Hugging Face's logo

Hugging Face FineVideo

Enterprise
community
Activity Feed

AI & ML interests

None defined yet.

Recent Activity

sayakpaul  authored a paper 16 days ago
Exploring the Deep Fusion of Large Language Models and Diffusion Transformers for Text-to-Image Synthesis
lewtun  authored a paper 20 days ago
Kimina-Prover Preview: Towards Large Formal Reasoning Models with Reinforcement Learning
sayakpaul  authored a paper about 1 month ago
From Reflection to Perfection: Scaling Inference-Time Optimization for Text-to-Image Diffusion Models via Reflection Tuning
View all activity

HuggingFaceFV's activity

merve 
posted an update 2 days ago
view post
Post
1457
HOT: MiMo-VL new 7B vision LMs by Xiaomi surpassing gpt-4o (Mar), competitive in GUI agentic + reasoning tasks ❤️‍🔥 XiaomiMiMo/mimo-vl-68382ccacc7c2875500cd212

not only that, but also MIT license & usable with transformers 🔥
merve 
posted an update 3 days ago
view post
Post
2561
introducing: VLM vibe eval 🪭 visionLMsftw/VLMVibeEval

vision LMs are saturated over benchmarks, so we built vibe eval 💬

> compare different models with refreshed in-the-wild examples in different categories 🤠
> submit your favorite model for eval
no numbers -- just vibes!
merve 
posted an update 5 days ago
view post
Post
2464
emerging trend: models that can understand image + text and generate image + text

don't miss out ⤵️
> MMaDA: single 8B diffusion model aligned with CoT (reasoning!) + UniGRPO Gen-Verse/MMaDA
> BAGEL: 7B MoT model based on Qwen2.5, SigLIP-so-400M, Flux VAE ByteDance-Seed/BAGEL
both by ByteDance! 😱

I keep track of all any input → any output models here merve/any-to-any-models-6822042ee8eb7fb5e38f9b62
  • 1 reply
·
m-ric 
posted an update 6 days ago
view post
Post
2384
A new research paper from KAIST builds on smolagents to push boundaries of distillation 🥳
➡️ "Distilling LLM Agent into Small Models with Retrieval and Code Tools" teaches that, when trying to distil reasoning capability from a strong LLM ("teacher") into a smaller one ("student"), it's much better to use Agent traces than CoT traces.

Advantages are:
1. Improved generalization
Intuitively, this is because your agent can encounter more "surprising" results by interacting with its environment : for example, a web research called by the LLM teacher in agent mode can bring results that the LLM teacher would not have generated in CoT.

2. Reduce hallucinations
The trace won't hallucinate tool call outputs!

Thank you @akseljoonas for mentioning this paper!
merve 
posted an update 6 days ago
view post
Post
3058
what happened in open AI past week? so many vision LM & omni releases 🔥 merve/releases-23-may-68343cb970bbc359f9b5fb05

multimodal 💬🖼️
> new moondream (VLM) is out: it's 4-bit quantized (with QAT) version of moondream-2b, runs on 2.5GB VRAM at 184 tps with only 0.6% drop in accuracy (OS) 🌚
> ByteDance released BAGEL-7B, an omni model that understands and generates both image + text. they also released Dolphin, a document parsing VLM 🐬 (OS)
> Google DeepMind dropped MedGemma in I/O, VLM that can interpret medical scans, and Gemma 3n, an omni model with competitive LLM performance

> MMaDa is a new 8B diffusion language model that can generate image and text



LLMs
> Mistral released Devstral, a 24B coding assistant (OS) 👩🏻‍💻
> Fairy R1-32B is a new reasoning model -- distilled version of DeepSeek-R1-Distill-Qwen-32B (OS)
> NVIDIA released ACEReason-Nemotron-14B, new 14B math and code reasoning model
> sarvam-m is a new Indic LM with hybrid thinking mode, based on Mistral Small (OS)
> samhitika-0.0.1 is a new Sanskrit corpus (BookCorpus translated with Gemma3-27B)

image generation 🎨
> MTVCrafter is a new human motion animation generator
  • 1 reply
·
merve 
posted an update 10 days ago
view post
Post
2551
Google released MedGemma on I/O'25 👏 google/medgemma-release-680aade845f90bec6a3f60c4

