Post
6,336
New LaTeX Templates free for PhD students, Masters students and PostDocs.
Download here - https://trybibby.com/latex/templates
Download here - https://trybibby.com/latex/templates
- 1 reply
Julien BLANCHON PRO
blanchon
AI & ML interests
None yet
Recent Activity
liked a model about 20 hours ago
Tongyi-MAI/MAI-Gym-9B liked a model 2 days ago
marianna13/prelude-9b-sft-v2Organizations
reacted to BibbyResearch's post with ❤️ 2 days ago
posted an update 20 days ago
Post
207
I read a lot of papers these days, and I was frustrated with every PDF reader. So I built the one I wanted instead.
Available on Mac, Windows, Linux, Chrome and the web.
Xivly is free and open source 👇
https://github.com/julien-blanchon/xivly
Available on Mac, Windows, Linux, Chrome and the web.
Xivly is free and open source 👇
https://github.com/julien-blanchon/xivly
reacted to danielhanchen's post with ❤️🚀 22 days ago
Post
4,120
Google releases EmbeddingGemma 2, a new open embedding model that runs locally on 0.5GB RAM.
The 740M parameter Apache 2.0 model combines a 270M text model with vision (170M) + audio (300M).
Run & train the model via Unsloth.
GGUF: unsloth/embeddinggemma-2-GGUF
Guide: https://unsloth.ai/docs/models/embeddinggemma-2
The 740M parameter Apache 2.0 model combines a 270M text model with vision (170M) + audio (300M).
Run & train the model via Unsloth.
GGUF: unsloth/embeddinggemma-2-GGUF
Guide: https://unsloth.ai/docs/models/embeddinggemma-2
reacted to BramVanroy's post with 👍 3 months ago
Post
2,855
**I benchmarked HF buckets against https access for Common Crawl.**
Took me a while to get round to do this but I benchmarked access to Common Crawl via https vs hf buckets. Both experiments were run at night in Europe. I do not think other hardware problems were impacting the speeds since CPU processing time of the non-download pipeline components were highly similar (within 2% identical) and below only the WarcReader speeds of datatrove are used.
Experiment: selected 5 disjoint samples of 64 files each (randomly from the latest crawl; 20,499 docs/file). Those five batches were then processed by 32 single-core tasks with 4GB/core (five batches to calculate CIs). Paired experiment between using https and hf bucket.
- https: 40.0 [39.3-40.6] (seconds per WARC file)
- hf bucket: 172.0 [122.7-221.2]
That is a difference of about 4x in streaming speed. You'll see that https is also more stable (smaller CI).
I also ran raw throughput tests to the endpoints to measure rate limiting (64MiB transfer at 8/32/128/256 concurrent readers) and rate limiting seems not an issue for either: at any of those parallel reader numbers, their respective speeds stay about the same.
Note that, given CC scale, this is still a small test. Rate limiting may become more obvious when processing a full crawl. I do not know whether the https endpoint vs HF bucket will shut you out earlier with which limits.
Took me a while to get round to do this but I benchmarked access to Common Crawl via https vs hf buckets. Both experiments were run at night in Europe. I do not think other hardware problems were impacting the speeds since CPU processing time of the non-download pipeline components were highly similar (within 2% identical) and below only the WarcReader speeds of datatrove are used.
Experiment: selected 5 disjoint samples of 64 files each (randomly from the latest crawl; 20,499 docs/file). Those five batches were then processed by 32 single-core tasks with 4GB/core (five batches to calculate CIs). Paired experiment between using https and hf bucket.
- https: 40.0 [39.3-40.6] (seconds per WARC file)
- hf bucket: 172.0 [122.7-221.2]
That is a difference of about 4x in streaming speed. You'll see that https is also more stable (smaller CI).
I also ran raw throughput tests to the endpoints to measure rate limiting (64MiB transfer at 8/32/128/256 concurrent readers) and rate limiting seems not an issue for either: at any of those parallel reader numbers, their respective speeds stay about the same.
Note that, given CC scale, this is still a small test. Rate limiting may become more obvious when processing a full crawl. I do not know whether the https endpoint vs HF bucket will shut you out earlier with which limits.
reacted to nyuuzyou's post with 👍🔥 5 months ago
Post
2,729
🌍 Street-Level Imagery Dataset nyuuzyou/streetview
934,191 image records index Eastern Europe and Northern Asia. Temporal links map historical views at identical coordinates across nine years.
Key Stats:
- 905,940 unique images
- Coverage spanning 2016 to 2025
- Average 14.3 historical links per location
Geographic bounds span 20.49° E to 152.32° E. Urban centers show higher data density.
934,191 image records index Eastern Europe and Northern Asia. Temporal links map historical views at identical coordinates across nine years.
