Join the conversation
Join the community of Machine Learners and AI enthusiasts.
Sign UpAll HF Hub posts
hypotheticalย
posted an update 1 day ago
Post
1937
The smallest and the highest quality in the world Gemma4 E2B and E4B models! 7x compression! From 9.3GB -> 1.4GB!
TheStageAI/gemma-4-E2B-it
TheStageAI/gemma-4-E4B-it
TheStageAI/gemma-4-E2B-it
TheStageAI/gemma-4-E4B-it
- 1 reply
sergiopaniegoย
posted an update 3 days ago
Post
6108
new banger blog alert ๐จ
@ariG23498 is starting a blog series about profiling in pytorch and part 1 just dropped
takes you from the simplest scenario to actually knowing what your gpu is doing. if you have never opened a profiler trace this is where you start
covers
find it here: https://huggingface.co/blog/torch-profiler
@ariG23498 is starting a blog series about profiling in pytorch and part 1 just dropped
takes you from the simplest scenario to actually knowing what your gpu is doing. if you have never opened a profiler trace this is where you start
covers
torch.profiler from scratch. reading tables and traces, overhead bound vs compute bound, the full dispatch chain from python to gpu kernels, and what torch.compile is actually fusing under the hoodfind it here: https://huggingface.co/blog/torch-profiler
- 1 reply
Post
1883
Hugging Face MCP Server v0.3.17
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
SEP-2640 "Skills Over MCP" support added (early access)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
SEP-2640 "Skills Over MCP" support added (early access)
- 1 reply
Post
1929
Qwen Image Edit 2511 Fast + LoRA โก
ovi054/Qwen-Image-Edit-2511-LoRA
QIE-2511 is an image editing model with integrated LoRA capabilities. You can add any custom LoRA to generate and edit images within this Space.
๐ Try it now: ovi054/Qwen-Image-Edit-2511-LoRA
ovi054/Qwen-Image-Edit-2511-LoRA
QIE-2511 is an image editing model with integrated LoRA capabilities. You can add any custom LoRA to generate and edit images within this Space.
๐ Try it now: ovi054/Qwen-Image-Edit-2511-LoRA
sergiopaniegoย
posted an update 2 days ago
Post
1869
most multi-turn RL loops have a silent bug: you decode the model's output to detect tool calls, then re-tokenize the conversation for the next turn. BPE isn't invertible, so decode then re-encode can land on different ids. gradient ends up on tokens the model never sampled. no crash, just quietly wrong math and broken training
@qgallouedec wrote a super educational blog on MITO (message-in, token-out) vs TITO (token-in, token-out) and how you might fix the problem above
go read it ๐ค
https://qgallouedec-tito.hf.space/
@qgallouedec wrote a super educational blog on MITO (message-in, token-out) vs TITO (token-in, token-out) and how you might fix the problem above
go read it ๐ค
https://qgallouedec-tito.hf.space/
Post
693
New blog post!
An introduction to a little-known but highly effective model reduction method: ๐ง๐ฟ๐ถ๐บ๐บ๐ถ๐ป๐ดโ๏ธ
We show how to reduce model size (we went up to 87.24% reduction) while preserving its performance.
We applied this technique to 16 different model families across several modalities to illustrate that it works on any architecture (as long as the embedding layer is the last one of the model) and on any modality involving text.
From these 16 families, we generated over ๐ฑ,๐ฑ๐ฌ๐ฌ ๐บ๐ผ๐ป๐ผ๐น๐ถ๐ป๐ด๐๐ฎ๐น ๐บ๐ผ๐ฑ๐ฒ๐น๐ ๐ถ๐ป ๐ญ๐ฎ๐ฐ ๐ฑ๐ถ๐ณ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ ๐น๐ฎ๐ป๐ด๐๐ฎ๐ด๐ฒ๐ ๐
Key takeaways from our experiments:
1๏ธโฃ Trimming does not require a GPU. Our models were obtained on a CPU.
2๏ธโฃ This method scales up to at least 4B parameters (we did not test beyond that).
3๏ธโฃ Trimmed model is smaller than the original while preserving its performance. If you observe a slight performance drop, just fine-tuned to recover or even surpass the original performance.
4๏ธโฃ For an equivalent compute budget, it is better to trim then fine-tune rather than fine-tuning the original model. Since the model is smaller, you can run more epochs/show more data and get in fine a better model than the original.
5๏ธโฃ Trimming is a competitive alternative to distillation and quantization. E.g. we obtained our alternative to DistilBERT in 9 minutes on CPU vs. 90 hours of GPU for the latter.
6๏ธโฃ Trimming could generate reasoning traces in the language of the trimmed model. This could be an alternative to generating traces in English and then translating them into the desired language.
