AiRukua/150M-Kilat-SMALL-10B
text-generation model by AiRukua
From the publisher
Model card & documentation
Source previewRead the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.
Kilat SMALL 10B is a 151M-parameter Base Pretrained Language Model developed using the Kilat v2.0 architecture and pretrained from scratch on 10 billion tokens of FineWeb-Edu. The model is built around a hybrid architecture that combines three complementary approaches: The model was trained using Kilat Macro-Synergy, a training framework developed for the Kilat architecture. This release represents the base pretrained stage of Kilat. It has not undergone supervised instruction fine-tuning (SFT), preference optimization, or conversational alignment. Kilat incorporates a Global Decay mechanism for efficient sequence modeling. The mechanism uses a recurrent state representation with a global decay factor, allowing information to propagate across sequence positions without maintaining a conventional full attention matrix. Key characteristics: The architecture incorporates Latent Multi-Head Attention, inspired by the latent attention design introduced in DeepSeek models. Instead of directly maintaining full-dimensional key-value representations, MLA uses compressed latent representations to reduce KV-cache memory requirements. Key characteristics: Kilat…
Read the full model card ↗ · Preview checked 2026-11-15T07:23:37.298Z
Inside the original model card — Document outline
- 150M - Kilat SMALL 10B
- Model Specifications
- Architecture Overview
- 1. Global Decay
- 2. Latent Multi-Head Attention (MLA)
- 3. DeepSeek-style Mixture-of-Experts
- Pretraining
- Training Performance
- Intended Use
- Recommended Applications
- Inference
- Model Limitations
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Documentation belongs to its respective authors. Reported project/model license: apache-2.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
Model overview
AiRukua/150M-Kilat-SMALL-10B is a text-generation model repository published by AiRukua on Hugging Face. A library was not reported.
This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.
Model facts
- Task
- text-generation
- Library
- Not reported
- Recent downloads (30 days)
- 0
- Cumulative likes
- 0
- Hugging Face trending score
- 0
- Reported safetensors parameters
- 150,950,244
- Architecture
- KilatForCausalLM
- License
- apache-2.0
- Access
- Not gated by Hugging Face
- Created
- 2026-11-15T06:07:11.000Z
- Last modified
- 2026-11-15T06:54:31.000Z
Compatibility and lineage
Language coverage was not reported.
Base-model lineage was not included.
Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.
Use and evaluate
Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.
No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.
Source and freshness
Source: Hugging Face Hub. Metadata observed 2026-11-15T06:59:14.047Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.