Hugging Face's logo
Activity Feed

AI & ML interests

None defined yet.
alvarobarttΒ 
posted an update 2 months ago
view post
Post
518
Open agents on AWS SageMaker AI with open models from the Hugging Face Hub!

> Deploy an open model from the Hugging Face Hub on SageMaker AI
> Connect the deployed model to Strands Agents
> Add built-in and custom tools for tool calling
> Expose external capabilities through MCP integration
> Bonus: talk to your agent and visualize traces with Gradio

https://alvarobartt.com/agents-on-aws-sagemaker
alvarobarttΒ 
posted an update 3 months ago
view post
Post
3367
Latest hf-mem release added a breakdown of Mixture-of-Experts (MoE) memory usage!

TL; DR MoEs can be misleading to reason about from active parameters alone, since each token only activates a subset of experts, while the serving setup still needs to account for the full resident memory footprint.

🧠 hf-mem now splits MoE memory into base model weights, routed experts, and KV cache
πŸ—οΈ Dense models usually load and use most weights every forward pass, while MoEs load many experts but only route each token to a few of them
⚑ Active params isn't the same as memory footprint, especially for sparse architectures
πŸ“¦ Runtime memory is about what is used per request/token, while loading memory also includes the expert weights that need to be resident
πŸ“š KV cache can still dominate depending on context length, batch size, and concurrency
πŸ”€ Expert Parallelism (EP) helps shard experts across accelerators when expert weights dominate
πŸš€ Data Parallelism (DP) + EP is often a good fit for throughput-oriented MoE serving

Check the repository at https://github.com/alvarobartt/hf-mem
mishigΒ 
posted an update 4 months ago
alvarobarttΒ 
posted an update 5 months ago
view post
Post
3769
Learn how to deploy Microsoft Research VibeVoice ASR on Microsoft Azure Foundry with Hugging Face to generate rich audio transcriptions with Who, When, and What! πŸ’₯

> πŸ•’ 60-minute single-pass processing, no chunking or stitching
> πŸ‘€ Customized hotwords to guide recognition on domain-specific content
> πŸ“ Rich transcription: joint ASR + diarization + timestamping in one pass
> 🌍 50+ languages with automatic detection and code-switching support
> πŸ€— Deployed on Microsoft Foundry via an OpenAI-compatible Chat Completions API

https://huggingface.co/docs/microsoft-azure/foundry/examples/deploy-vibevoice-asr
mryabΒ 
submitted a paper to Daily Papers 5 months ago
alvarobarttΒ 
posted an update 6 months ago
view post
Post
3304
πŸ’₯ hf-mem v0.4.1 now also estimates KV cache memory requirements for any context length and batch size with the --experimental flag!

uvx hf-mem --model-id ... --experimental will automatically pull the required information from the Hugging Face Hub to include the KV cache estimation, when applicable.

πŸ’‘ Alternatively, you can also set the --max-model-len, --batch-size and --kv-cache-dtype arguments (Γ  la vLLM) manually if preferred.
  • 1 reply
Β·
merveΒ 
posted an update 10 months ago
view post
Post
14196
deepseek-ai/DeepSeek-OCR is out! πŸ”₯ my take ‡️
> pretty insane it can parse and re-render charts in HTML
> it uses CLIP and SAM features concatenated, so better grounding
> very efficient per vision tokens/performance ratio
> covers 100 languages
  • 5 replies
Β·
jfchiΒ 
authored 6 papers 10 months ago
maheshp9Β 
authored 3 papers 10 months ago
merveΒ 
posted an update 11 months ago
view post
Post
7090
large AI labs open-sourced a ton of models last week πŸ”₯
here's few picks, find even more here merve/sep-16-releases-68d13ea4c547f02f95842f05 🀝
> IBM released a new Docling model with 258M params based on Granite (A2.0) πŸ“ ibm-granite/granite-docling-258M
> Xiaomi released 7B audio LM with base and instruct variants (MIT) XiaomiMiMo/mimo-audio-68cc7202692c27dae881cce0
> DecartAI released Lucy Edit, open Nano Banana 🍌 (NC) decart-ai/Lucy-Edit-Dev
> OpenGVLab released a family of agentic computer use models (3B/7B/32B) with the dataset πŸ’» OpenGVLab/scalecua-68c912cf56f7ff4c8e034003
> Meituan Longcat released thinking version of LongCat-Flash πŸ’­ meituan-longcat/LongCat-Flash-Thinking
  • 2 replies
Β·
merveΒ 
posted an update 11 months ago
view post
Post
3589
IBM just released small swiss army knife for the document models: granite-docling-258M on Hugging Face πŸ”₯

> not only a document converter but also can do document question answering, understand multiple languages 🀯
> best part: released with Apache 2.0 license πŸ‘ use it with your commercial projects!
> it supports transformers, vLLM and MLX from the get-go! πŸ€—
> built on SigLIP2 & granite-165M

model: ibm-granite/granite-docling-258M
demo: ibm-granite/granite-docling-258m-demo πŸ’—
lysandreΒ 
posted an update 11 months ago
view post
Post
9455
We're kick-starting the process of Transformers v5, with @ArthurZ and @cyrilvallez !

v5 should be significant: we're using it as a milestone for performance optimizations, saner defaults, and a much cleaner code base worthy of 2025.

Fun fact: v4.0.0-rc-1 came out on Nov 19, 2020, nearly five years ago!
  • 6 replies
Β·
merveΒ 
posted an update 11 months ago
view post
Post
1322
a ton of image/video generation models and LLMs from big labs πŸ”₯

> Meta released facebook/mobilellm-r1-68c4597b104fac45f28f448e, smol LLMs for on-device use πŸ’¬
> Tencent released tencent/SRPO, high res image generation model and tencent/POINTS-Reader, cutting edge OCR πŸ“
> ByteDance released bytedance-research/HuMo, video generation from any input ⏯️

find more models, datasets, demos here merve/sep-11-releases-68c7dbfa26bea8cd921fa0ac