Key Stats:
- 905,940 unique images
- Coverage spanning 2016 to 2025
- Average 14.3 historical links per location
Geographic bounds span 20.49° E to 152.32° E. Urban centers show higher data density.
- 5 replies
replied to their post 6 months ago
Link to X thread/post: https://x.com/JulienBlanchon/status/2054519347574350115
reacted to spillai's post with 🔥 6 months ago
Post
8,804
mm-ctx – fast, multimodal context for agents.
LLM-based agents handle text incredibly well, but images, videos, or PDFs with visual content are hard to interpret. mm-ctx gives your CLI agent multi-modal skills.
Try it interactively in Spaces: vlm-run/mm-ctx
Readme: https://vlm-run.github.io/mm/
PyPI: https://pypi.org/project/mm-ctx
SKILL.md: https://github.com/vlm-run/skills/blob/main/skills/mm-cli-skill/SKILL.md
mm-ctx is meant to feel familiar: the UNIX tools we already love (find/cat/grep/wc), rebuilt for file types LLMs can't read natively and designed to work with agents via the CLI.
- mm grep "invoice #1234" ~/Downloads searches across PDFs and returns line-numbered matches
- mm cat <document>.pdf returns a metadata description of the file
- mm cat <photo>.jpg returns a caption of the photo
- mm cat <video>.mp4 returns a caption of the video
A few things we obsessed over:
⚡ Speed: Rust core for the hot paths
🏠 Local-first, BYO model: Uses any OpenAI-compatible endpoint: Ollama, vLLM/SGLang, LMStudio with any multimodal LLM (Gemma4, Qwen3.5, GLM-4.6V).
🔗 Composable: stdin + structured outputs
🤖 Drops into any agent via mm-cli-skills: Claude Code, Codex, Gemini CLI, OpenClaw.
We’d love to hear your feedback! Especially on the CLI and what file types and workflows you would like to see next.
LLM-based agents handle text incredibly well, but images, videos, or PDFs with visual content are hard to interpret. mm-ctx gives your CLI agent multi-modal skills.
Try it interactively in Spaces: vlm-run/mm-ctx
Readme: https://vlm-run.github.io/mm/
PyPI: https://pypi.org/project/mm-ctx
SKILL.md: https://github.com/vlm-run/skills/blob/main/skills/mm-cli-skill/SKILL.md
mm-ctx is meant to feel familiar: the UNIX tools we already love (find/cat/grep/wc), rebuilt for file types LLMs can't read natively and designed to work with agents via the CLI.
- mm grep "invoice #1234" ~/Downloads searches across PDFs and returns line-numbered matches
- mm cat <document>.pdf returns a metadata description of the file
- mm cat <photo>.jpg returns a caption of the photo
- mm cat <video>.mp4 returns a caption of the video
A few things we obsessed over:
⚡ Speed: Rust core for the hot paths
🏠 Local-first, BYO model: Uses any OpenAI-compatible endpoint: Ollama, vLLM/SGLang, LMStudio with any multimodal LLM (Gemma4, Qwen3.5, GLM-4.6V).
🔗 Composable: stdin + structured outputs
🤖 Drops into any agent via mm-cli-skills: Claude Code, Codex, Gemini CLI, OpenClaw.
We’d love to hear your feedback! Especially on the CLI and what file types and workflows you would like to see next.
- 2 replies
posted an update 6 months ago
Post
3,223
I'm releasing OpenCS2 a 11TB dataset of around 5000 hours of counter strike gameplay recording.
- HD resolution - 1280×720 · 32 fps
- For each frame keyboard and mouse + world state (player position, velocity, weapon ...)
- HD Stereo audio
- All 10 players perspective
https://huggingface.co/collections/blanchon/opencs2
- HD resolution - 1280×720 · 32 fps
- For each frame keyboard and mouse + world state (player position, velocity, weapon ...)
- HD Stereo audio
- All 10 players perspective
https://huggingface.co/collections/blanchon/opencs2
- 1 reply
reacted to projectlosangeles's post with 🔥❤️ 9 months ago
Post
1,918
Check out Orpheus Karaoke! Turn any MIDI into a unique Karaoke MIDI!
projectlosangeles/Orpheus-Karaoke
projectlosangeles/Orpheus-Karaoke
Hey @Tonic , I'm absolutly not related with the Liquid AI team. But happy to chat anytime (you can PM me on X maybe) !
reacted to vikhyatk's post with 🔥 11 months ago
Post
5,739
Announcing RefCOCO-M, a refreshed RefCOCO with pixel-accurate masks and the problematic prompts removed.
