SAVRN Model Hub
Open-Weight Models
An open-weight model is an AI model whose trained weights are published for anyone to download. The weights are what the model learned in training. With a copy of them you can run the model on hardware you control and train it further on your own data.
Open weights are not the same as open source. Many publishers release the weights without the training data or code, and the license sets what you may do with the model. This library puts each model's full card, architecture, files, license and published evaluations on one page.
Updated 2026-10-23 · How the library is built
12,579 models, sorted by most downloaded.
This repository contains the released online weights for MGFlow's ImageNet-256 and text-to-image experiments. Each checkpoint is a PyTorch dictionary with exactly two top-level entries: - model: the online model state dict used for sampling; - step: the checkpoint's training step. EMA weights, optimizer state, random-number-generator state, and training metadata are not included. Load with torch.load(path, maplocation="cpu", weightsonly=True) and use the model state dict with the corresponding architecture. ImageNet/ contains 12 final checkpoints: one W2 and one KL checkpoint for each of JiT-B/L/H and pMF-B/L/H. - W2-k1+4.pth: MGFlow-W2 with K=1+4. - KL-k1+4+16.pth: MGFlow-KL with K=1+4+16…
Model checkpoints, reference distributions, and evaluation assets for MGFlow, presented in Unifying Distributional Training for One-Step Visual Generation. The release covers ImageNet class-conditional generation at 256×256 with JiT and pMF (B/L/H), and text-to-image generation at 512×512 with FLUX.2 [klein] 4B. Both generate images in one step after MGFlow post-training. The T2I checkpoints use the MGFlow-T2I reference dataset, which contains synthetic COCO and GenEval images paired with their prompts. The dataset link in the metadata applies to this T2I branch. If you use MGFlow or these assets, please cite
Built with Llama. Quantized GGUF build of slseanwu/MIDI-LLMLlama-3.2-1B for local inference in MIDI Forge via llama.cpp. - Do not use truncated GGUFs that drop the MIDI embedding rows. This model is a derivative of Meta Llama 3.2 and is subject to the and Acceptable Use Policy. 1. Creative → Generate with MIDI-LLM… 2. Accept the Llama 3.2 license in the dialog 3. Download model (this file) or the app fetches it automatically from: https://huggingface.co/tazaleth/MIDI-LLMLlama-3.2-1B-GGUF/resolve/main/midi-llm-llama32-1b-q5km.gguf AMT event vocabulary: Thickstun et al., Anticipatory Music Transformer (TMLR 2024).
Behaviorally modified MXFP4 research checkpoint derived from XiaomiMiMo/MiMo-V2.6-Flash-RL. DERISKED identifies the Blackfrost release family. It is not a claim of zero refusals, complete safety, harmlessness, or production readiness. Refusal measurements will be published only after evaluation of the exact uploaded artifact is complete. MIMO-V2.6-DERISKED-MXFP4 is an independent Blackfrost research derivative of XiaomiMiMo/MiMo-V2.6-Flash-RL. The checkpoint retains the upstream MiMo V2.6 architecture, tokenizer, multimodal components, native chat template, and speculative-decoding assets while preserving its mixed-precision MXFP4 deployment layout. The behavioral modification is encoded in…
This is a Stage-B research weight release: three trained input hidden→hidden Rank-4 LRARC slow initializations (rho), together with the original, always-enabled semantic adapter omega. The exact MiniCPM5-1B-SFT backbone must be obtained separately. This is a finite four-program selection policy, not a standalone chat model or a completed PDSA system. 中文：本目录保存 B 阶段选定静态目标臂的三个训练后 rho 和必需的原 omega。三种子整体保留，未挑最好的一次。A/B 已完成本轮验收；权威记忆、预测写入 Q、双状态联合效应和严格 RSI 尚未由这些权重验证。 Junyi Zou and Avrova Donz. MMLA: Memory-Mediated Learning Architecture for Predictive Dual-State Adaptation, arXiv:2606.28876v4, 2026. MMLA is a mixer-agnostic learning architecture. Its identity is not defined by the temporarily reused…
static quants of https://huggingface.co/SvalTek/MN-CharThink-12B-Base For a convenient overview and download list, visit our model page for this model. weighted/imatrix quants are available at https://huggingface.co/mradermacher/MN-CharThink-12B-Base-i1-GGUF If you are unsure how to use GGUF files, refer to one of TheBloke's READMEs for more details, including on how to concatenate multi-part files. (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) Here is a handy graph by ikawrakow comparing some lower-quality quant And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 See…