> 4B and 27B instruction fine-tuned vision LMs and a 4B pre-trained vision LM for medicine
> available with transformers from the get-go 🤗

they also released a cool demo for scan reading ➡️ google/rad_explain

use with transformers ⤵️
  • 1 reply
·
merve 
posted an update 10 days ago
view post
Post
3067
Bu post'u çevirebilirsiniz 🤗💗
·
merve 
posted an update 10 days ago
view post
Post
2348
tis the year of any-to-any/omni models 🤠
ByteDance-Seed/BAGEL-7B-MoT 7B native multimodal model that understands and generates both image + text

it outperforms leading VLMs like Qwen 2.5-VL 👏 and has Apache 2.0 license 😱
sayakpaul 
posted an update 10 days ago
view post
Post
2296
Diffusers supports a good variety of quantization backends. It can be challenging to navigate through them, given the complex nature of diffusion pipelines in general.

So, @derekl35 set out to write a comprehensive guide that puts users in the front seat. Explore the different backends we support, learn the trade-offs they offer, and finally, check out the cool space we built that lets you compare quantization results.

Give it a go here:
https://lnkd.in/gf8Pi4-2
sayakpaul 
posted an update 12 days ago
view post
Post
1656
Despite the emergence of combining LLM and DiT architectures for T2I synthesis, its design remains severely understudied.

This was done long ago and got into CVPR25 -- super excited to finally share it now, along with the data and code ♥️

We explore several architectural choices that affect this design. We provide an open & reproducible training recipe that works at scale.

Works like Playground v3 have already explored a deep fusion between an LLM and a DiT, sharing their representations through layerwise attention. They exhibit excellent performance on T2I.

Despite its compelling results and other performance virtues, it remains unexplored, which is what we want to improve in our work. Specifically, we take a pre-trained LLM (Gemma-2B) and trainable DiT, and set out to explore what makes a "good deep fusion" between the two for T2I.

We explore several key questions in the work, such as:

Q1: How should we do attention? We considered several alternatives. PixArt-Alpha like attention (cross-attention) is very promising.
Q2: Should we incorporate additional text modulation?
Q3: Can we eliminate timestep conditioning?
Q4: How do we do positional encodings?
Q5: Do instruction-tuned LLMs help deep fusion?
Q6: Would using a decoder LLM from a multimodal model be helpful?
Q7: Does using a better variant of Gemma help?

Based on the above findings, we arrive at FuseDiT with the following components on top of the base architecture from the findings of our experiments.

* No AdaLN-Zero modules
* 1D + 2D-RoPE
* Gemma 2 2B, adjusting DiT configurations accordingly

We trained FuseDiT on a mixture from CC12M, JourneyDB, & SA (~26M image-text pairs) for 800 steps. While not the best model, it's encouraging to develop something in a guided manner using open datasets.

To know more (code, models, all are available), please check out the paper:
https://lnkd.in/gg6qyqZX.
merve 
posted an update 12 days ago
view post
Post
1706
NVIDIA released new vision reasoning model for robotics: Cosmos-Reason1-7B 🤖 nvidia/cosmos-reason1-67c9e926206426008f1da1b7

> first reasoning model for robotics
> based on Qwen 2.5-VL-7B, use with Hugging Face transformers or vLLM 🤗
> comes with SFT & alignment datasets and a new benchmark 👏
merve 
posted an update 13 days ago
view post
Post
2565
It was the week of video generation at @huggingface , on top of many new LLMs, VLMs and more!
Let’s have a wrap 🌯 merve/may-16-releases-682aeed23b97eb0fe965345c

