Any-to-Any
MLX
Safetensors
gemma4
mlx-vlm
rlcd
multimodal
classification
parallel-inference
image-text-to-text
audio
video
4-bit precision
Instructions to use larkooo/gemma-e2b-rlcd with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use larkooo/gemma-e2b-rlcd with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download larkooo/gemma-e2b-rlcd --local-dir gemma-e2b-rlcd
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
|
Download docs/assets/README.md from larkooo/gemma-e2b-rlcd: direct link, hf CLI and curl.
- Browser
- Download file 845 Bytes
-
https://huggingface.co/larkooo/gemma-e2b-rlcd/resolve/main/docs/assets/README.md
- Command line
-
hf download hf://larkooo/gemma-e2b-rlcd/docs/assets/README.md
-
curl -L -o README.md https://huggingface.co/larkooo/gemma-e2b-rlcd/resolve/main/docs/assets/README.md
845 Bytes
| # Demo assets | |
| `demo.mp4` is a 24-second, 1080p, silent replay of the measured support-triage comparison. Both clocks use real elapsed-time scale; each answer appears when its measured median completion time elapses. It is a designed replay rather than a screen recording. Eight representative fields are shown; all 28 outputs are in `demo-metrics.json`. | |
| The source is [`support_28` in the workload report](../../reports/demo-workload-benchmark.json): four measured repetitions per path after warmup, 4.4145 seconds parallel versus 15.6597 seconds normal generation. All 28 outputs match; both paths pass 15/17 annotated checks. The metrics file retains the report hash and complete methodology. | |
| The video uses IBM Plex Sans under the [SIL Open Font License](Plex-OFL.txt). The repository URL and measured results appear in the closing frame. | |