Post
133
Lily 2 models will be private in the next 2 days, Lily 2 will move to Lily AI website with 🤗 Hugging Face!
Sorry of the inconvenience caused this!
Sorry of the inconvenience caused this!
Blog-explorers
AI & ML interests
None defined yet.
Recent Activity
View all activity
blog-explorers's activity
kulia-moon
posted
an
update
about 12 hours ago
Post
2239
Meet our new agentic model : 𝗗𝗲𝘃𝘀𝘁𝗿𝗮𝗹
Devstral is an open-source LLM built software engineering tasks built under a collaboration between Mistral AI and All Hands AI 🙌.
𝗞𝗲𝘆 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀 :
• 🤖 𝗔𝗴𝗲𝗻𝘁𝘀 : perfect for Agentic coding
• 🍃 𝗹𝗶𝗴𝗵𝘁𝘄𝗲𝗶𝗴𝗵𝘁: Devstral is a 𝟮𝟰𝗕 parameter based on Mistral small.
• ©️ 𝗔𝗽𝗮𝗰𝗵𝗲 𝟮.𝟬, meaning fully open-source !
• 📄 A 𝟭𝟮𝟴𝗸 context window.
📚Blog : https://mistral.ai/news/devstral
⚡API : The model is also available on our API under the name 𝗱𝗲𝘃𝘀𝘁𝗿𝗮𝗹-𝘀𝗺𝗮𝗹𝗹-𝟮𝟱𝟬𝟱
🤗 repo : mistralai/Devstral-Small-2505
Can't wait to see what you will build with it !
Devstral is an open-source LLM built software engineering tasks built under a collaboration between Mistral AI and All Hands AI 🙌.
𝗞𝗲𝘆 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀 :
• 🤖 𝗔𝗴𝗲𝗻𝘁𝘀 : perfect for Agentic coding
• 🍃 𝗹𝗶𝗴𝗵𝘁𝘄𝗲𝗶𝗴𝗵𝘁: Devstral is a 𝟮𝟰𝗕 parameter based on Mistral small.
• ©️ 𝗔𝗽𝗮𝗰𝗵𝗲 𝟮.𝟬, meaning fully open-source !
• 📄 A 𝟭𝟮𝟴𝗸 context window.
📚Blog : https://mistral.ai/news/devstral
⚡API : The model is also available on our API under the name 𝗱𝗲𝘃𝘀𝘁𝗿𝗮𝗹-𝘀𝗺𝗮𝗹𝗹-𝟮𝟱𝟬𝟱
🤗 repo : mistralai/Devstral-Small-2505
Can't wait to see what you will build with it !
- 1 reply
KennyUTC
authored
a
paper
3 days ago
[Support] Community Articles
🚀
🤝
1
85
#5 opened about 1 year ago
by
victor
kulia-moon
posted
an
update
9 days ago
hesamation
posted
an
update
10 days ago
Post
3171
60+ Generative AI projects for your resume. grind this GitHub repo if you want to level up:
> LLM fine-tuning and applications
> advanced RAG apps
> Agentic AI projects
> MCP and A2A (new)
GitHub: https://github.com/aishwaryanr/awesome-generative-ai-guide/blob/main/resources/60_ai_projects.md
> LLM fine-tuning and applications
> advanced RAG apps
> Agentic AI projects
> MCP and A2A (new)
GitHub: https://github.com/aishwaryanr/awesome-generative-ai-guide/blob/main/resources/60_ai_projects.md
kulia-moon
posted
an
update
11 days ago
Post
1885
🦢 Lily 2.0 is coming! Currently testing Lily 2.0 Preview models, Using Lily 2.x models will:
- Allow gated user
- Full Datasets (maybe, 1M?)
- 2.x will train longer, smarter, and more like a man/lady (Premium)
- Using famous models like Qwen or Llama etc, DeepSeek. we will do our best!
Coming soon 2.x at kulia-moon/lily-2x-68231b410d7bdd4e2f998181
- Allow gated user
- Full Datasets (maybe, 1M?)
