microsoft/VibeVoice-ASR-HF
Audio-Text-to-Text • 8B • Updated • 158k • 59
Tom Whitworth
twhitworth
AI & ML interests
Low Level Infrastructure, Tooling, Busyforking
Recent Activity
liked a model 20 days ago
microsoft/VibeVoice-ASR-HF liked a model 20 days ago
LocoreMind/LocoOperator-4B liked a model 24 days ago
nvidia/nemotron-speech-streaming-en-0.6bOrganizations
liked 2
models 20 days ago
liked 2
models 24 days ago
liked 3
models 25 days ago
reactedto OzTianlu's post with 🚀 about 1 month ago
Post
3461
O(1) inference is the foundational design of Spartacus-1B-Instruct 🛡️ !
NoesisLab/Spartacus-1B-Instruct
We have successfully replaced the KV-cache bottleneck inherent in Softmax Attention with Causal Monoid State Compression. By defining the causal history as a monoid recurrence, , the entire prefix is lossily compressed into a fixed-size state matrix per head.
The technical core of this architecture relies on the associativity of the monoid operator:
Training: parallel prefix scan using Triton-accelerated JIT kernels to compute all prefix states simultaneously.
Inference: True sequential updates. Memory and time complexity per token are decoupled from sequence length.
Explicit Causality: We discard RoPE and attention masks. Causality is a first-class citizen, explicitly modeled through learned, content-dependent decay gates.
Current zero-shot benchmarks demonstrate that Spartacus-1B-Instruct (1.3B) is already outperforming established sub-quadratic models like Mamba-1.4B and RWKV-6-1.6B on ARC-Challenge (0.3063). Recent integration of structured Chain-of-Thought (CoT) data has further pushed reasoning accuracy to 75%.
The "Spartacus" era is about scaling intelligence, not the memory wall ♾️.
NoesisLab/Spartacus-1B-Instruct
We have successfully replaced the KV-cache bottleneck inherent in Softmax Attention with Causal Monoid State Compression. By defining the causal history as a monoid recurrence, , the entire prefix is lossily compressed into a fixed-size state matrix per head.
The technical core of this architecture relies on the associativity of the monoid operator:
Training: parallel prefix scan using Triton-accelerated JIT kernels to compute all prefix states simultaneously.
Inference: True sequential updates. Memory and time complexity per token are decoupled from sequence length.
Explicit Causality: We discard RoPE and attention masks. Causality is a first-class citizen, explicitly modeled through learned, content-dependent decay gates.
Current zero-shot benchmarks demonstrate that Spartacus-1B-Instruct (1.3B) is already outperforming established sub-quadratic models like Mamba-1.4B and RWKV-6-1.6B on ARC-Challenge (0.3063). Recent integration of structured Chain-of-Thought (CoT) data has further pushed reasoning accuracy to 75%.
The "Spartacus" era is about scaling intelligence, not the memory wall ♾️.
reactedto ajibawa-2023's post with 🔥 about 1 month ago
Post
3254
JavaScript-Code-Large
ajibawa-2023/JavaScript-Code-Large
JavaScript-Code-Large is a large-scale corpus of JavaScript source code comprising around 5 million JavaScript files. The dataset is designed to support research in large language model (LLM) pretraining, code intelligence, software engineering automation, and program analysis for the JavaScript ecosystem.
By providing a high-volume, language-specific corpus, JavaScript-Code-Large enables systematic experimentation in JavaScript-focused model training, domain adaptation, and downstream code understanding tasks.
JavaScript-Code-Large addresses the need for a dedicated JavaScript-only dataset at substantial scale, enabling focused research across frontend, backend, and full-stack JavaScript environments. .
ajibawa-2023/JavaScript-Code-Large
JavaScript-Code-Large is a large-scale corpus of JavaScript source code comprising around 5 million JavaScript files. The dataset is designed to support research in large language model (LLM) pretraining, code intelligence, software engineering automation, and program analysis for the JavaScript ecosystem.
By providing a high-volume, language-specific corpus, JavaScript-Code-Large enables systematic experimentation in JavaScript-focused model training, domain adaptation, and downstream code understanding tasks.
JavaScript-Code-Large addresses the need for a dedicated JavaScript-only dataset at substantial scale, enabling focused research across frontend, backend, and full-stack JavaScript environments. .
liked 8
models about 1 month ago
inclusionAI/ZwZ-4B
Image-Text-to-Text • 5B • Updated • 2.2k • 29
MiniMaxAI/MiniMax-M2.5
Text Generation • 229B • Updated • 491k • • 1.26k
ai21labs/Jamba-tiny-reward-dev
Updated • 7.3k • 2
ai21labs/Jamba-tiny-dev
0.3B • Updated • 316k • 14
rednote-hilab/dots.ocr
Image-Text-to-Text • 3B • Updated • 256k • 1.28k
nvidia/Qwen3-Next-80B-A3B-Thinking-NVFP4
Text Generation • Updated • 115k • 53
microsoft/InfoAgent
Updated • 14
ibm-granite/granite-vision-3.3-2b-chart2csv-preview
Image-Text-to-Text • 3B • Updated • 3.46k • 14
liked 3
models about 2 months ago