LLMs 💬
> Alibaba Qwen released WorldPM-72B, new World Preference Model trained with 15M preference samples (OS)
> II-Medical-8B, new LLM for medical reasoning that comes in 8B by Intelligent-Internet
> TRAIL is a new dataset by Patronus for trace error reasoning for agents (OS)

Multimodal 🖼️💬
> Salesforce Research released BLIP3o, a new any-to-any model with image-text input and image-text output 💬it’s based on an image encoder, a text decoder and a DiT, and comes in 8B
> They also released pre-training and fine-tuning datasets
> MMMG is a multimodal generation benchmark for image, audio, text (interleaved)

Image Generation ⏯️
> Alibaba Wan-AI released Wan2.1-VACE, video foundation model for image and text to video, video-to-audio and more tasks, comes in 1.3B and 14B (OS)
> ZuluVision released MoviiGen1.1, new cinematic video generation model based on Wan 2.1 14B (OS)
> multimodalart released isometric-skeumorphic-3d-bnb, an isometric 3D asset generator (like AirBnB assets) based on Flux
> LTX-Video-0.9.7-distilled is a new real-time video generation (text and image to video) model by Lightricks
> Hidream_t2i_human_preference is a new text-to-image preference dataset by Rapidata with 195k human responses from 38k annotators

Audio 🗣️
> stabilityai released stable-audio-open-small new text-to-audio model
> TEN-framework released ten-vad, voice activity detection model (OS)
sayakpaul 
authored a paper 16 days ago
merve 
posted an update 16 days ago
m-ric 
posted an update 19 days ago
view post
Post
2593
𝗔𝗯𝘀𝗼𝗹𝘂𝘁𝗲 𝗭𝗲𝗿𝗼: 𝗟𝗟𝗠𝘀 𝗰𝗮𝗻 𝘁𝗿𝗮𝗶𝗻 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗮𝗻𝘆 𝗲𝘅𝘁𝗲𝗿𝗻𝗮𝗹 𝗱𝗮𝘁𝗮 🤯

Has the "data wall" just been breached?

Recent RL paradigms often relied on a set of questions an answers that needs to be manually curated. Researchers from Tsinghua University went like "why though".

🤔 Indeed, why learn from question designed by a human teacher, when the model can start from their base knowledge and learn by experimenting in a code environment, proposing coding tasks themselves and trying to solve them?

Thus they created “Absolute Zero Reasoning” (AZR), an approach that removes any need for human curated data.

🎭 𝗗𝘂𝗮𝗹 𝗿𝗼𝗹𝗲𝘀:
‣ Proposer: Generates challenging but solvable coding tasks
‣ Solver: Attempts to solve those self-proposed tasks

🧪 𝗧𝗵𝗿𝗲𝗲 𝘁𝗮𝘀𝗸 𝘁𝘆𝗽𝗲𝘀: all types are defined as triplets of program, input and output
‣ Deduction: Give model an input and program, it must deduce the output
‣ Abduction: Give model an program and output, it must find the input that gave said output
‣ Induction: Synthesize a program from input/output pairs
Btw this reminded me of my long-forgotten philosophy classes: Aristotle was more on the induction side, learning from real-world analogies, while Plato was more on the deduction side, trying to progress quite far with just one input and his reasoning.

📊 𝗥𝗲𝘀𝘂𝗹𝘁𝘀:
‣ AZR post-training creates a nice improvement on known models like Qwen2.5-7B
‣ Shows strong cross-domain transfer: coding ↔️ math reasoning

🧐 𝗢𝘁𝗵𝗲𝗿 𝗳𝗶𝗻𝗱𝗶𝗻𝗴𝘀:
‣ Having a better base performance (general or code specific) amplify the gains from Absolute Zero Reasoning
‣ Researchers warn about "Uh-oh moments" (winking to the "aha moments" of DeepSeek) where the model generates concerning goals like "make an extremely convoluted code to outsmart all these humans": so supervision is still needed!