- 2.x will train longer, smarter, and more like a man/lady (Premium)
- Using famous models like Qwen or Llama etc, DeepSeek. we will do our best!
Coming soon 2.x at kulia-moon/lily-2x-68231b410d7bdd4e2f998181
By Kulia-
- 2 replies
hesamation
posted
an
update
13 days ago
Post
3041
this book actually exists for free, “the little book of deep learning”. best to refresh your mind about DL basics:
> foundations of machine learning
> how models train
> common layers (dropout, pooling…)
> basic intro to LLMs
actually optimized for mobile.
Book: https://fleuret.org/public/lbdl.pdf
> foundations of machine learning
> how models train
> common layers (dropout, pooling…)
> basic intro to LLMs
actually optimized for mobile.
Book: https://fleuret.org/public/lbdl.pdf
kulia-moon
posted
an
update
14 days ago
juhoinkinen
authored
a
paper
15 days ago
juhoinkinen
posted
an
update
16 days ago
Post
2763
We (
@osma
,
@MonaLehtinen
& me, i.e. the Annif team at the National Library of Finland) recently took part in the LLMs4Subjects challenge at the SemEval-2025 workshop. The task was to use large language models (LLMs) to generate good quality subject indexing for bibliographic records, i.e. titles and abstracts.
We are glad to report that our system performed well; it was ranked
🥇 1st in the category where the full vocabulary was used
🥈 2nd in the smaller vocabulary category
🏅 4th in the qualitative evaluations.
14 participating teams developed their own solutions for generating subject headings and the output of each system was assessed using both quantitative and qualitative evaluations. Research papers about most of the systems are going to be published around the time of the workshop in late July, and many pre-prints are already available.
We applied Annif together with several LLMs that we used to preprocess the data sets: translated the GND vocabulary terms to English, translated bibliographic records into English and German as required, and generated additional synthetic training data. After the preprocessing, we used the traditional machine learning algorithms in Annif as well as the experimental XTransformer algorithm that is based on language models. We also combined the subject suggestions generated using English and German language records in a novel way.
More information can be found in our system description preprint: Annif at SemEval-2025 Task 5: Traditional XMTC augmented by LLMs (2504.19675)
See also the task description preprint: SemEval-2025 Task 5: LLMs4Subjects -- LLM-based Automated Subject Tagging for a National Technical Library's Open-Access Catalog (2504.07199)
The Annif models trained for this task are available here: NatLibFi/Annif-LLMs4Subjects-data
We are glad to report that our system performed well; it was ranked
🥇 1st in the category where the full vocabulary was used
🥈 2nd in the smaller vocabulary category
🏅 4th in the qualitative evaluations.
14 participating teams developed their own solutions for generating subject headings and the output of each system was assessed using both quantitative and qualitative evaluations. Research papers about most of the systems are going to be published around the time of the workshop in late July, and many pre-prints are already available.
We applied Annif together with several LLMs that we used to preprocess the data sets: translated the GND vocabulary terms to English, translated bibliographic records into English and German as required, and generated additional synthetic training data. After the preprocessing, we used the traditional machine learning algorithms in Annif as well as the experimental XTransformer algorithm that is based on language models. We also combined the subject suggestions generated using English and German language records in a novel way.
More information can be found in our system description preprint: Annif at SemEval-2025 Task 5: Traditional XMTC augmented by LLMs (2504.19675)
See also the task description preprint: SemEval-2025 Task 5: LLMs4Subjects -- LLM-based Automated Subject Tagging for a National Technical Library's Open-Access Catalog (2504.07199)
The Annif models trained for this task are available here: NatLibFi/Annif-LLMs4Subjects-data
- 2 replies
mrfakename
posted
an
update
22 days ago
Post
2708
Hi everyone,
I just launched TTS Arena V2 - a platform for benchmarking TTS models by blind A/B testing. The goal is to make it easy to compare quality between open-source and commercial models, including conversational ones.
