small-models
39 posts tagged small-models.
- Tether’s BotPsy points to smaller phone agents, but the proof is still thin
- The Label Wording Problem Nobody Tests in Agent Routers
- LoopVQ and the Hidden Memory Cost of Looped Transformers
- Memdoor's Claim That Five Small Models Beat One Big One on Hard Coding
- The Economics of Proactive AI: When a Single 1B Decoder Beats a Two-Model Pipeline
- A 0.6B Persian Embedding Model That Catches a 4B: What Yalda Teaches About Distillation
- A 4B Qwen fine-tune points at local 3D scene generation
- Nanbeige4.3-33B is really a small-model bet
- What Samsung's On-TV TensorFlow Lite Tracker Tells You About Edge AI
- When a Benchmark Claim Has No Benchmark Behind It
- TS2Vec-FM shows small convolutional time-series models still have legs
- What DeepSeek V4.1 Flash on Dwarfstar tells us about SSD-streaming local models
- What greedy sampling and expert-skip defaults cost you on a single-4090 Qwen3.8 rig
- DoubtLess uses a small model to make a big one stop overthinking
- On-device agents are a systems bet, not just a model release
- 100B-parameter phone models are a battery and memory bet
- The missing model is becoming a product requirement
- Multi-teacher distillation should copy the delta, not the whole teacher
- EmbeddingGemma 2 and the push toward smaller multimodal retrieval
- Paradee makes single-voice TTS small enough for a laptop
- Running Anomaly Detection on the Satellite Instead of the Ground
- KaliBench Measures Whether LLMs Can Actually Type the Right Security Command
- Telescopic Language Models: one training run that serves every compute budget
- Ling 3.0 Tiny makes old CPUs interesting again
- When a Model Gets Math Right but Reads the Same Problem Differently
- A Small Model That Steers a Bigger One's Reasoning
- Xiaomi's MiMo V2.6 Ships in Three Flavors, and the Split Matters
- Mini-AGI makes local training the interesting part
- When the Teacher Knows to Quit: RetireOPD and Distillation for Agents
- What llama.cpp v0.4.1 tells you about where local AI is heading
- Local LLMs are getting useful because constraints are back
- Compiling a Prompt Into a Small Neural Function You Can Version
- Small dialogue agents need repair loops, not just bigger training
- Puro-2B makes small-model pretraining feel operator-sized
- Smaller AI models need phone-level proof
- Daedalus-150M treats CPU inference as the design target
- Relay-OPD and the prefix failure problem in on-policy distillation
- MiniCPM5-1B points at the phone-sized agent layer
- Compiling a Prompt Into Weights: What Program-as-Weights Actually Changes