Text-to-Image
MLX
MiniMax H3
Safetensors
English
apple-silicon
macos
image-to-image
image-editing
image-generation
Instructions to use devin-lai/MiniMax-H3-Image-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use devin-lai/MiniMax-H3-Image-mlx with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download devin-lai/MiniMax-H3-Image-mlx --local-dir MiniMax-H3-Image-mlx
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Download benchmarks/data/bf16/comparison.json from devin-lai/MiniMax-H3-Image-mlx: direct link, hf CLI and curl.
- Browser
- Download file 1.82 kB
-
https://huggingface.co/devin-lai/MiniMax-H3-Image-mlx/resolve/refs%2Fpr%2F1/benchmarks/data/bf16/comparison.json
- Command line
-
hf download hf://devin-lai/MiniMax-H3-Image-mlx@refs/pr/1/benchmarks/data/bf16/comparison.json
-
curl -L -o comparison.json https://huggingface.co/devin-lai/MiniMax-H3-Image-mlx/resolve/refs%2Fpr%2F1/benchmarks/data/bf16/comparison.json
1.82 kB
| { | |
| "schema_version": 1, | |
| "baseline": "r3-deployment-control", | |
| "candidate": "r3-deployment-qk-fc2-native-norm", | |
| "decision": "measured_improvement", | |
| "quality_pass": true, | |
| "quality_checks": { | |
| "max_abs_error": true, | |
| "nonfinite_values": true | |
| }, | |
| "feasible": true, | |
| "objective": "latency", | |
| "latency_reduction_pct": 5.068505095942822, | |
| "speedup": 1.053391185939559, | |
| "peak_memory_reduction_pct": 0.9451447766078203, | |
| "latency_range_screen_pass": true, | |
| "baseline_summary": { | |
| "cases": { | |
| "product": { | |
| "median_seconds": 66.80325495894067, | |
| "min_seconds": 64.55974454106763, | |
| "max_seconds": 67.02427170914598, | |
| "repeats": 3, | |
| "max_peak_memory_bytes": 42924694280 | |
| }, | |
| "portrait": { | |
| "median_seconds": 64.86302212486044, | |
| "min_seconds": 64.2339057920035, | |
| "max_seconds": 65.15299695800059, | |
| "repeats": 3, | |
| "max_peak_memory_bytes": 43080719280 | |
| } | |
| }, | |
| "geomean_seconds": 65.82599034131171, | |
| "max_peak_memory_bytes": 43080719280 | |
| }, | |
| "candidate_summary": { | |
| "cases": { | |
| "product": { | |
| "median_seconds": 63.90098704188131, | |
| "min_seconds": 61.27363566705026, | |
| "max_seconds": 64.23788916692138, | |
| "repeats": 3, | |
| "max_peak_memory_bytes": 42598718264 | |
| }, | |
| "portrait": { | |
| "median_seconds": 61.10937987500802, | |
| "min_seconds": 60.65490966709331, | |
| "max_seconds": 62.00676687504165, | |
| "repeats": 3, | |
| "max_peak_memory_bytes": 42673544112 | |
| } | |
| }, | |
| "geomean_seconds": 62.48959666640749, | |
| "max_peak_memory_bytes": 42673544112 | |
| }, | |
| "scope": "Equal-weight geometric mean of declared timing-case medians; maximum observed memory. Range screen is descriptive, not statistical confidence; no automatic deployment." | |
| } | |