What's new in V2:
- **Conversational Arena**: Evaluate models like CSM-1B, Dia 1.6B, and PlayDialog in multi-turn settings
- **Personal Leaderboard**: Optional login to see which models you tend to prefer
- **Multi-speaker TTS**: Random voices per generation to reduce speaker bias
- **Performance Upgrade**: Rebuilt from Gradio → Flask. Much faster with fewer failed generations.
- **Keyboard Shortcuts**: Vote entirely via keyboard
Also added models like MegaTTS 3, Cartesia Sonic, and ElevenLabs' full lineup.
I'd love any feedback, feature suggestions, or ideas for models to include.
TTS-AGI/TTS-Arena-V2
I just launched TTS Arena V2 - a platform for benchmarking TTS models by blind A/B testing. The goal is to make it easy to compare quality between open-source and commercial models, including conversational ones.
What's new in V2:
- **Conversational Arena**: Evaluate models like CSM-1B, Dia 1.6B, and PlayDialog in multi-turn settings
- **Personal Leaderboard**: Optional login to see which models you tend to prefer
- **Multi-speaker TTS**: Random voices per generation to reduce speaker bias
- **Performance Upgrade**: Rebuilt from Gradio → Flask. Much faster with fewer failed generations.
- **Keyboard Shortcuts**: Vote entirely via keyboard
Also added models like MegaTTS 3, Cartesia Sonic, and ElevenLabs' full lineup.
I'd love any feedback, feature suggestions, or ideas for models to include.
TTS-AGI/TTS-Arena-V2
- 2 replies
alielfilali01
posted
an
update
23 days ago
Post
563
Great efforts from
@AtlasIA
folks to adapt text2image models (ghibli style) for Moroccan Context
Read the blog is here : https://huggingface.co/blog/atlasia/creating-your-custom-ghibli-text-to-image-model
Read the blog is here : https://huggingface.co/blog/atlasia/creating-your-custom-ghibli-text-to-image-model
Post
3380
𝗜 𝘁𝗿𝗮𝗶𝗻𝗲𝗱 𝗮 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹 𝘁𝗼 𝘀𝗰𝗵𝗲𝗱𝘂𝗹𝗲 𝗲𝘃𝗲𝗻𝘁𝘀 𝘄𝗶𝘁𝗵 𝗚𝗥𝗣𝗢! 👑 🗓️
✍️ Blog post: https://huggingface.co/blog/anakin87/qwen-scheduler-grpo
I experimented with GRPO lately.
I am fascinated by models learning from prompts and rewards - no example answers needed like in Supervised Fine-Tuning.
After the DeepSeek boom, everyone is trying GRPO with GSM8K or the Countdown Game...
I wanted a different challenge, like 𝘁𝗲𝗮𝗰𝗵𝗶𝗻𝗴 𝗮 𝗺𝗼𝗱𝗲𝗹 𝘁𝗼 𝗰𝗿𝗲𝗮𝘁𝗲 𝗮 𝘀𝗰𝗵𝗲𝗱𝘂𝗹𝗲 𝗳𝗿𝗼𝗺 𝗮 𝗹𝗶𝘀𝘁 𝗼𝗳 𝗲𝘃𝗲𝗻𝘁𝘀 𝗮𝗻𝗱 𝗽𝗿𝗶𝗼𝗿𝗶𝘁𝗶𝗲𝘀.
Choosing an original problem forced me to:
🤔 Think about the problem setting
🧬 Generate data
🤏 Choose the right base model
🏆 Design reward functions (and experiencing reward hacking)
🔄 Run multiple rounds of training, hoping that my model would learn something.
A fun and rewarding 😄 experience.
I learned a lot of things, that I want to share with you. 👇
✍️ Blog post: https://huggingface.co/blog/anakin87/qwen-scheduler-grpo
💻 Code: https://github.com/anakin87/qwen-scheduler-grpo
🤗 Hugging Face collection (dataset and model): anakin87/qwen-scheduler-grpo-680bcc583e817390525a8837
✍️ Blog post: https://huggingface.co/blog/anakin87/qwen-scheduler-grpo
I experimented with GRPO lately.
