271 28 62

Joao Gante

joaogante

https://github.com/gante

AI & ML interests

None yet

Recent Activity

upvoted an article 1 day ago

Tiny Agents in Python: a MCP-powered agent in ~70 lines of code

upvoted a changelog 1 day ago

Filter by MCP compatibility available in HF Spaces

upvoted a changelog 1 day ago

AI-generated Abstract summaries on Hugging Face Papers

View all activity

Organizations

joaogante's activity

upvoted an article 1 day ago

Article

Tiny Agents in Python: a MCP-powered agent in ~70 lines of code

and 3 others •

1 day ago

• 42

upvoted 2 changelogs 1 day ago

Changelog

Filter by MCP compatibility available in HF Spaces

3 days ago

• 57

Changelog

AI-generated Abstract summaries on Hugging Face Papers

2 days ago

• 49

updated a model 1 day ago

transformers-community/sink_cache

Updated 1 day ago

published a model 2 days ago

transformers-community/sink_cache

Updated 1 day ago

posted an update 3 days ago

Post

354

Let's go! Custom generation code has landed in transformers 🚀

Have you designed a new cool KV cache? Maybe you're comparing new test-time compute ideas you've been researching? Have you found a way to do diffusion with existing models? You can now easily share your findings with the community with custom generation code, sharing the well-known generate interface 🤓

In a nutshell, we have expanded the support of custom modeling code on the Hub with *model-agnostic* custom generation code. Write for one model, reuse with any model -- hopefully, this will democratize access to new generation ideas 🫡

As a creator, you gain the ability to get your ideas in transformers with minimal effort. You'll also have access to all Hub features: a landing page for your creation, discussions, usage metrics, ... 🤓

💎 Resources 💎
- docs: https://huggingface.co/docs/transformers/generation_strategies#custom-decoding-methods
- minimal example: transformers-community/custom_generate_example
- discussion: transformers-community/support#10