readme = """ basemodel: - meta-llama/Meta-Llama-3.1-8B - nvidia/Llama-3.1-Nemotron-70B-Instruct-HF - NousResearch/Hermes-3-Llama-3.1-8B - code - llama-3.1 - merge - slerp - gguf - ollama Merge de Llama 3.1 8B + Nemotron + Hermes 3 usando SLERP + TIES. Optimizado para código, razonamiento y español. - Cuantizado a Q4KM para correr en cualquier laptop from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.frompretrained("fiel1986/Andromeda-8B-Coder", devicemap="auto") tokenizer = AutoTokenizer.frompretrained("fiel1986/Andromeda-8B-Coder") prompt = "Escribe una API en FastAPI con auth JWT" inputs = tokenizer(prompt, returntensors="pt").to("cuda")…
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.
12,870Models
3,655Datasets
291Papers
7,691Publishers
12,980Sourced relationships
Updated 2026-10-24 · How the library is built
12,870 models, sorted by most downloaded.
https://civitai.red/models/2458426/anima?modelVersionId=2836417
https://github.com/suzukimain/AnimeSeg-demo config.json at the repository root lists the checkpoints. Each folder contains one safetensors file. The dev suffix marks development checkpoints.
The three-language base with a 2,048-token context: AnuLM-Base-400M continued for 10,000 steps at block 2,048 with YaRN, on 46M tokens of the same Hindi / English / Python proportions it was originally trained on. Five hours on one RTX 5070 Ti. Full log and the honest reading of what it bought: docs/RESULTS.md §28, with the zero-shot measurement it is compared against Hindi and English Wikipedia are CC BY-SA, C4 is ODC-BY, and the Python slice is codeparrot-clean, de-duplicated GitHub Python with mixed licences. Not affiliated with Sarvam AI, AI4Bharat, BharatGen or the Government of India. Every other checkpoint in this project is trained at 512 tokens. This one answers what happens if you…
No final model weights are uploaded to this repository yet. This page reserves the final model's release address; it is not a loadable model. The intended student is based on Qwen3.5-9B and learns from spatial specialists through selective multi-teacher distillation. Source-native G, S, A and joint SFT preparation adapters are separately staged in a private repository. They are not relabeled as the final MOPD model. Full native-output evaluation remains in progress. Final MOPD training and a weight-hash-bound evaluation report are required before weights appear here. Manuscript-reported results and controlled mechanism studies should not be interpreted as measurements of weights in this…
weighted/imatrix quants of https://huggingface.co/YOON1v/Apex-2-Base
Ciru v4.1.2 — October 5, 2026. Applies the shared output-budget correction; Apodex retains its own native MTP4 implementation. Explicit 262,144-token output ceilings now use the fully rendered prompt's remaining context. Weights, pinned engine wheels and sampling defaults are unchanged. Patch notes · Upgrade existing installations. OMP clients must advertise input: [text, image], reasoning: true, contextWindow: 262144 and maxTokens: 262144 to use vision and the full output ceiling. Ciru v4.1.1 — September 30, 2026. Qualified repairs and a retained subset of community PR #2. Apodex retains its own native MTP4, model-specific state layout and released recurrence/attention implementation. The…
Aqua is a 100-million-parameter model trained on vast amounts of curated, organized data. It is designed for general-purpose responses and is retrained daily. Its intended use is as a base model for fine-tuning by companies or individuals, and it is released under the Apache 2.0 license for flexible use. Although this version has been trained on highly refined data up to 2026, it is the smallest variant. We will soon launch Aqua-3.0.0, a model optimized for more fluent Spanish conversations, with future plans to add support for English, Chinese, Arabic, and Russian.
本仓库（Repository）所包含的所有人工神经网络（Artificial Neural Network）权重文件（.pth/.index）、训练日志及相关代码，均为计算声学（Computational Acoustics）与深度学习（Deep Learning）领域的技术研究实验产物。 这些文件本质上是高维张量（High-dimensional Tensors）的数值序列，通过随机梯度下降（SGD）与反向传播算法（Backpropagation）对公开可获取的音频数据进行统计建模（Statistical Modeling）得到。其技术形态与数字图像处理中的卷积核（Convolution Kernels）、自然语言处理中的词向量（Word Embeddings）并无本质差异。 本仓库不构成对任何第三方知识产权的故意侵犯，所有代码遵循 MIT License 开源协议，模型权重文件仅作为技术实现的副产品（By-products）存在。 本技术实验所使用的训练数据集（Training Dataset）包含以下角色的语音样本： - Mon3tr（モンスター）、Mon3tr 中文语音 等 - 声源版权归属：上海鹰角网络科技有限公司（Hypergryph Network Technology Co., Ltd.）及其关联公司 - 原始作品：《明日方舟》（Arknights） 上述角色的声音版权、肖像权、姓名权及相关知识产权均完全归属于鹰角网络及其合法授权方。本实验仅基于已公开发布的游戏内语音资源进行技术层面的信号处理（Signal Processing）与特征提取（Feature…
This is a calibrated NVIDIA NVFP4 conversion of TheDrummer/Artemis-31B-v1.2, a fine-tune of google/gemma-4-31B. It is provided as a compressed-tensors safetensors checkpoint for runtimes that support Gemma 4 and nvfp4-pack-quantized. The original fine-tune is credited to TheDrummer; this repository contains a precision conversion. The checkpoint has two safetensors shards totaling 20,446,828,448 bytes (19.04 GiB), plus its configuration, tokenizer, chat template, processor configuration, recipe, and conversion manifest. Download the whole repository to use this format. The two weight shards alone are insufficient. - The tokenizer and chat template match the source checkpoint by SHA-256.…
This is a text-model GGUF conversion of TheDrummer/Artemis-31B-v1.2, a fine-tune of google/gemma-4-31B. The language model's eligible linear weights were calibrated and quantized to NVIDIA NVFP4, then repacked into GGUF. The original fine-tune is credited to TheDrummer; this repository contains a format and precision conversion. Load Artemis-31B-v1.2-NVFP4.gguf as the model in a Gemma 4 and NVFP4-capable GGUF runtime. KoboldCpp 1.121 loaded this file and generated text successfully in a local test. The source model's Gemma 4 chat template is embedded in the GGUF. Use the Gemma4 31B vision tower if you want image recognition. On the 299-question ARC-Challenge validation split, using…
Drummer proudly presents... - Gemma 4 31B template, thinking or non-thinking. - Sampler wrangling may be required (or be conservative) Artemis v1.2 landed as the strongest Artemis yet — and, for several testers, the strongest Gemma 4 tune they've used. The recurring themes: - Writing that doesn't feel like G4. Immersive, creative prose with genuinely human-sounding dialogue; multiple testers made it their daily driver, and one got so absorbed in the character they forgot to test darker themes at all. - Reasoning that actually works. The think block pulls the right details and the response stays faithful to it — no cross-card personality bleed at ~32k, scenario adherence holding at 20k…
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.