I am fascinated by models learning from prompts and rewards - no example answers needed like in Supervised Fine-Tuning.
After the DeepSeek boom, everyone is trying GRPO with GSM8K or the Countdown Game...
I wanted a different challenge, like 𝘁𝗲𝗮𝗰𝗵𝗶𝗻𝗴 𝗮 𝗺𝗼𝗱𝗲𝗹 𝘁𝗼 𝗰𝗿𝗲𝗮𝘁𝗲 𝗮 𝘀𝗰𝗵𝗲𝗱𝘂𝗹𝗲 𝗳𝗿𝗼𝗺 𝗮 𝗹𝗶𝘀𝘁 𝗼𝗳 𝗲𝘃𝗲𝗻𝘁𝘀 𝗮𝗻𝗱 𝗽𝗿𝗶𝗼𝗿𝗶𝘁𝗶𝗲𝘀.
Choosing an original problem forced me to:
🤔 Think about the problem setting
🧬 Generate data
🤏 Choose the right base model
🏆 Design reward functions (and experiencing reward hacking)
🔄 Run multiple rounds of training, hoping that my model would learn something.
A fun and rewarding 😄 experience.
I learned a lot of things, that I want to share with you. 👇
✍️ Blog post: https://huggingface.co/blog/anakin87/qwen-scheduler-grpo
💻 Code: https://github.com/anakin87/qwen-scheduler-grpo
🤗 Hugging Face collection (dataset and model): anakin87/qwen-scheduler-grpo-680bcc583e817390525a8837
- 2 replies
Post
4454
BOOOOM: Today I'm dropping TINY AGENTS
the 50 lines of code Agent in Javascript 🔥
I spent the last few weeks working on this, so I hope you will like it.
I've been diving into MCP (Model Context Protocol) to understand what the hype was all about.
It is fairly simple, but still quite powerful: MCP is a standard API to expose sets of Tools that can be hooked to LLMs.
But while doing that, came my second realization:
Once you have a MCP Client, an Agent is literally just a while loop on top of it. 🤯
➡️ read it exclusively on the official HF blog: https://huggingface.co/blog/tiny-agents
the 50 lines of code Agent in Javascript 🔥
I spent the last few weeks working on this, so I hope you will like it.
I've been diving into MCP (Model Context Protocol) to understand what the hype was all about.
It is fairly simple, but still quite powerful: MCP is a standard API to expose sets of Tools that can be hooked to LLMs.
But while doing that, came my second realization:
Once you have a MCP Client, an Agent is literally just a while loop on top of it. 🤯
➡️ read it exclusively on the official HF blog: https://huggingface.co/blog/tiny-agents
- 1 reply
Post
1964
If you haven't had the chance to test the latest open model from Meta, Llama 4 Maverick, go try it on AMD MI 300 on Hugging Face!
amd/llama4-maverick-17b-128e-mi-amd
amd/llama4-maverick-17b-128e-mi-amd
Post
891
https://www.youtube.com/watch?v=AN-iZblyZNE
Discover how to harness the power of NOMIC Embed Vision v1.5 to find similar images within GNU Emacs Dired mode. With this innovative embeddings model, you can search for images based on semantic similarities using simple keywords. This is possible because the text model of NOMIC shares the same vector space as the Embed Vision model.
In this video, we'll show you how to run the script on your computer and explore the capabilities of this groundbreaking model. You'll learn how to find similar pictures and enjoy the convenience of searching for images using just a few words.
Don't miss out on this exciting opportunity to enhance your image searching experience with NOMIC Embed Vision v1.5 in Emacs Lisp.
Script to run model:
https://gitea.com/gnusupport/LLM-Helpers/src/branch/main/bin/nomic-embed-vision-v1.5-api.py
Discover how to harness the power of NOMIC Embed Vision v1.5 to find similar images within GNU Emacs Dired mode. With this innovative embeddings model, you can search for images based on semantic similarities using simple keywords. This is possible because the text model of NOMIC shares the same vector space as the Embed Vision model.