And many other things (such as how much data are needed, the impact of the database used, the order in which it should be done, etc.) are available in the blogpost!
Blogpost: https://huggingface.co/blog/lbourdois/introduction-to-trimming
Models: alphaedge-ai/Trimming_models_search
An introduction to a little-known but highly effective model reduction method: ๐ง๐ฟ๐ถ๐บ๐บ๐ถ๐ป๐ดโ๏ธ
We show how to reduce model size (we went up to 87.24% reduction) while preserving its performance.
We applied this technique to 16 different model families across several modalities to illustrate that it works on any architecture (as long as the embedding layer is the last one of the model) and on any modality involving text.
From these 16 families, we generated over ๐ฑ,๐ฑ๐ฌ๐ฌ ๐บ๐ผ๐ป๐ผ๐น๐ถ๐ป๐ด๐๐ฎ๐น ๐บ๐ผ๐ฑ๐ฒ๐น๐ ๐ถ๐ป ๐ญ๐ฎ๐ฐ ๐ฑ๐ถ๐ณ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ ๐น๐ฎ๐ป๐ด๐๐ฎ๐ด๐ฒ๐ ๐
Key takeaways from our experiments:
1๏ธโฃ Trimming does not require a GPU. Our models were obtained on a CPU.
2๏ธโฃ This method scales up to at least 4B parameters (we did not test beyond that).
3๏ธโฃ Trimmed model is smaller than the original while preserving its performance. If you observe a slight performance drop, just fine-tuned to recover or even surpass the original performance.
4๏ธโฃ For an equivalent compute budget, it is better to trim then fine-tune rather than fine-tuning the original model. Since the model is smaller, you can run more epochs/show more data and get in fine a better model than the original.
5๏ธโฃ Trimming is a competitive alternative to distillation and quantization. E.g. we obtained our alternative to DistilBERT in 9 minutes on CPU vs. 90 hours of GPU for the latter.
6๏ธโฃ Trimming could generate reasoning traces in the language of the trimmed model. This could be an alternative to generating traces in English and then translating them into the desired language.
And many other things (such as how much data are needed, the impact of the database used, the order in which it should be done, etc.) are available in the blogpost!
Blogpost: https://huggingface.co/blog/lbourdois/introduction-to-trimming
Models: alphaedge-ai/Trimming_models_search
- 4 replies
RakshitAralimattiย
posted an update 3 days ago
Post
442
Reading engineering and research blogs from OpenAI, Anthropic, DeepMind, Meta and others has genuinely leveled up my understanding of AI systems and helped me in my day-to-day work. But keeping track of 20+ sites manually is a pain.
So I built AI Blogs Tracker โ a Streamlit app that scrapes the actual blog listing pages (not search) of 20+ top AI companies and surfaces titles, dates, and links in one clean feed. Filter by source, by date, star posts to a reading list, or add your own custom sources.
One click. ~30 seconds. Everything in one place.
๐ GitHub link - https://github.com/rakshit2020/Tech-Blogs-Tracker-of-Top-AI-Companies-Agent
So I built AI Blogs Tracker โ a Streamlit app that scrapes the actual blog listing pages (not search) of 20+ top AI companies and surfaces titles, dates, and links in one clean feed. Filter by source, by date, star posts to a reading list, or add your own custom sources.
One click. ~30 seconds. Everything in one place.
๐ GitHub link - https://github.com/rakshit2020/Tech-Blogs-Tracker-of-Top-AI-Companies-Agent
- 1 reply
AxionLab-officialย
posted an update about 15 hours ago
Post
555
We're happy to announce that we released a Reasoning tuned version of Supra-50M!
SupraLabs/Supra-50M-Reasoning
SupraLabs/Supra-50M-Reasoning
salma-remyxย
posted an update about 16 hours ago
Post
847
In that benchmark comparison, do you even have the sample size to distinguish two models, or are you making decisions based on statistical noise?
"Resolution Diagnostics for Paired LLM Evaluation" offers a simple check: a per-pair resolution ratio q = N/N* that flags when a displayed ranking sits below the resolution floor regardless of p-value.
arXiv: https://arxiv.org/abs/2605.30315v1
Outrider automatically matched this paper to our fork of lm-evaluation-harness and opened a PR implementing the diagnostic.
Configure the action to find new methods tailored to your repo: https://github.com/remyxai/outrider
"Resolution Diagnostics for Paired LLM Evaluation" offers a simple check: a per-pair resolution ratio q = N/N* that flags when a displayed ranking sits below the resolution floor regardless of p-value.
arXiv: https://arxiv.org/abs/2605.30315v1
Outrider automatically matched this paper to our fork of lm-evaluation-harness and opened a PR implementing the diagnostic.
Configure the action to find new methods tailored to your repo: https://github.com/remyxai/outrider