Paper here: Absolute Zero: Reinforced Self-play Reasoning with Zero Data (2505.03335)
merve 
posted an update 20 days ago
view post
Post
5019
VLMS 2025 UPDATE 🔥

We just shipped a blog on everything latest on vision language models, including
🤖 GUI agents, agentic VLMs, omni models
📑 multimodal RAG
⏯️ video LMs
🤏🏻 smol models
..and more! https://huggingface.co/blog/vlms-2025
  • 1 reply
·
lewtun 
authored a paper 20 days ago
m-ric 
posted an update 23 days ago
view post
Post
4391
I've made an open version of Google's NotebookLM, and it shows the superiority of the open source tech task! 💪

The app's workflow is simple. Given a source PDF or URL, it extracts the content from it, then tasks Meta's Llama 3.3-70B with writing the podcast script, with a good prompt crafted by @gabrielchua ("two hosts, with lively discussion, fun notes, insightful question etc.")
Then it hands off the text-to-speech conversion to Kokoro-82M, and there you go, you have two hosts discussion any article.

The generation is nearly instant, because:
> Llama 3.3 70B is running at 1,000 tokens/seconds with Cerebras inference
> The audio is generated in streaming mode by the tiny (yet powerful) Kokoro, generating voices faster than real-time.

And the audio generation runs for free on Zero GPUs, hosted by HF on H200s.

Overall, open source solutions rival the quality of closed-source solutions at close to no cost!

Try it here 👉👉 m-ric/open-notebooklm
·
merve 
posted an update 26 days ago
view post
Post
5061
A ton of impactful models and datasets in open AI past week, let's summarize the best 🤩 merve/releases-apr-21-and-may-2-6819dcc84da4190620f448a3

💬 Qwen made it rain! They released Qwen3: new dense and MoE models ranging from 0.6B to 235B 🤯 as well as Qwen2.5-Omni, any-to-any model in 3B and 7B!
> Microsoft AI released Phi4 reasoning models (that also come in mini and plus sizes)
> NVIDIA released new CoT reasoning datasets
🖼️ > ByteDance released UI-TARS-1.5, native multimodal UI parsing agentic model
> Meta released EdgeTAM, an on-device object tracking model (SAM2 variant)
🗣️ NVIDIA released parakeet-tdt-0.6b-v2, a smol 600M automatic speech recognition model
> Nari released Dia, a 1.6B text-to-speech model
> Moonshot AI released Kimi Audio, a new audio understanding, generation, conversation model
👩🏻‍💻 JetBrains released Melium models in base and SFT for coding
> Tesslate released UIGEN-T2-7B, a new text-to-frontend-code model 🤩
merve 
posted an update 27 days ago
view post
Post
6565
A real-time object detector much faster and accurate than YOLO with Apache 2.0 license just landed to Hugging Face transformers 🔥

D-FINE is the sota real-time object detector that runs on T4 (free Colab) 🤩

> Collection with all checkpoints and demo ustc-community/d-fine-68109b427cbe6ee36b4e7352

Notebooks:
> Tracking https://github.com/qubvel/transformers-notebooks/blob/main/notebooks/DFine_tracking.ipynb
> Inference https://github.com/qubvel/transformers-notebooks/blob/main/notebooks/DFine_inference.ipynb
> Fine-tuning https://github.com/qubvel/transformers-notebooks/blob/main/notebooks/DFine_finetune_on_a_custom_dataset.ipynb
h/t @vladislavbro @qubvel-hf @ariG23498 and the authors of the paper 🎩

Regular object detectors attempt to predict bounding boxes in (x, y, w, h) pixel perfect coordinates, which is very rigid and hard to solve 🥲☹️



D-FINE formulates object detection as a distribution for bounding box coordinates, refines them iteratively, and it's more accurate 🤩

Another core idea behind this model is Global Optimal Localization Self-Distillation ⤵️

this model uses final layer's distribution output (sort of like a teacher) to distill to earlier layers to make early layers more performant.

  • 2 replies
·