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 2026-10-18 · How the library is built
11,122 models, sorted by most downloaded.
UniMate (SIGGRAPH Asia 2026) generates motion from text for skeletons of any topology: animals, humanoids and rigged objects. A flow-matching transformer, conditioned on a flan-t5-base encoding of the prompt and on the skeleton's T-pose, hierarchy and joint names, turns noise into 60 frames (2 s at 30 fps) of motion. This repository holds the unimateuniml3df60v3 checkpoint converted to MLX for Apple silicon, for use with mlx-vlm. mlx-community/unimate-uniml3d-f60-v3-cross-attn-bf16 holds the full attention x cross-attention variant (unimateuniml3df60v3crossattn), converted the same way. This is an unofficial conversion, not made or endorsed by the UniMate - Extracted the EMA weights…
UniMate (SIGGRAPH Asia 2026) generates motion from text for skeletons of any topology: animals, humanoids and rigged objects. A flow-matching transformer, conditioned on a flan-t5-base encoding of the prompt and on the skeleton's T-pose, hierarchy and joint names, turns noise into 60 frames (2 s at 30 fps) of motion. This repository holds the unimateuniml3df60v3crossattn checkpoint converted to MLX for Apple silicon, for use with mlx-vlm. mlx-community/unimate-uniml3d-f60-v3-bf16 holds the recommended graph attention x adaLN model (unimateuniml3df60v3), converted the same way. This is an unofficial conversion, not made or endorsed by the UniMate - Extracted the EMA weights…
This model is a fine-tuned version of google/embeddinggemma-300m on an unknown dataset. It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 2e-05 - trainbatchsize: 16 - evalbatchsize: 16 - distributedtype: multi-GPU - numdevices: 2 - totaltrainbatchsize: 32 - totalevalbatchsize: 32 - lrschedulertype: linear - trainingsteps: 20000 - Transformers 5.17.0 - Pytorch 2.10.0+cu128 - Datasets 5.0.0 - Tokenizers 0.23.2
Three fine-tunes of Cosmos3-Edge (3.37 B parameters, 1.42 B trainable; the vision–language tower stays frozen) on LIBERO-10, differing only in which tokens are given to the model as clean conditions during training. Every intermediate checkpoint is included: 1,000 · 2,000 · … · 9,000 iterations, so the whole training trajectory is available, not just the endpoint. A 17-frame clip at 20 fps. Frame 0 is the observation (two 256×256 cameras side by side) and frames 1–16 are the future. The action lane holds 16 tokens, one per future frame, each a 10-D end-effector delta: [Δpos(3), Δrotation as 6-D(6), gripper(1)]. A per-sample prefix length k decides how many action tokens are handed over…
An unmodified copy of Gemma3-1B-ITmulti-prefill-seqq4ekv4096.litertlm from litert-community/Gemma3-1B-IT, used by the Ur-Cars-Voice app's on-device chat. Hosted here so the app can download it directly. SHA-256: 1325ae366d31950f137c9c357b9fa89448b176d76998180c08ceaca78bba98be Gemma is provided under and subject to the Gemma Terms of Use found at Use of this model is subject to the By downloading or using this model you agree to those terms and restrictions.
uratori-ja-2b is a Japanese decision model. Given a text state (for example a source document and a claim) and one or more typed questions, it returns a probability distribution over the allowed answers in a single forward pass, without generating text. Three question types are supported: noul (yes/no), choice (pick one of 2 to 8 labelled options) and score (an ordinal scale with 2 to 10 levels). It is trained for three jobs: checking whether a claim or an answer is supported by supplied evidence (grounding), comparing two documents (contradictions, meaning-changing edits, added claims) and judging retrieval results and answers in a RAG pipeline (relevance, sufficiency, faithfulness). The…
uratori-ja-310m is a Japanese decision model. Given a text state (for example a source document and a claim) and one or more typed questions, it returns a probability distribution over the allowed answers in a single forward pass, without generating text. Three question types are supported: noul (yes/no), choice (pick one of 2 to 8 labelled options) and score (an ordinal scale with 2 to 10 levels). It is trained for three jobs: checking whether a claim or an answer is supported by supplied evidence (grounding), comparing two documents (contradictions, meaning-changing edits, added claims) and judging retrieval results and answers in a RAG pipeline (relevance, sufficiency, faithfulness). The…
uratori-ja-4b is a Japanese decision model. Given a text state (for example a source document and a claim) and one or more typed questions, it returns a probability distribution over the allowed answers in a single forward pass, without generating text. Three question types are supported: noul (yes/no), choice (pick one of 2 to 8 labelled options) and score (an ordinal scale with 2 to 10 levels). It is trained for three jobs: checking whether a claim or an answer is supported by supplied evidence (grounding), comparing two documents (contradictions, meaning-changing edits, added claims) and judging retrieval results and answers in a RAG pipeline (relevance, sufficiency, faithfulness). The…
Byte-level GPTs that dream up login-style usernames (darkphoenix, Cargan66), trained on the ks46/usernames corpus. Every folder is one trained model: its checkpoint (ckpt.pt), the export for the single-threaded x86-64 batch kernel ndgen (model-int7.ndq, model.ndq), the kernel's golden reference NLLs, eval.json, manifest.json, samples.txt and a card with the details. Vocabulary is 256 bytes with byte 0 as the stop token; the held-out split is xxh364(name) % 64 == 0, so bits/char are comparable only between models trained on the same corpus.
Work-in-progress checkpoints of the UTT sensor-agnostic tactile encoder/decoder, published during training cosmosframework/model/generator/tokenizers/utt/. - /index.json: uploaded steps with val/loss, lateststep, beststep (lowest val/loss). - /stepXXXXXXX/model.pt: model weights only (fp32, no optimizer) + UTTConfig + val metrics. - /stepXXXXXXX/metrics.json: validation metrics at that step. Stages: s0a = Sparsh-DINO image stem frozen; s0b = stem fine-tuned (0.1x LR), initialized from s0a. Training data: easyminnn/utt-tactile-prepared v0 pretrain split. Several source datasets carry non-commercial / research-only licenses (see each SOURCE.md there); use these weights accordingly.
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.