.net/aimle2e/scipy-python-cuda:1.7.0-3.8-12.1` | | Jupyter | py3.10-cuda12.1 | 3.10 | 12.1 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/jupyter-python-cuda:3.10-12.1` | | Jupyter | py3.8-cuda12.1 | 3.8 | 12.1 | 20.04 | No | Yes | `registry.e2enetworks.net/aimle2e/jupyter-python-cuda:3.8-12.1` | | Python 3.10 | CUDA - 12.4 | 3.10 | 12.4 | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu-cuda:22.04-12.4` | | Python 3.10 | CUDA - 12.5 | 3.10 | 12.5 | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu-cuda:22.04-12.5` | | Python 3.10 | CUDA - 12.6 | 3.10 | 12.6 | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu-cuda:22.04-12.6` | | Python 3.10 | CUDA - 13.1 | 3.10 | 13.1 | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu-cuda:22.04-13.1` | | Python 3.10 | CUDA - 13.0 | 3.10 | 13.0 | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu-cuda:22.04-13.0` | | Python 3.10 | CUDA - 12.9 | 3.10 | 12.9 | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu-cuda:22.04-12.9` | | Python 3.10 | CUDA - 13.1 | 3.10 | 13.1 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-cuda-jupyter:22.04-13.1` | | Python 3.10 | CUDA - 13.0 | 3.10 | 13.0 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-cuda-jupyter:22.04-13.0` | | Python 3.10 | CUDA - 12.9 | 3.10 | 12.9 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-cuda-jupyter:22.04-12.9` | | Python 3.10 | CUDA - 12.8 | 3.10 | 12.8 | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu-cuda:22.04-12.8` | | Python 3.10 | CUDA - 12.8 | 3.10 | 12.8 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-cuda-jupyter:22.04-12.8` | | Python 3.10 | CUDA - 12.6 | 3.10 | 12.6 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-cuda-jupyter:22.04-12.6` | | Python 3.10 | CUDA - 12.5 | 3.10 | 12.5 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-cuda-jupyter:22.04-12.5` | | Python 3.10 | CUDA - 12.4 | 3.10 | 12.4 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-cuda-jupyter:22.04-12.4` | | Python 3.10 | CUDA - 12.2 | 3.10 | 12.2 | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu-cuda:22.04-12.2` | | Python 3.10 | CUDA - 12.3 | 3.10 | 12.3 | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu-cuda:22.04-12.3` | | Python 3.10 | CUDA - 12.2 | 3.10 | 12.2 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-cuda-jupyter:22.04-12.2` | | Python 3.10 | CUDA - 12.3 | 3.10 | 12.3 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-cuda-jupyter:22.04-12.3` | | Ubuntu | 24.04-cuda13.0 | Not stated | 13.0 | 24.04 | Yes | No | `golden-ubuntu2404-cuda13` | | Ubuntu | 22.04 | Not stated | None | 22.04 | Yes | No | `golden-ubuntu2204` | | Ubuntu | 24.04 | Not stated | None | 24.04 | Yes | No | `golden-ubuntu2404` | | Ubuntu | 22.04 | 3.10 | None | 22.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu:22.04` | | Ubuntu | 20.04 | 3.8 | None | 20.04 | Yes | No | `registry.e2enetworks.net/aimle2e/ubuntu:20.04` | | Ubuntu | 22.04 | 3.10 | None | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-jupyter:22.04` | | Ubuntu | 20.04 | 3.8 | None | 20.04 | No | Yes | `registry.e2enetworks.net/aimle2e/ubuntu-jupyter:20.04` | | TensorRT-LLM Engine Builder | v0.12.0 | Not stated | None | Not stated | No | Yes | registry.e2enetworks.net/aimle2e/triton_trt_llm_jupyter:0.12.0 | | TensorRT-LLM Engine Builder | v0.11.0 | Not stated | None | Not stated | No | Yes | registry.e2enetworks.net/aimle2e/triton_trt_llm_jupyter:0.11.0 | | TensorRT-LLM Engine Builder | v0.10.0 | Not stated | None | Not stated | No | Yes | registry.e2enetworks.net/aimle2e/triton_trt_llm_jupyter:0.10.0 | | TensorRT-LLM Engine Builder | v0.9.0 | Not stated | None | Not stated | No | Yes | registry.e2enetworks.net/aimle2e/triton_trt_llm_jupyter:0.9.0 | | TensorRT-LLM Engine Builder | v0.7.2 | Not stated | None | Not stated | No | Yes | registry.e2enetworks.net/aimle2e/triton_trt_llm_jupyter:0.7.2 | | CLIP | 2.2.2-12.1 | 3.10 | None | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/pytorch-cuda-clip:2.2.2-12.1` | | Detectron | v0.6 | 3.10 | None | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/pytorch-detectron:2.2.2-0.6` | | vLLM | v0.15.1 | 3.12.12 | 12.9 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/vllm-jupyter:v0.15.1` | | vLLM | v0.14.0 | 3.12.12 | 12.9 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/vllm-jupyter:v0.14.0` | | vLLM | v0.13.0 | 3.12.12 | 12.9 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/vllm-jupyter:v0.13.0` | | vLLM | v0.12.0 | 3.12.12 | 12.9 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/vllm-jupyter:v0.12.0` | | vLLM | v0.11.2 | 3.12.12 | 12.9 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/vllm-jupyter:v0.11.2` | | vLLM | v0.10.2 | 3.12.11 | 12.8 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/vllm-jupyter:v0.10.2` | | vLLM | 0.9.0-pre-release | 3.12.10 | 12.8 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/vllm-jupyter:v0.9.0-pre-release` | | vLLM | v0.5.4 | 3.10 | 12.4 | 20.04 | No | Yes | `registry.e2enetworks.net/aimle2e/vllm-jupyter:0.5.4` | | vLLM | v0.5.0 | 3.10 | 12.4 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/vllm-jupyter:0.5.0` | | Julia | v1.9.3 | 3.11 | None | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/julia-jupyter:1.9.3` | | ComfyUI | 0.11.1 | Not stated | None | 22.04 | No | No | `registry.e2enetworks.net/aimle2e/ubuntu-comfyui:22.04-0.11.1` | | ComfyUI | 0.3.5 | Not stated | None | 22.04 | No | No | `registry.e2enetworks.net/aimle2e/ubuntu-comfyui:22.04-0.3.35` | | FramePack | 25.04.18 | Not stated | None | 22.04 | No | No | `registry.e2enetworks.net/aimle2e/ubuntu-framepack:22.04-apr18` | | NVIDIA Riva-Speech | 2.17.0-servicemaker | 3.10 | 12.5 | 24.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-riva-speech:2.17.0-servicemaker` | | NVIDIA Riva-Speech | 2.19.0 | 3.12 | 12.8 | 24.04 | No | Yes | `registry.e2enetworks.net/aimle2e/nvidia-riva-speech:2.19.0` | | Triton | 2.59.0 | 3.12 | 12.9 | 24.04 | No | Yes | `registry.e2enetworks.net/aimle2e/triton-server:25.06-2.59.0` | | Debian | 12 | Not stated | None | Not stated | Yes | No | `golden-debian12` | | SGLang | 0.4.10-cu126 | 3.10.12 | 12.6 | 22.04 | No | Yes | `registry.e2enetworks.net/aimle2e/sglang-jupyter:v0.4.10cu126-4.6.3` | | Alma Linux | 10 | Not stated | None | Not stated | Yes | No | `golden-alma10` | | Alma Linux | 9 | Not stated | None | Not stated | Yes | No | `golden-alma9` | | Rocky Linux | 10 | Not stated | None | Not stated | Yes | No | `golden-rocky10` | | Rocky Linux | 9 | Not stated | None | Not stated | Yes | No | `golden-rocky9` | You can add SSH keys to an instance from any image. The images marked **SSH key required** need one to launch, because you reach them only over SSH. "None" means the image's description names no CUDA version, as on a CPU-only or an operating-system image. "Not stated" means the description leaves that value out. ## What a pre-built image includes What follows holds for four pre-built images: TensorFlow v2.15.0, Transformers v4.51.3, Jupyter py3.10-cuda12.1, and PyTorch v2.8. On any other image, check these points on the instance itself. On all four, JupyterLab and its terminal run as `root`, `sudo` is passwordless, and you install `conda` or `mamba` yourself if you need them. `rsync` comes with the TensorFlow and Transformers images. On the Jupyter and PyTorch images, install it yourself, as [Copy files to and from an instance](/tir/tir-instance/how-to/copy-files-to-and-from-an-instance/) shows. Of the four, TensorBoard comes with the PyTorch, TensorFlow, and Transformers images; see [Use TensorBoard on an instance](/tir/tir-instance/how-to/use-tensorboard-on-an-instance/). To launch one of these images, see [Create an instance](/tir/tir-instance/how-to/create-an-instance/).