Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up

All HF Hub posts
yanghaojin 
posted an update 2 days ago
view post
Post
4,748
DeepSeek-V4.1-Flash-1017, running locally on a Mac

Our 4-bit paged build is up: GreenBitAI/DeepSeek-V4.1-Flash-1017-4bit-paged

On one M3 Ultra Mac Studio with gbx:
• ~22 tok/s standard decode
• ~44 tok/s on code with DeepSeek's official draft head (1.99×)
• ~33 tok/s on chat with the draft head

Why run 1017: DeepSeek reports 77.0 on DeepSWE v1.1 (up from 74.2) and 42.3 on Terminal-Bench 4.0 (up from 31.2), with shorter outputs (https://api-docs.deepseek.com/news/news261017).

It reads text and images, and plugs straight into Claude Code and Codex; the setup is in the model card. The paged format also loads on Macs with less memory, and on an NVIDIA DGX Spark with experts paged from the SSD;

Engine: https://github.com/GreenBitAI/gbx-lm/releases
DedeProGames 
posted an update 3 days ago
view post
Post
8,744
🚀 Introducing Kiyo-230M-Instruct — my compact English & Portuguese chat model!

Built on Kiyo-230M-Preview, a base model I pretrained from scratch, Kiyo now has an instruction-tuned release for everyday conversations, rewriting, and lightweight assistant experiments.

🧠 229.7M parameters
🌍 English + Portuguese
💬 Native chat template for multi-turn conversations
📏 2,048-token context window
📦 459 MB of BF16 safetensors weights
🔓 Apache-2.0 license

For instruction tuning, I used 165,735 conversations from SmolTalk2, smol-smoltalk, UltraChat, and Dolci-Instruct-SFT. Training ran for two complete epochs with assistant-only loss, using a roughly 70% English / 30% Portuguese token mix.

The updated model card includes benchmarks against SmolLM2-135M-Instruct, BananaMind, and Supra2, plus example conversations, CPU measurements, training details, and a Transformers quickstart.

The results are mixed: Kiyo is competitive with these small models across the reported evaluations, while SmolLM2 and BananaMind score higher on the instruction-following benchmarks. Portuguese was a highlight in the small conversational sanity test.

This is an experimental 230M model, with plenty of room to improve reasoning, factual accuracy, and strict instruction following. I’m excited to share it with people exploring how much we can get from small, accessible models!

🤗 Model & full README:
DedeProGames/Kiyo-230M-Instruct

💬 Try the chat demo:
DedeProGames/kiyo-230m-chat

Try it in English or Portuguese — I’d love to see your examples, feedback, and experiments!
  • 7 replies
·
gianlucamazza 
posted an update 1 day ago
view post
Post
2,107
GGUF models from xllama's catalogue, my independent research project for local LLM inference on an Xbox Series S. Each note gives the measured figures or test status, with model, quant and setup. For example, LFM2.5-350M (QAD Q4_0, Dev Mode, llama.cpp CPU) decodes at 101.6 tok/s. Source for every number: xllama's model matrix. Not affiliated with Microsoft.
gianlucamazza/xllama-models-measured-on-xbox-series-s-6aca0b32275cbe99dbf02d89
Mati83moni 
posted an update 1 day ago
view post
Post
3,880
Is "evaluation awareness" in a language model one direction, or eight?

We held the task fixed, toggled the eight cue factors of EvalAwareBench one at a time, and looked at the residual stream of Qwen3 (0.6B–32B), Gemma 3 and PLLuM-12B, plus a first Polish cross-lingual pilot.

What we found (descriptive, same instrument and author throughout):
- Cue directions overlap at about 1/5–1/4 of what two estimates of the same cue reach: not one shared axis.
- The standard all-cues direction is mostly an explicit-disclosure detector (cosine 0.81–0.89 with that cue).
- It transfers poorly to stimulus artificiality (held-out AUROC ≈ 0.5–0.6), though the models do represent that cue.
- No statistically detectable MMLU framing effect in the tested settings. That does not establish absence of an effect or of sandbagging.
- The Polish pilot is inconclusive: the formal verdict says so, and its own length control was not met.

The article carries its corrections (three early findings were withdrawn at scale; one traced to a bug in our hook code), an open validation problem for the add interventions, and all 23 figures with sources.

