Downloads · 30 days
198
27% of all-time downloads
Sawfwair/MiniMax-H3-FL2VA-MLX-8bit
MiniMax-H3-FL2VA-MLX-8bit is a text-to-video model from Sawfwair. Use it when you need video from a text prompt. It is set up for mlx. The card lists the license as other.
This is Mere's high-quality compact MLX runtime artifact for MiniMax-H3 FL2VA. Every weight input came from the pinned official checkpoint MiniMaxAI/MiniMax-H3@ec19cc6daf5d8add9417c18e86b6b58cc6c55027; no third-party…
Downloads · 30 days
198
27% of all-time downloads
All-time downloads
730
Public
Repo size
61 GB
Likes
0
Public
Click a slice to open those files.
.safetensors58.3 GB · 100%
From the Hugging Face model README
This is Mere's high-quality compact MLX runtime artifact for MiniMax-H3 FL2VA.
Every weight input came from the pinned official checkpoint
MiniMaxAI/MiniMax-H3@ec19cc6daf5d8add9417c18e86b6b58cc6c55027; no
third-party converted or quantized checkpoint was used.
Eligible linears in the 20.11B-parameter active denoising core use MLX affine INT8/group-64. The Qwen3-VL conditioner also uses MLX affine INT8/group-64, the video VAE uses FP16, and the audio VAE remains FP32. Schedule-only AdaLN projections, the timestep MLP, and reconstructed RoPE tensors are omitted after generating a source-bound pack of exact 5, 9, 12, 16, 21, and 31-point tables at video/audio shifts 12/3 plus the LightX2V 5- and 9-point 6/3 tables.
The original cache pack was evaluated by mere.run model optimize on MLX
Metal. The 9-point 6/3 table was added by the source-range refresh recorded in
adaln_cache.refresh.json, also on MLX Metal. The refresh reproduced the
tensor closure of every previously published table exactly before the new
table was accepted. The pack remains bound to the official projection-tensor
closure plus the original bit-identical 9- and 21-point real-generation
receipts. The denoising core was reproduced and quantized on MLX CUDA;
CUDA-evaluated AdaLN tables are not accepted because backend BF16 reduction
order is not bit-identical to Apple Silicon.
SOURCE_MANIFEST.json, transformer.conversion.json, MODIFICATIONS.md, and
SHA256SUMS provide the conversion and byte provenance. The artifact is
intended for video-minimax-h3-fl2va-8bit-mlx in mere.run. Q8 is published as
a storage and memory option; no speed claim is made without matched hardware
measurements.
These weights are governed by the MiniMax H3 Community License Agreement in
LICENSE, not the mere.run source license. At the pinned source revision the
license excludes use, distribution, and display in the United States, European
Union, United Kingdom, and Republic of Korea and carries additional notice and
safeguard obligations. Review the complete license before downloading or using
the artifact.