Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up

All HF Hub posts
danielhanchenΒ 
posted an update 2 days ago
DedeProGamesΒ 
posted an update 2 days ago
view post
Post
2958
πŸš€ OxCoder-9B β€” a lightweight agentic coding model, now on HF!

Introducing OxCoder-9B, a 9B parameter model built for long-horizon tasks, agentic coding, and agentic reasoning. Despite its compact size, it delivers frontier-level performance in Agentic Terminal and Agentic Coding, rivaling models many times its size.

Highlights:
- Trained on frontier agent traces β€” distilled from Fable-5.1 and GLM-5.3 agentic coding trajectories across Claude Code, OpenCode, and Codex
- 262K native context β€” handles complex, multi-file codebases and long-horizon reasoning tasks with ease
- Error recovery β€” learns read-before-write patterns, responds to LSP diagnostics, and applies minimal edit diffs instead of full rewrites
- Strong front-end reasoning β€” deep understanding of UI logic, component architecture, and web-native patterns, rare in sub-10B models

Benchmarks (vs. Ornith-1.5-9B, Ornith-1.0-9B, Qwen3.5-9B, and Gemma-4-31B):
- Terminal-Bench 2.1 (Terminus-2): 49.6
- Terminal-Bench 2.1 (Claude Code): 50.8
- SWE-bench Verified: 73.5
- SWE-bench Pro: 49.1
- NL2Repo: 36.2
- HLE (no tools): 21.2
- HLE (with tools): 32.8
- GPQA Diamond: 86.9
- MCP-Atlas: 56.7
- BrowseComp: 57.4
- ClawEval: 67.8

All OxCoder-9B results are averaged over five independent runs. Built on Qwen/Qwen3.5-9B, released under Apache 2.0.

πŸ”— OrionLLM/OxCoder-9B
  • 2 replies
Β·
anakin87Β 
posted an update 1 day ago
view post
Post
2245
I made a 1.1M ModernBERT encoder play Doom in real time on a CPU

Some time ago, VAGO Solutions released SauerkrautLM-Doom-MultiVec-1.3M, a tiny model trained to play Doom Defend the Center scenario from 31k human gameplay examples.

My first thought: cool! I love both Doom and Small Language Models.

Then another idea: I bet I can do better :-)

What I did?
- evaluated the original model and found it's better than reported
- changed a bit the architecture
- generated SFT data with a scripted oracle
- SFT + PPO refinement on consumer hardware

Got a smaller, faster and killer model
Can even fit a floppy with int8 quantization πŸ’Ύ

Watch it play/read the article: anakin87/tiny-doom-defender
nwaughachukwumaΒ 
posted an update 2 days ago
view post
Post
2630
It’s easy to get distracted by benchmarks, throughput (tok/s), and all the hype around frontier model releases.

This is Shiny Model Syndrome, which makes engineers and teams forget the basic physics of production software, i.e., using the right tool for the job and optimizing for ease of integration.

- Teams spend huge amounts of money on frontier models for document parsing, OCR, detection, segmentation, and other task-specific visual AI workflows.

- Inference marketplaces don’t find it profitable to list task-specific models like glm-ocr, paddleocr, or dots.mocr, even though they’re all superior to frontier VLMs for document parsing and OCR.

- Engineers stitch together multiple endpoints for different use cases across the long tail of visual AI. Those who choose to self-host instead deal with painful infrastructure and GPU ops.

At VLM Run, we wanted one place to run OCR models, VLMs, and ViTs that we could confidently use for our own internal agents and evals. The gateway was born out of that need, and we’ve since opened it to the public.

The gateway exposes a single OpenAI-compatible endpoint for the long tail of visual AI across OCR, document parsing, VQA, detection, segmentation, embeddings, and transcription. Simply point the base_url of your OpenAI SDK at gateway.vlm.run/v1/openai, or ask your agent to connect via MCP (gateway.vlm.run/mcp).

You can swap the model name to compare glm-ocr, dots.mocr, paddleocr-vl-1.6, qwen3.8-27b, gemma4-26b-a4b, and more. We handle serving, runtime, and pipelining behind the scenes to give you high-quality visual intelligence.

- https://vlm.run/gateway
- https://huggingface.co/blog/vlm-run/introducing-gateway
- https://www.vlm.run/blog/introducing-gateway
GoktugDΒ 
posted an update 2 days ago
view post
Post
2247
πŸ‡ΉπŸ‡· One of our small Turkish models quietly reached **500+ monthly downloads** on Hugging Face.

**Werea-TR-TextRestore β€” only 300M parameters.**

Its job is simple:

istanbulda hava cok guzel
β†’ Δ°stanbul'da hava Γ§ok gΓΌzel.

A lightweight model for restoring Turkish text:
β€’ diacritics
β€’ punctuation
β€’ casing
β€’ corrupted text

**96.5% word accuracy** on real Turkish news sentences.

And it runs without sending your text to a cloud API.

πŸ€— Try the model:
Werea-co/Werea-TR-TextRestore

πŸ‡ΉπŸ‡· Built in TΓΌrkiye. Open source.