Full article: https://huggingface.co/blog/Mati83moni/is-evaluation-awareness-one-direction-or-eight

Spaces:
Mati83moni/evaluation-awareness-one-direction-or-eight
Mati83moni/evaluation-awareness-pllum-12b
  • 2 replies
·
eaddario 
posted an update 2 days ago
view post
Post
4,810
Experimental global target bits‑per‑weight quantization of IFM/K2-Horizon-7B and IFM/K2-Horizon-MoVA-36B-A4B.

Unlike standard llama.cpp quantization that rely on fixed type heuristics (e.g., Q4_K_M), the Target BPW approach automatically optimizes per-tensor precision where it matters the most, and produces high quality models that meet a precise global size target.

Key Advantages:
- VRAM Maximization: Can generate high quality models sized exactly to fit hardware constraints (e.g., fitting the model into exactly 24GB VRAM).
- Data-Driven Precision: Quantization mix is determined by actual weight error sensitivity rather than hardcoded rules, often yielding better PPL/KLD size trade-offs.

Full benchmarks (PPL, KLD, ARC, GPQA, MMLU, etc.) and methodology in the model's card.

eaddario/K2-Horizon-7B-GGUF
eaddario/K2-Horizon-MoVA-36B-A4B-GGUF
DavidAU 
posted an update 3 days ago
view post
Post
3,204
TURBO-BRILLIANCE 50X - Qwen 3.8 27B, Qwen-Flash Next (120B) Reg and Ablit/Uncensored, NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB

15 AI CONSTRUCTS, SOTA Models + Turbo Brilliance:

Each a precision engineered "consciousness" to help you in every task, using fewer tokens and generating better answers.

This model has significant internal upgrades under user control that are "on the fly" (you set in your prompt(s) and/or API).

Each mode / construct reduces over all token usage and increases answer quality...

Please read this model card carefully to get the most from it - including all the reasoning and instruct modes (27+27+9 xhigh) that feature model suggestions ("context help") for all use cases.

The Turbo Brilliance system will begin appearing in other model sizes/types shortly (IE: Qwen, Gemma, Mistral, GLM, Llama, ... including our strongest fine tunes too). The "rollout" is now in progress...

Qwen-Flash-Next 120B, and run on 12GB GPU too (special project) at high speed, with RSQ-RCO BF16 level quants :

REG Version:
DavidAU/Qwen3.8-Flash-Next-Turbo-Brilliance-50x-Reasoning-Instruct-GSQ-RCO-GGUF

Abliterated/Uncensored:
DavidAU/Qwen3.8-Flash-Next-Abliterated-Uncensored-Turbo-Brilliance-50x-Reasoning-Instruct-GSQ-RCO-GGUF

NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB: (2 bit, QAT, at close to BF16 performance , only 22GB!)

DavidAU/NVIDIA-Nemotron-3.5-Bolt-Turbo-Brilliance-50x-74B-A9B-24GB-GGUF

Qwen 3.8, 27B version:
DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF

PS: Qwen Next, LFM2.5 24B-A2B, Qwen 3.6 35B-A3B and others in testing now / coming soon.

COLLECTION:
https://huggingface.co/collections/DavidAU/turbo-brilliance-ai-constructs-50x-reasoning-instruct

Please see our support page for breakdown of models (sort), info on projects and a lot more information:

DavidAU/support
Banaxi-Tech 
posted an update 4 days ago
view post
Post
7,257
hi everyone
we at
klein-ml are announcing KleinMem

These models will have Native memory, without any system-prompt or prompt, it still remembers exactly.

Get early access at https://tinyurl.com/klein-eap limited to 15 users.


Architecture details are coming soon toon.

If you know alot about architecture and want to help build it, join
klein-ml .
danielhanchen 
posted an update about 8 hours ago
alibidaran 
posted an update about 18 hours ago
view post
Post
526
I have recently worked on a personal project called progressive thinking, which enables LLMs to generate chains of thought in a progressive pattern. In addition, the model learned to generate COT before choosing the appropriate tools, making it useful for complex and deeper agentic tasks. This model is well trained on QWEN3-14B using the SFT algorithm and needs GRPO to make feasible outputs.
alibidaran/qwen_progressive_thinking_16bit
OpenDataArena/Spark-234K
IFM/TxT360-3efforts
etemiz 
posted an update 1 day ago
view post
Post
1,310
same machine, different intention
there is nothing inherently wrong about AI
  • 2 replies
·
Recently active users