huggingface-projects/Deep-RL-Course-Certification
Viewer
•
Updated
•
1.49k
•
307
•
14
AI & ML interests
None defined yet.
Recent Activity
View all activity
ThomasSimonini
updated
a
dataset
about 11 hours ago
giadap
authored
a
paper
1 day ago
Post
3123
🧠👁️ Can AI visualize solutions?
Humans often solve visual problems by sketching ideas in our minds. What if Vision-Language Models (VLMs) could do something similar, not by generating full images, but by using internal “mental sketches”?
That’s the idea behind Mirage, a new framework that empowers VLMs to reason using latent visual tokens. Instead of just thinking in words, Mirage mixes in abstract visual representations that help the model solve complex tasks.
These aren't photorealistic images. They're compact, internal representations optimized purely to support reasoning.
🔧 Mirage is trained in two phases:
1) Grounding: It learns to produce latent tokens anchored in real images.
2) Refinement: The model drops the images and learns to generate visual tokens on its own.
📈 And yes, it works!
On challenging benchmarks like Visual Spatial Planning, Jigsaw puzzles, and Spatial Attention Tasks, Mirage clearly outperforms GPT-4o and other strong baselines.
Smart sketches > empty words.
By mimicking the way humans visualize solutions, Mirage gives AI a new kind of imagination, one that’s faster, more efficient, and more human-like.
Kudos to the teams at UMass Amherst and MIT behind this exciting work.
Check the paper: Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens (2506.17218)
Humans often solve visual problems by sketching ideas in our minds. What if Vision-Language Models (VLMs) could do something similar, not by generating full images, but by using internal “mental sketches”?
That’s the idea behind Mirage, a new framework that empowers VLMs to reason using latent visual tokens. Instead of just thinking in words, Mirage mixes in abstract visual representations that help the model solve complex tasks.
These aren't photorealistic images. They're compact, internal representations optimized purely to support reasoning.
🔧 Mirage is trained in two phases:
1) Grounding: It learns to produce latent tokens anchored in real images.
2) Refinement: The model drops the images and learns to generate visual tokens on its own.
📈 And yes, it works!
On challenging benchmarks like Visual Spatial Planning, Jigsaw puzzles, and Spatial Attention Tasks, Mirage clearly outperforms GPT-4o and other strong baselines.
Smart sketches > empty words.
By mimicking the way humans visualize solutions, Mirage gives AI a new kind of imagination, one that’s faster, more efficient, and more human-like.
Kudos to the teams at UMass Amherst and MIT behind this exciting work.
Check the paper: Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens (2506.17218)
- 1 reply
Post
1434
The Chinese Open Source Heatmap is live 🔥
You can now track the companies/ research labs/ communities powering China’s open source AI movement.
zh-ai-community/model-release-heatmap-zh
Some highlights:
✨Giant Tech are investing more in open source.
-Alibaba: Full stack open ecosystem
-Tecent: Hunyuan image/video/3D
-Bytedance: Catching up fast in 2025
-Baidu: New player in open LLM
✨New players emerging post–DeepSeek moment.
-Xiaomi
-Red Note
-Bilibili
-MiniMax
-Moonshot AI
✨Startup list is shifting fast! Those who find a direction aligned with their strengths are the ones who endure.
-DeepSeek
-MiniMax
-StepFun
-Moonshot AI
-Zhipu AI
-OpenBMB
✨Research Lab & Community are making key contributions.
-BAAI
-Shanghai AI Lab
-OpenMOSS
-MAP
You can now track the companies/ research labs/ communities powering China’s open source AI movement.
zh-ai-community/model-release-heatmap-zh
Some highlights:
✨Giant Tech are investing more in open source.
-Alibaba: Full stack open ecosystem
-Tecent: Hunyuan image/video/3D
-Bytedance: Catching up fast in 2025
-Baidu: New player in open LLM
✨New players emerging post–DeepSeek moment.
-Xiaomi
-Red Note
-Bilibili
-MiniMax
-Moonshot AI
✨Startup list is shifting fast! Those who find a direction aligned with their strengths are the ones who endure.
