VoxCPM2 is a tokenizer-free, diffusion autoregressive Text-to-Speech model — 2B parameters, 30 languages, 48kHz audio output, trained on over 2 million hours of multilingual speech data. - 30-Language Multilingual — No language tag needed; input text in any supported language directly - Voice Design — Generate a novel voice from a natural-language description alone (gender, age, tone, emotion, pace…); no reference audio required - Controllable Cloning — Clone any voice from a short clip, with optional style guidance to steer emotion, pace, and expression while preserving timbre - Ultimate Cloning — Provide reference audio + its transcript for audio-continuation cloning; every vocal nuance…
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.
12,283Models
3,510Datasets
290Papers
7,436Publishers
12,632Sourced relationships
Updated 2026-10-22 · How the library is built
12,283 models, sorted by most downloaded.
An INT8-quantized repack of openbmb/VoxCPM2, prepared so the model can be loaded and run entirely inside a web browser by shi3z/voxcpm2web (a WASM + WebGPU port of Nothing about the model architecture was changed. This repository exists to solve 1. wasm32 has a 4 GB address space. The upstream checkpoint is 4.27 GiB of weights plus a 359 MB audiovae.pth. 2. Quantizing during load costs VRAM you may not have. Doing it ahead of time means the browser never materializes the full-precision tensors at all. There is no separate audiovae file: the AudioVAE weights are already merged in, so a browser needs exactly one large download instead of two. Both rows produce applied=632 missing=0 unused=0…
GGUF conversions of mistralai/Voxtral-Mini-3B-2507 for use with transcribe.cpp. Ported from upstream commit pinned 2026-06-06. Validated against the Transformers reference at transcribe.cpp commit Offline audio-LLM speech-to-text and speech translation. A Whisper-large-v3 bidirectional audio encoder feeds a 4-frame-group projector (375 audio tokens per 30 s chunk) into a Ministral-3B causal LM (30 layers, GQA 32/8, NEOX RoPE, SwiGLU) via audio-token injection. Takes a 16 kHz mono WAV and produces a transcript via greedy decoding; speech translation runs through the mistral-common instruct template. The smaller sibling of Voxtral Small 24B — same encoder, projector, log-mel frontend, and…
Исторический идентификатор ветви и файлов — Vol2. WAL2 использует новую исходную Qwen3.8-FP8; старый WAL RC1 на Qwen3.6 со sparse/WALB2 не является parent этих весов. Основной свежий артефакт этого снимка — E022-Fixed128, потомок E020-B1024. Завершённая диагностика: 69/136 (50,74%), лучше родителя 64/136 по сумме, но AIME25 0/30 против 2/30. E020-B1024 сохранён отдельно как контрольный кандидат. Утверждения о превосходстве E022 по всем задачам нет. Полный GSM8K 84,69% относится только к E017. Слабое AIME25 и регрессии более новых вариантов сохранены в отчёте. Артефакт текстовый. Визуальная часть исходной conditional-generation модели не преобразовывалась, не обучалась и не выпускается.…
This repository contains nine weight-only checkpoints. Each checkpoint has a resolved training config.yaml and a datasetstats.json: The midtraining checkpoint and all seven HumanoidArena checkpoints were initialized directly from the full-pretraining checkpoint. HumanoidArena training uses the official SONIC reference actions at their native 50 Hz rate, with separate normalization statistics for each task. Training code and instructions are available in the WB-WAM repository.
This repository contains seven HumanoidArena post-train checkpoints under humanoidarenanative50/: boxshelf, hammer, kickfootball, obstaclenavigation, opendoor, punchmarkers, and sitsofa. Each task directory has a checkpoint, its resolved training config.yaml, and datasetstats.json. Post-training uses official SONIC reference actions at their native 50 Hz rate, with separate normalization statistics for each task. Download the HumanoidArena checkpoints into a local directory: Training and evaluation instructions are in the WB-WAM repository. These results were obtained with Isaac Sim 5.1.0 on an NVIDIA RTX 5090 GPU, using three seeds and 20 episodes per seed for each task. Values are the…
This repository contains the full pre-train checkpoint and the PICO mid-train checkpoint. The mid-train model was initialized directly from the pre-train model. Each checkpoint has its resolved training config.yaml and datasetstats.json. Training code and instructions are in the WB-WAM repository.
Gemma 4 26B-A4B, post-trained with GRPO against a reward model learned from 1.2 million double-blind votes cast by HiWaifu users inside their own role-play conversations. Put back into the same arena, blind, it met GLM-5.1 in 1,430 battles and won 49.6% of the decided votes; against a 13-model field including Gemini, DeepSeek-v4 and Qwen's character models it won 54.7%. Most open role-play models are tuned on preferences that come from an LLM judge, from a handful of annotators, or from synthetic pairs. We had something rarer: a live arena where, inside ordinary chats on our platform, a user is occasionally shown two candidate replies and asked which one they want to continue with. Those…
weighted/imatrix quants of https://huggingface.co/hiwaifu-research/WaifuGemma4-26b-a4b-v1 For a convenient overview and download list, visit our model page for this model. static quants are available at https://huggingface.co/mradermacher/WaifuGemma4-26b-a4b-v1-GGUF This is a vision model - mmproj files (if any) will be in the static repository. 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…
Walker is a decoder-only language model that adds trajectory modeling to the Transformer. Attention asks "what should I look at?" by comparing queries with keys: \text{Attention}(Q,K,V)=\text{softmax}\!\left(\frac{QK^\top}{\sqrt{d}}\right)V Walker keeps attention and asks a second question: "where is my representation going?" Treat the hidden states x₁, x₂, x₃, … as a path. Walker measures how that path moves: This is a discrete derivative of the hidden state. Attention averages over content. The difference signal captures direction and change, which averaging does not. Attention can look around the path. Walker also walks along it. Figure 2. The hidden states form a path. Walker reads its…
In this repository, we present Wan2.1, a comprehensive and open suite of video foundation models that pushes the boundaries of video generation. Wan2.1 offers these key features: This repository features our T2V-14B model, which establishes a new SOTA performance benchmark among both open-source and closed-source models. It demonstrates exceptional capabilities in generating high-quality visuals with significant motion dynamics. It is also the only video model capable of producing both Chinese and English text and supports video generation at both 480P and 720P resolutions. Your browser does not support the video tag. - Wan2.1 Text-to-Video - [x] Multi-GPU Inference code of the 14B and 1.3B…
This model is a StarVLA StarVLA-WM-Wan-PI (WanPI) policy jointly trained on 12 SIMPLE Unitree G1 humanoid tasks. It uses one shared checkpoint across all tasks; there is no task-specific fine-tuning or task-specific checkpoint selection. The training job was configured through 120,000 steps. Results in this card refer specifically to the 100,000-step checkpoint, not finalmodel or the 120,000-step checkpoint. The image is encoded by the Wan VAE and the instruction by UMT5. Hidden features from all 30 Wan transformer blocks are projected to 1,024 dimensions and paired layer by layer with the Action DiT. Even-indexed Action-DiT blocks cross-attend to the matching Wan features; odd-indexed…
Wan 2.2 T2V model merged with LoRA UnifiedReward Flex. BF16 quantization with excluded layers.
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.