Open-Weight Models

Ru-Small-Instruct — экспериментальная компактная русскоязычная языковая модель класса SLM (Small Language Model) с объемом параметров ~0.2B (~165M). Разработана с упором на суверенность весов (Zero-Fingerprint): модель обучена с нуля без заимствования базовых чекпоинтов у сторонних корпоративных сетей (Llama 3 от Meta, Qwen от Alibaba, Mistral). Модель предназначена для исследований локального инференса, работы на маломощном оборудовании, CPU и мобильных чипах, где критичны нулевая задержка (Time-To-First-Token) и полная независимость весов. Для компактной модели в 165M параметров, обученной на одном домашнем GPU за 48 часов, способность держать роль, грамотно формулировать сложные термины…

Model

S-AGI

Convergent Intelligence

Part of the Standalone Models by Convergent Intelligence LLC: Research Division This model is part of the Convergent Intelligence LLC: Research Division portfolio. All models in this portfolio are developed under the Discrepancy Calculus (DISC) framework — a measure-theoretic approach to understanding and controlling the gap between what a model should produce and what it actually produces. DISC treats training singularities (loss plateaus, mode collapse, catastrophic forgetting) not as failures to be smoothed over, but as structural signals that reveal the geometry of the learning problem. Key concepts: For the full mathematical treatment, see Discrepancy Calculus: Foundations and Core…

Converted with mlx-lm 0.31.3. A six-case deterministic sanity check matched three upstream reference strings exactly. The remaining differences included a retained leading “So,” punctuation/ordinal variation, and omission of “tomorrow” in one correction case. This is a small functional check, not the upstream 7,519-case evaluation; assess the 4-bit build on your own transcripts. A 0.6B-parameter text normalizer for speech-to-text output. It takes a raw ASR transcript and rewrites it as clean written text: fillers removed, false starts and self-corrections resolved to the value the speaker landed on, punctuation and capitalization applied, and spoken numbers, dates, times, currency and email…

Converted with mlx-lm 0.31.3. A six-case deterministic sanity check matched four upstream reference strings exactly. The two differences were a retained leading “So,” and one comma variation. This is a small functional check, not the upstream 7,519-case evaluation. A 0.6B-parameter text normalizer for speech-to-text output. It takes a raw ASR transcript and rewrites it as clean written text: fillers removed, false starts and self-corrections resolved to the value the speaker landed on, punctuation and capitalization applied, and spoken numbers, dates, times, currency and email addresses rendered in written form. On a held-out set of 7,519 English cases it reaches 94.8% token accuracy, and…

Model · Robotics

SOMA-X

NVIDIA

SOMA (Unifying Parametric Human Body Models) is a unified framework that decouples identity representation from pose parameterization by mapping supported parametric models to canonical body and hand topologies and skeletons, enabling shared Linear Blend Skinning (LBS) pipelines across backends. The full-body layer supports six identity backends (SOMA-shape, SMPL, SMPL-X, MHR, ANNY, and GarmentMeasurements). SOMA-X v0.3 also includes wrist-local left/right hand layers with native SOMA identity and articulation priors plus interoperability with user-supplied MANO models. This model is ready for commercial use. SOMA is released under the Global SOMA is intended for use by computer vision…

Detector weights and derived bases from a controlled study of occlusion compensation in aerial wildlife monitoring, along two axes: single view vs multi view, and sensor (pixel) space vs embedding space. container it ran in, every scored metric behind the paper's tables, and a step-by-step in REPLICATION.md. Run directory names here are the run identifiers used there, so a checkpoint joins directly against metrics/.json under the arms key. Backbones are not included. DINOv3 and V-JEPA 2.1 are used frozen and unmodified — obtain them from their original sources. Only self-trained artefacts are here. Nearly everything here is cross-validated. A run directory whose name carries f belongs to…

This is Qwen3.8-27B that gets to the answer faster. AgentionAI Signal is a minimally invasive fine-tune of Qwen3.8-27B designed for lower generation latency and better token efficiency. On our held-out general-prompt evaluation, Signal produces 57% fewer answer tokens and uses 52% fewer thinking tokens, while matching or improving the measured answer quality of the base model. The percentages above were measured on the first release. The weights updated on 2026-09-13 trade a little of that reduction for stability; their re-measurement on the same prompt set is in progress and will replace these numbers. The result is substantially faster end-to-end generation: on typical chat prompts…

Model · Robotics

SmolVLA-Base-0.45B

Mithilesh Nakade

This repository contains weights or code derived from the SmolVLA foundational architecture developed by Hugging Face and the LeRobot Authors. This is SmolVLA-Base model cloned from Hugginface "lerobot/smolvlabase" repository. This was createed for ready-to-use custom model for easy inference during Hackathon challenge.

Private backup of trained policy weights (.pt only, no optimizer states).

Model

Structure-Over-Scale

Convergent Intelligence

Roy C March 2026 We present a methodology for training small language models on CPU at FP32 precision that achieves capability-per-dollar efficiency orders of magnitude beyond GPU-based training. Across 15 models spanning four novel architecture families — Mixture of Attentions (MoA), cross-architecture fusion (Qemma), swarm intelligence (SAGI), and metric-space causal language models (DiscoverLM) — total compute cost was $24 on a single AMD EPYC 9454P processor. We introduce seven methodological pillars: (1) FP32 precision preservation, with experiments demonstrating 5,810x single-operation error and 23,225x compounding error ratio for FP16 at network depth; (2) sparse cognitive…

An INT8 W8A8 quantization of (the abliterated BF16 build of a merged LoRA finetune of Qwen/Qwen3.8-27B), for fast serving on GPUs where native FP8 is unavailable or undesirable. int-quantized, native CUTLASS INT8 tensor-core path. embedtokens, the vision tower, GatedDeltaNet recurrent gates, all norms. Note these are FP16, not the source's BF16 — the 16-bit residual precision of this build is float16 throughout (config.json: "dtype": "float16"). (llm-compressor), with the MTP drafter ablated in the rotated basis. The MTP head is unquantized but is not byte-identical to the BF16 build's: it carries the same residual-stream rotation and RMSNorm fold as the rest of the model, so the two heads…

Refusal-ablated Ternary Bonsai 2 27B, edited directly on the ternary lattice and shipped in PrismML's native PTQ10 pack. - 5,946,648,928 bytes — byte-for-byte the size of the official PTQ10 release - No requantization, no BF16 bake, no runtime steering, no control vector at serve time - Requires the PrismML llama.cpp fork Differs from PrismML's release in exactly 98 tensors and nowhere else. The other 753 tensors are byte-identical, including tokenembd and output. Anyone with both files can check that without trusting us. Paired against the untouched PrismML release: same card, same runtime, same flags, same seed, enablethinking: false on both arms. Judge is Qwen3.8-27B-OBLITERATED-Q80 with…