If you're working on Turkish NLP, I'd love to hear what we should build next.

#TurkishNLP #HuggingFace #OpenSourceAI #NLP
  • 2 replies
Β·
RN0311Β 
posted an update 1 day ago
view post
Post
2043
SONAR is now open-source! πŸ”Š

SONAR is an evaluation toolkit for multilingual ASR that goes beyond WER/CER. It combines semantic similarity, the Poseidon Score, and analysis across dialect, demographic, and metadata-based failure modes. 🌍

Our goal is to make it easier for everyone to understand why an ASR model fails, not just how often. πŸ” You can plug in your own models + audio, extend it to new languages and datasets, or contribute directly. πŸ› οΈ

MIT licensed. Would love feedback from the HF community! πŸ€—

πŸ”— GitHub: https://github.com/PSDN-AI/SONAR-OSS/
πŸ”— Blog: https://www.psdn.ai/blog/open-source-multilingual-asr-evaluation
OppaAIΒ 
posted an update 1 day ago
view post
Post
1966
Another small 4B model comes out yesterday.
NeoHorse 1 4B
TokenRhythm/NeoHorse-1-4B

There are quite a few good smaller parameter models that are capable for Agentic tasks:

The ones from the chart, I have tried a few already in my Jetson Orin Nano,
❌Gemma4 E2B IT - cannot fit my RAM usage if use with TTS and embedder
❓Qwen3.5 4B - just barely fit my RAM usage, need to add think/no_think
❌Spark X2.5 4B - need to build the forked llama.cpp; no vision
➑️Nanbeige 4.2 3B - need to build the forked llama.cpp; slower than Ministral3-3B by 25%; no vision but good for coding; maybe run this is separate server for doing coding tasks
➑️Agents A1 4B - This one is quite interesting. Another Qwen3.5 4B base. I just learnt this right now. This model may surpassed the Ministral3-3B that I'm currently running.
➑️NeoHorse 1 4B - wait for GGUF version comes out; Qwen 3.5 4B base with vision striped
➑️Needle2 45M - need to use separately from llama.cpp server; currently testing to see if it can be used as spawning sub-agents to do parallel tasks
  • 7 replies
Β·
Hoglet-33Β 
posted an update 3 days ago
view post
Post
3139
Today, we planned to release Pebble-50M and Pebble-50M-Chat to the world. Unfortunately, due to a few issues, that didn't go quite as planned.

What happened:
- Some data and benchmark results were lost or corrupted
- The models performed worse on benchmarks than our other Pebble models

Despite that, you can still find both models here:
Pebble-50M-beta: basically-experimental/Pebble-50M-beta
Pebble-50M-Chat-beta: basically-experimental/Pebble-50M-Chat-beta

There are still some interesting improvements in these models:
- Compatible with non-CUDA devices
- Vocabulary increased to 16K tokens
- Context length increased to 16K tokens

For now, there won't be any more Pebble releases for a while. We're going to take some time to experiment with other approaches and hopefully make the next generation a monumental leap over this one.

Follow for updates:
@Hoglet-33
basically-ai
basically-experimental
  • 4 replies
Β·
SoulInPsyAbstractΒ 
posted an update 2 days ago
view post
Post
2179
Loss went from 2.35 to 0.27 in 50 steps. Clean, textbook convergence curve.
Held-out score: 0/10 before fine-tuning. 0/10 after.
Ran a before/after LoRA fine-tune on IFM/K2-Horizon-0.9B (Apache 2.0, released this week) on a binary fabrication-detection gate β€” entirely on a free CPU tier, no GPU. The training loss says it learned something real. The eval says it learned nothing that generalizes.
Looked at the actual raw outputs instead of trusting the score. Both before and after, the model never once emits TRUE or FALSE β€” it just continues the system prompt as text: "The user is asking me..." before, "The user is asking for..." after. Fine-tuning moved the failure string by two words. It didn't touch the failure.
Most likely cause: a 0.9B model that doesn't follow "answer in one word" untrained won't learn to from 50 gradient steps on 194 rows β€” it'll memorize the training set's surface form (which is exactly what the loss curve shows) without the instruction-following ever kicking in. The fix isn't more steps, it's checking base instruction-following before spending the run at all.
Writing up a clean negative is worth exactly as much as a clean positive if you actually open the raw generations instead of stopping at the number. Full run, scripts, and raw JSON:
SoulInPsyAbstract/sipa-os-governance
  • 2 replies
Β·
Bc-AIΒ 
posted an update 2 days ago
view post
Post
2554
Hello everyone!
I have 2 announcements today!
The first one is the launch of our new API platform! You can make a account and get 5 dollars free credits. No credits card needed because i have no idea how to set up a payment's thing. If you want more credits just email me at smilyai@outlook.com .
The platform currently features G1-Preview a preview of G1 and the older Mira-1-Large.
2nd announcement is we have started working on G1-MINI so expect a late October Ish launch
- Bc-AI on behalf of Smilyai-Labs
  • 2 replies
Β·
Recently active users