-DeepSeek
-MiniMax
-StepFun
-Moonshot AI
-Zhipu AI
-OpenBMB
✨Research Lab & Community are making key contributions.
-BAAI
-Shanghai AI Lab
-OpenMOSS
-MAP
sergiopaniego
posted
an
update
2 days ago
Post
1438
Updated my HF Space for vibe testing smol VLMs on object detection, visual grounding, keypoint detection & counting! 👓
🆕 Compare Qwen2.5 VL 3B vs Moondream 2B side-by-side with annotated images & text outputs.
Try examples or test your own images! 🏃
📱Space: sergiopaniego/vlm_object_understanding
🆕 Compare Qwen2.5 VL 3B vs Moondream 2B side-by-side with annotated images & text outputs.
Try examples or test your own images! 🏃
📱Space: sergiopaniego/vlm_object_understanding
Post
3162
🔥 June highlights from China’s open source ecosystem.
zh-ai-community/june-2025-open-works-from-the-chinese-community-683d66c188f782dc5570ba15
✨Baidu & MiniMax both launched open foundation models
- Baidu: Ernie 4.5 ( from 0.3B -424B ) 🤯
- MiniMax: MiniMax -M1 ( Hybrid MoE reasoning model )
✨Multimodal AI is moving from fusion to full-stack reasoning: unified Any-to-Any pipelines across text, vision, audio, and 3D
- Baidu: ERNIE-4.5-VL-424B
- Moonshot AI: Kimi-VL-A3B
- Alibaba: Ovis-U1
- BAAI: Video-XL-2/OmniGen2
- AntGroup: Ming-Lite-Omni
- Chinese Academy of Science: Stream-Omni
- Bytedance: SeedVR2-3B
- Tencent: Hunyuan 3D 2.1/ SongGeneration
- FishAudio: Openaudio-s1-mini
✨Domain specific models are rapidly emerging
- Alibaba DAMO: Lingshu-7B (medical MLLM)
- BAAI: RoboBrain (Robotics)
✨ So many small models!
- OpenBMB: MiciCPM4 ( on device )
- Qwen: Embedding/Reranker (0.6B)
- Alibaba: Ovis-U1-3B
- Moonshot AI: Kimi-VL-A3B
- Bytedance: SeedVR2-3B
zh-ai-community/june-2025-open-works-from-the-chinese-community-683d66c188f782dc5570ba15
✨Baidu & MiniMax both launched open foundation models
- Baidu: Ernie 4.5 ( from 0.3B -424B ) 🤯
- MiniMax: MiniMax -M1 ( Hybrid MoE reasoning model )
✨Multimodal AI is moving from fusion to full-stack reasoning: unified Any-to-Any pipelines across text, vision, audio, and 3D
- Baidu: ERNIE-4.5-VL-424B
- Moonshot AI: Kimi-VL-A3B
- Alibaba: Ovis-U1
- BAAI: Video-XL-2/OmniGen2
- AntGroup: Ming-Lite-Omni
- Chinese Academy of Science: Stream-Omni
- Bytedance: SeedVR2-3B
- Tencent: Hunyuan 3D 2.1/ SongGeneration
- FishAudio: Openaudio-s1-mini
✨Domain specific models are rapidly emerging
- Alibaba DAMO: Lingshu-7B (medical MLLM)
- BAAI: RoboBrain (Robotics)
✨ So many small models!