weighted/imatrix quants of https://huggingface.co/SvalTek/MN-CharThink-12B-Base For a convenient overview and download list, visit our model page for this model. static quants are available at https://huggingface.co/mradermacher/MN-CharThink-12B-Base-GGUF If you are unsure how to use GGUF files, refer to one of TheBloke's READMEs for more details, including on how to concatenate multi-part files. (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) Here is a handy graph by ikawrakow comparing some lower-quality quant And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 See…
Checkpoint for MORAD: Reinforcement Learning with Multi-Objective Rewards for model after 390 updates of online reinforcement learning with the six-component MORAD reward. A single shared policy designs nucleotide sequences for any target 3D backbone, with no per-target optimization. SHA-256 checksums are listed in SHA256SUMS. Structural and pairing metrics use candidate 0 of each target (structures predicted with RhoFold+, pairs with EternaFold); NED is from the thermodynamic evaluation (ViennaRNA). All tables of the paper are in the downloadassets.sh stores this file as artifacts/models/morad/moradupdate390.pt. Any MORAD checkpoint can be evaluated with RIDECHECKPOINT=path/to/policy.pt…
The released model of MORAD: Reinforcement Learning with Multi-Objective diffusion model, post-trained online for 390 updates with the six-objective MORAD reward. One shared policy designs nucleotide sequences for any target 3D backbone, with no per-target optimization. EternaFold, normalized ensemble defect (NED) with ViennaRNA. gRNAde at T = 0.1. Training this model from pretrained RIDE takes about 2 hours on 4 × RTX 3090 (bash scripts/trainmorad.sh).
Unofficial Research Turbo LoRA for MOSS-Image using Mean Flow Distillation MOSS-Image-Base-Turbo-MFD-v0.1 - 50 aux/20 student, 4-6 steps inference ComfyUI version is.comfy.safetensors Thank you HuggingFace, OpenMOSS Team, and An Zhao and Shengyuan Zhang and Zhongjian Sun and Yixiang Zhou and Zejian Li and Ling Yang and Tianrun Chen and Lingyun Sun
Model Collections
Hand-picked starting points, each with the reason it exists.
Collection · 4 entries
Embedding models for retrieval
Sentence and document embedding models used to build retrieval systems. Dimension and sequence length matter more than size here, and both come from the publisher.
Collection · 4 entries
Models that fit on one accelerator
Models whose publisher-reported parameter count puts them within reach of a single accelerator at common precisions. Memory needed depends on precision and serving configuration, so treat the parameter count as the starting point, not the answer.
Collection · 6 entries
Open-weight text models worth knowing
Widely used open-weight language models, chosen because each one is a distinct family rather than a variant of the one above it. Selection, not a ranking.
Collection · 3 entries
Speech and audio models
Recognition and synthesis models, grouped so the two directions are easy to compare.
Open-Weight Models Explained
What is an open-weight model?
An AI model whose trained weights are published for anyone to download, so it can be run, tested and fine-tuned on hardware the user controls.
Is an open-weight model the same as open source?
Not always. Open weights means the trained model can be downloaded. Open source usually also means the training code and data are available and the license allows broad reuse. Many open-weight models release the weights only.
Can I use an open-weight model commercially?
It depends on the license. Apache 2.0 and MIT allow commercial use. Other licenses limit it, for example to non-commercial use or below a set number of users. Every model page here shows its license.
How much memory does an open-weight model need?
About two bytes per parameter at 16-bit precision, so a 7-billion-parameter model needs roughly 14 GB for its weights, plus memory for the context it processes. Each model page lists its parameter count and the size of its files.
Related SAVRN Research
The hub sits beside SAVRN's market data and infrastructure research: what models cost to run, and what it takes to run them.
SAVRN Index
What open models cost to run
The same open-weight model priced by every host that serves it, per million tokens.
Research Hub
Data center trackers and maps
Moratoriums, permits, power, water and capital behind the facilities that run these models.
Method
How the Model Hub is built
Sources, evidence labels, refresh behaviour, and the limits of every comparison here.