In this video, we'll show you how to run the script on your computer and explore the capabilities of this groundbreaking model. You'll learn how to find similar pictures and enjoy the convenience of searching for images using just a few words.
Don't miss out on this exciting opportunity to enhance your image searching experience with NOMIC Embed Vision v1.5 in Emacs Lisp.
Script to run model:
https://gitea.com/gnusupport/LLM-Helpers/src/branch/main/bin/nomic-embed-vision-v1.5-api.py
- 1 reply
hesamation
posted
an
update
about 1 month ago
Post
2969
The best researchers from DeepSeek, OpenAI, Microsoft, and ByteDance explored RL and Reasoning in LLMs,
Here's some of their key findings:
1/ RL can further improve distilled models. These models are essentially SFT fine-tuned with the data generated by larger models, and the SFT+RL combo does not disappoint.
This is verified in the DeepSeek-R1 paper.
2/ both GRPO and PPO algorithms suffer from length bias; they encourage longer responses. This can be tackled by introducing explicit rewards based on the length of the answer.
3/Most reasoning research is focused on code and math. But training models on logic puzzles improves them for mathematical tasks too.
This shows the RL reasoning is generalized beyond the specific domain knowledge.
Previous research also shows RL can be a great generalizer.
4/The reasoning might not be only induced by RL; it might already be hidden in the base models due to the pre-training and CoT data they were trained on.
So while RL does wake up the reasoning beast, maybe it's not the only solution (e.g. other methods such as distillation)
5/ back to the length bias; reasoning models tend to generate longer responses for wrong answers. RL might be the culprit.
RL favours longer answers when the reward is negative, to dilute the penalty per individual token and lower the loss.
This might explain the "aha" moments!
6/ OpenAI's competitive programming paper showed an interesting finding:
o3 can learn its own test-time strategies (like writing an inefficient but correct solution to verify the answer of an optimized solution)
RL helps LLMs develop their own reasoning & verification methods.
The recent article by @rasbt helped me a lot in getting a broad view of the recent research on reasoning models.
He also lists more influential papers on this topic, It's a must-read if you're interested.
check it out 👇
https://magazine.sebastianraschka.com/p/the-state-of-llm-reasoning-model-training
Here's some of their key findings:
1/ RL can further improve distilled models. These models are essentially SFT fine-tuned with the data generated by larger models, and the SFT+RL combo does not disappoint.
This is verified in the DeepSeek-R1 paper.
2/ both GRPO and PPO algorithms suffer from length bias; they encourage longer responses. This can be tackled by introducing explicit rewards based on the length of the answer.
3/Most reasoning research is focused on code and math. But training models on logic puzzles improves them for mathematical tasks too.
This shows the RL reasoning is generalized beyond the specific domain knowledge.
Previous research also shows RL can be a great generalizer.
4/The reasoning might not be only induced by RL; it might already be hidden in the base models due to the pre-training and CoT data they were trained on.
So while RL does wake up the reasoning beast, maybe it's not the only solution (e.g. other methods such as distillation)
5/ back to the length bias; reasoning models tend to generate longer responses for wrong answers. RL might be the culprit.
RL favours longer answers when the reward is negative, to dilute the penalty per individual token and lower the loss.
This might explain the "aha" moments!
6/ OpenAI's competitive programming paper showed an interesting finding:
o3 can learn its own test-time strategies (like writing an inefficient but correct solution to verify the answer of an optimized solution)
RL helps LLMs develop their own reasoning & verification methods.
The recent article by @rasbt helped me a lot in getting a broad view of the recent research on reasoning models.
He also lists more influential papers on this topic, It's a must-read if you're interested.
check it out 👇
https://magazine.sebastianraschka.com/p/the-state-of-llm-reasoning-model-training
Post
2091
Back to LLM integration.