- OpenBMB: MiciCPM4 ( on device )
- Qwen: Embedding/Reranker (0.6B)
- Alibaba: Ovis-U1-3B
- Moonshot AI: Kimi-VL-A3B
- Bytedance: SeedVR2-3B
Post
558
SOOOO MANY MODEL RELEASES 😍
Here's some picks from past week 🤗
> ByteDance/XVerse is a new identity preserving image generation model 🖼️
> google/gemma-3n-E4B-it, any-to-text model supported by transformers 🤗
> nvidia/llama-nemoretriever-colembed-3b-v1 two new state-of-the-art visual document retrievers 📑
> New version of Dia TTS model is up nari-labs/Dia-1.6B-0626
> Black Forest Labs releases Kontext benchmark black-forest-labs/kontext-bench
Find more here merve/releases-june-27-6864e8eb17f7e3a8b444083c
Here's some picks from past week 🤗
> ByteDance/XVerse is a new identity preserving image generation model 🖼️
> google/gemma-3n-E4B-it, any-to-text model supported by transformers 🤗
> nvidia/llama-nemoretriever-colembed-3b-v1 two new state-of-the-art visual document retrievers 📑
> New version of Dia TTS model is up nari-labs/Dia-1.6B-0626
> Black Forest Labs releases Kontext benchmark black-forest-labs/kontext-bench
Find more here merve/releases-june-27-6864e8eb17f7e3a8b444083c
Post
222
MTVCraft 🔥 Veo3 style Audio-Video model by BAAI
Model:
BAAI/MTVCraft
Demo:
BAAI/MTVCraft
✨ Text > [Speech + SFX + BGM] > Synchronized Video
✨ Built with Qwen3 + ElevenLabs + MTV
Model:
BAAI/MTVCraft
Demo:
BAAI/MTVCraft
✨ Text > [Speech + SFX + BGM] > Synchronized Video
✨ Built with Qwen3 + ElevenLabs + MTV
Post
2822
visual reasoning is now in transformers 🔥
THUDM/GLM-4.1V-9B-Thinking is just released and merged into transformers, we gave it a vibe test run 🤠
it's very good, comes with 64k context length and MIT license 😍
it supports 4k image tokens and any aspect ratio as well!
Notebook: http://colab.research.google.com/drive/1atODIiV57hOZLv16Bjzwd6fwx0yoTorj?usp=sharing
Demo: THUDM/GLM-4.1V-9B-Thinking-Demo
THUDM/GLM-4.1V-9B-Thinking is just released and merged into transformers, we gave it a vibe test run 🤠
it's very good, comes with 64k context length and MIT license 😍
it supports 4k image tokens and any aspect ratio as well!
Notebook: http://colab.research.google.com/drive/1atODIiV57hOZLv16Bjzwd6fwx0yoTorj?usp=sharing
Demo: THUDM/GLM-4.1V-9B-Thinking-Demo
Post
2211
GLM-4.1V-Thinking 🔥 New open vision reasoning model by Zhipu AI
THUDM/glm-41v-thinking-6862bbfc44593a8601c2578d
✨ 9B base & Thinking - MIT license
✨ CoT + RL with Curriculum Sampling
✨ 64k context, 4K image, any aspect ratio
✨ Support English & Chinese
✨ Outperforms GPT 4O -2024/11/20
THUDM/glm-41v-thinking-6862bbfc44593a8601c2578d
✨ 9B base & Thinking - MIT license
✨ CoT + RL with Curriculum Sampling
✨ 64k context, 4K image, any aspect ratio
✨ Support English & Chinese
✨ Outperforms GPT 4O -2024/11/20
sergiopaniego
posted
an
update
5 days ago
Post
959
📣 CALL FOR CONTRIBUTORS! 📣
Following last week’s full release of Gemma 3n, we launched a dedicated recipes repo to explore and share use cases. We already added some! 🧑🍳
Now we’re inviting the community to contribute and showcase how these models shine! ✨
Let them cook.
Check it out: https://github.com/huggingface/huggingface-gemma-recipes/issues/4
Following last week’s full release of Gemma 3n, we launched a dedicated recipes repo to explore and share use cases. We already added some! 🧑🍳
Now we’re inviting the community to contribute and showcase how these models shine! ✨
Let them cook.
