Datasets:

juiceb0xc0de
/

gemma-4-e2b-it-cu128-torch291-py311

Size:

Tags:

License:

Dataset card Files Files and versions

xet

Community

The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.

`juiceboxdocks/gemma-4-e2b-it-base` — Docker Image Card

Not a data dataset. This repo is a landing page for the Docker image juiceboxdocks/gemma-4-e2b-it-base. If you're here because you're struggling to get Gemma-4-E2B-it running with the right CUDA/PyTorch stack, just pull the image and skip the environment setup entirely.

Quick start

docker pull juiceboxdocks/gemma-4-e2b-it-base:cu128-torch291-py311

docker run --rm -it --gpus all --ipc=host \
  -v gemma-workspace:/workspace \
  -e HF_TOKEN \
  juiceboxdocks/gemma-4-e2b-it-base:cu128-torch291-py311

That drops you at a bash shell in /workspace with everything already installed. Nothing downloads automatically. The image contains no model weights.

GPaaS

Don't want to deal with Docker? Deploy directly on RunPod:

RunPod template: gemma-4-e2b-it-base cu128-torch291-py311

Image tags

Tag	When to use
`cu128-torch291-py311`	Stable — use this for normal work
`cu128-torch291-py311-<short-sha>`	Pin to an exact build for reproducibility
`latest`	Tracks the most recent push

Docker Hub: juiceboxdocks/gemma-4-e2b-it-base

What's in the image

Base: nvidia/cuda:12.8.1-cudnn-devel-ubuntu22.04

Stack:

Component	Version
Python	3.11
PyTorch	2.9.1+cu128
TorchVision	0.24.1
TorchAudio	2.9.1
Transformers	5.5.4
CUDA	12.8.1

Pinned source trees (editable installs):

Package	Repo	SHA
`gemma-triton-flash-attn`	zzhhjjj/gemma-triton-flash-attn	`c6a0cf8`
`event-aware-SAE-trainer`	JuiceB0xC0de/event-aware-SAE-trainer	`319ba61`

Source trees live at /opt/src/ and are installed as editable (pip install -e).

Dev tools included: tmux, Git, nvcc, cuda-gdb, compute-sanitizer, Nsight Systems (nsys), btop, nvitop, W&B, modern hf CLI, Prometheus client, NVTX, libaio-dev, Ninja, numactl.

Caches all live under /workspace — mount a persistent volume there and Triton kernels, HuggingFace downloads, and pip cache survive restarts.

Explicitly NOT included: model weights, checkpoints, datasets, credentials, automatic downloads, automatic jobs.

Triton cache

TRITON_CACHE_DIR is set to /workspace/.cache/triton. With a persistent mount, compiled kernels survive container restarts so you skip recompilation after the first run.

Warm the cache without weights:

python /opt/image-tools/warm_triton_cache.py

Run it twice on the same GPU + volume and look for new_files: 0 on the second pass — that's a real cache hit. A100 (sm_80) and H100 (sm_90) compile different binaries; both coexist under the same cache root.

GPU smoke test

No weights required:

python /opt/image-tools/gpu_smoke.py

Reports GPU/driver/stack, imports the SAE trainer, registers the Triton attention adapter, and runs forward + backward comparisons against PyTorch references for:

Gemma full causal attention: D=512, 32 query / 4 KV heads
Gemma sliding-window attention: D=256, 32 query / 16 KV heads, window 512

Load Gemma-4-E2B-it (after pulling weights)

Once you're inside the container with HF_TOKEN set:

import torch
from transformers import AutoModelForImageTextToText, AutoProcessor

model = AutoModelForImageTextToText.from_pretrained(
    "google/gemma-4-E2B-it",
    dtype=torch.bfloat16,
    device_map="auto",
)
processor = AutoProcessor.from_pretrained("google/gemma-4-E2B-it")
print(type(model).__name__)

Needs ~4–5 GB VRAM in bf16. A single A100/H100/RTX 3090+ is fine.

Extend

FROM juiceboxdocks/gemma-4-e2b-it-base:cu128-torch291-py311

RUN python -m pip install --no-cache-dir jupyterlab

COPY . /workspace/my-project
WORKDIR /workspace/my-project
CMD ["python", "train.py"]

No VOLUME directive on /workspace means your downstream build-time writes persist into the layer.