ClickDefine.sh -- quickly define or explain anything within your whole desktop environment
You only need to run the model locally, maybe with the **llama.cpp** or **ollama**
- https://github.com/ggml-org/llama.cpp
- https://ollama.com/download
And you get universal explaining tool that works anywhere on your X Org Desktop (on operating systems which are usually Fully Free Software like Debian GNU/Linux)
ClickDefine - Interactive Text Processor Script for Iterative LLM Query Handling:
https://hyperscope.link/9/6/0/9/8/ClickDefine-Interactive-Text-Processor-Script-for-Iterative-LLM-Query-Handling-96098.html
Watch the demonstration here: https://www.youtube.com/watch?v=mQxCYAiReu0&t=2s
ClickDefine.sh -- quickly define or explain anything within your whole desktop environment
You only need to run the model locally, maybe with the **llama.cpp** or **ollama**
- https://github.com/ggml-org/llama.cpp
- https://ollama.com/download
And you get universal explaining tool that works anywhere on your X Org Desktop (on operating systems which are usually Fully Free Software like Debian GNU/Linux)
ClickDefine - Interactive Text Processor Script for Iterative LLM Query Handling:
https://hyperscope.link/9/6/0/9/8/ClickDefine-Interactive-Text-Processor-Script-for-Iterative-LLM-Query-Handling-96098.html
Watch the demonstration here: https://www.youtube.com/watch?v=mQxCYAiReu0&t=2s
hesamation
posted
an
update
about 1 month ago
Post
2169
OpenAI just released a 34-page practical guide to building agents,
Here's 10 things it teaches us:
1➜ agents are different from workflows: they are complete autonomous systems that perform tasks on your behalf. many applications use LLMs for workflows, but this is not an agent.
2➜ use them for tricky stuff: complex decision making, dynamic rules, unstructured data
3➜ core recipe: each agent has three main components: Model (the brain), Tools, Instructions on how to behave
4➜ choose the right brain: set up evals to get a baseline performance, use a smart model to see what's possible, gradually downgrade the model for cost and speed
5➜ tools are key: choose well-defined and tested tools. an agent needs tools to retrieve data and context, and take actions.
6➜ instruction matters A LOT: be super clear telling the agent its goals, steps, and rules. Vague instructions = unpredictable agent. Be explicit.
7➜ start simple, then scale: often a single agent with several tools is ok. don't jump to complex multi-agent systems immediately.
8➜ if you use multi-agents: you can have a "manager" agent directing traffic to specialist agents, or have agents hand off tasks to each other.
9➜ gaurdrails are a MUST: check user input for weird stuff, make sure the agent isn't about to do something risky, filter out private info, block harmful content. Don't let it run wild.
10➜ build and plan for humans: start small, test, improve. always have a plan for when the agent gets stuck or is about to do something high-risk.
Download: https://t.co/fJaCkgf7ph
Here's 10 things it teaches us:
1➜ agents are different from workflows: they are complete autonomous systems that perform tasks on your behalf. many applications use LLMs for workflows, but this is not an agent.
2➜ use them for tricky stuff: complex decision making, dynamic rules, unstructured data
3➜ core recipe: each agent has three main components: Model (the brain), Tools, Instructions on how to behave
4➜ choose the right brain: set up evals to get a baseline performance, use a smart model to see what's possible, gradually downgrade the model for cost and speed
5➜ tools are key: choose well-defined and tested tools. an agent needs tools to retrieve data and context, and take actions.
6➜ instruction matters A LOT: be super clear telling the agent its goals, steps, and rules. Vague instructions = unpredictable agent. Be explicit.
7➜ start simple, then scale: often a single agent with several tools is ok. don't jump to complex multi-agent systems immediately.
8➜ if you use multi-agents: you can have a "manager" agent directing traffic to specialist agents, or have agents hand off tasks to each other.
9➜ gaurdrails are a MUST: check user input for weird stuff, make sure the agent isn't about to do something risky, filter out private info, block harmful content. Don't let it run wild.
10➜ build and plan for humans: start small, test, improve. always have a plan for when the agent gets stuck or is about to do something high-risk.
Download: https://t.co/fJaCkgf7ph
-
- 3 replies