Check it out: https://github.com/huggingface/huggingface-gemma-recipes/issues/4
- 1 reply
Post
2445
so many multimodal releases these days 🤠
> ERNIE-4.5-VL: new vision language MoE models by Baidu https://huggingface.co/models?search=ernie-4.5-vl
> new visual document retrievers by NVIDIA (sota on ViDoRe!) nvidia/llama-nemoretriever-colembed-3b-v1 nvidia/llama-nemoretriever-colembed-1b-v1
> Ovis-3b: new image-text in image-text out models by Alibaba ⤵️ https://huggingface.co/spaces/AIDC-AI/Ovis-U1-
> ERNIE-4.5-VL: new vision language MoE models by Baidu https://huggingface.co/models?search=ernie-4.5-vl
> new visual document retrievers by NVIDIA (sota on ViDoRe!) nvidia/llama-nemoretriever-colembed-3b-v1 nvidia/llama-nemoretriever-colembed-1b-v1
> Ovis-3b: new image-text in image-text out models by Alibaba ⤵️ https://huggingface.co/spaces/AIDC-AI/Ovis-U1-
Post
920
Pangu Pro MoE 🔥 Huawei's first open model!
Paper:
Pangu Pro MoE: Mixture of Grouped Experts for Efficient Sparsity (2505.21411)
Model:
https://gitcode.com/ascend-tribe/pangu-pro-moe-model
✨ MoGE: Mixture of Grouped Experts
✨ 16B activated params - 48 layers
✨ Trained on 15T tokens
✨ Natively optimized for Ascend hardware
Paper:
Pangu Pro MoE: Mixture of Grouped Experts for Efficient Sparsity (2505.21411)
Model:
https://gitcode.com/ascend-tribe/pangu-pro-moe-model
✨ MoGE: Mixture of Grouped Experts
✨ 16B activated params - 48 layers
✨ Trained on 15T tokens
✨ Natively optimized for Ascend hardware
- 1 reply
Post
302
Baidu kept its promise, releasing 10 open models on the very last day of June🚀 Let's meet ERNIE 4.5 🔥
baidu/ernie-45-6861cd4c9be84540645f35c9
✨ From 0.3B to 424B total params
✨ Includes 47B & 3B active param MoE models + a 0.3B dense model
✨ Apache 2.0
✨ 128K context length
✨ Text+Vision co-training with ViT & UPO
baidu/ernie-45-6861cd4c9be84540645f35c9
✨ From 0.3B to 424B total params
✨ Includes 47B & 3B active param MoE models + a 0.3B dense model
✨ Apache 2.0
✨ 128K context length
✨ Text+Vision co-training with ViT & UPO
Add video example
🔥
1
#2 opened 8 days ago
by
sergiopaniego
Post
3063
Hunyuan-A13B 🔥 New MoE LLM by TencentHunyuan
tencent/Hunyuan-A13B-Instruct
✨80B total / 13B active params
✨256K context window
✨Dual-mode reasoning: fast & slow thinking
✨Efficient inference (GQA + quantization)
tencent/Hunyuan-A13B-Instruct
✨80B total / 13B active params
✨256K context window
✨Dual-mode reasoning: fast & slow thinking
✨Efficient inference (GQA + quantization)
hysts
updated
a
Space
8 days ago
hysts
published
a
Space
9 days ago
Post
559
Dataset Viewer for PDFs just landed on Hugging Face 📖🤗 you can now preview all the PDFs easier than before!
on top of this, there's PdfFolder format to load the PDF datasets quicker 💨
> to use it, your dataset should follow a directory format like folder/train/doc1.pdf, folder/train/doc1.pdf
> if you want to include bounding boxes, labels etc. you can keep them in a metadata.csv file in the same folder 🤝
read document dataset docs https://huggingface.co/docs/datasets/main/en/document_dataset
check all the document datasets here https://huggingface.co/datasets?modality=modality:document&sort=trending 📖
on top of this, there's PdfFolder format to load the PDF datasets quicker 💨
> to use it, your dataset should follow a directory format like folder/train/doc1.pdf, folder/train/doc1.pdf
> if you want to include bounding boxes, labels etc. you can keep them in a metadata.csv file in the same folder 🤝
read document dataset docs https://huggingface.co/docs/datasets/main/en/document_dataset
check all the document datasets here https://huggingface.co/datasets?modality=modality:document&sort=trending 📖
- 1 reply
freddyaboulton
posted
an
update
10 days ago