moondream/refcoco-m
moondream/refcoco-m
reacted to Kseniase's post with 🔥 12 months ago
Post
11,281
11 Fascinating new Policy Optimization techniques
Policy optimization (PO) algorithms are central to training AI models with preference-based feedback. In recent weeks, numerous new PO methods have emerged that build on or replace the popular PPO and GRPO, solving their issues. Here are 11 of them:
1. BAlanced Policy Optimization (BAPO) → BAPO: Stabilizing Off-Policy Reinforcement Learning for LLMs via Balanced Policy Optimization with Adaptive Clipping (2510.18927)
Dynamically adjusting the clipping bounds in PPO-style updates to balance positive and negative gradients and prevent entropy collapse
2. Training-Free GRPO → Training-Free Group Relative Policy Optimization (2510.08191)
Instead of using numeric rewards, it compares rollouts semantically to distill useful knowledge as a token prior, which is then applied during inference to guide the model’s behavior
3. Asymmetric Importance Sampling Policy Optimization (ASPO) → ASPO: Asymmetric Importance Sampling Policy Optimization (2510.06062)
Fixes imbalanced token weighting in LLM training. It flips the importance sampling ratios for positive tokens to correct over- and under-updates, and adds a soft dual-clipping step to keep gradients stable
4. In-Context Steered Policy Optimization (ICPO) → https://arxiv.org/abs/2510.26519
Uses a model’s own in-context learning ability to guide training with existing data. It combines Mixed-Policy GRPO with Implicit Expert Forcing to expand exploration and adds Expert Region Reject Sampling and Annealed Expert-Bonus Reward Shaping to ensure stability and balanced expert influence
5. Graph-Enhanced Policy Optimization (GEPO) → https://arxiv.org/abs/2510.26270
Builds a graph of an agent’s experiences to understand how different states connect, guide exploration and assign rewards more effectively
6. Information Gain-based Policy Optimization (IGPO) → Information Gain-based Policy Optimization: A Simple and Effective Approach for Multi-Turn LLM Agents (2510.14967)
Uses the model’s own belief updates to create dense, informative feedback for smoother multi-turn learning
Read further below ⬇️
If you like this, also subscribe to the Turing post: https://www.turingpost.com/subscribe
Policy optimization (PO) algorithms are central to training AI models with preference-based feedback. In recent weeks, numerous new PO methods have emerged that build on or replace the popular PPO and GRPO, solving their issues. Here are 11 of them:
1. BAlanced Policy Optimization (BAPO) → BAPO: Stabilizing Off-Policy Reinforcement Learning for LLMs via Balanced Policy Optimization with Adaptive Clipping (2510.18927)
Dynamically adjusting the clipping bounds in PPO-style updates to balance positive and negative gradients and prevent entropy collapse
2. Training-Free GRPO → Training-Free Group Relative Policy Optimization (2510.08191)
Instead of using numeric rewards, it compares rollouts semantically to distill useful knowledge as a token prior, which is then applied during inference to guide the model’s behavior
3. Asymmetric Importance Sampling Policy Optimization (ASPO) → ASPO: Asymmetric Importance Sampling Policy Optimization (2510.06062)
Fixes imbalanced token weighting in LLM training. It flips the importance sampling ratios for positive tokens to correct over- and under-updates, and adds a soft dual-clipping step to keep gradients stable
4. In-Context Steered Policy Optimization (ICPO) → https://arxiv.org/abs/2510.26519
Uses a model’s own in-context learning ability to guide training with existing data. It combines Mixed-Policy GRPO with Implicit Expert Forcing to expand exploration and adds Expert Region Reject Sampling and Annealed Expert-Bonus Reward Shaping to ensure stability and balanced expert influence
5. Graph-Enhanced Policy Optimization (GEPO) → https://arxiv.org/abs/2510.26270
Builds a graph of an agent’s experiences to understand how different states connect, guide exploration and assign rewards more effectively
6. Information Gain-based Policy Optimization (IGPO) → Information Gain-based Policy Optimization: A Simple and Effective Approach for Multi-Turn LLM Agents (2510.14967)
Uses the model’s own belief updates to create dense, informative feedback for smoother multi-turn learning
Read further below ⬇️
If you like this, also subscribe to the Turing post: https://www.turingpost.com/subscribe
- 2 replies
Amazing ! Any spaces to try this out quickly ?
reacted to piercus's post with 🔥👍 12 months ago
Post
4,098
Starts erasing! 🎉 🎉 🎉
This is made with a one-step SD1.5 LBM [1] eraser !
Data is open. Data pipeline is open. Training code is open.
On our LBM fork : https://github.com/finegrain-ai/LBM
[1] LBM: Latent Bridge Matching for Fast Image-to-Image Translation (2503.07535)
This is made with a one-step SD1.5 LBM [1] eraser !
Data is open. Data pipeline is open. Training code is open.
On our LBM fork : https://github.com/finegrain-ai/LBM
[1] LBM: Latent Bridge Matching for Fast Image-to-Image Translation (2503.07535)
- 1 reply
replied to mrfakename's post 12 months ago
LAION data is all you need xd
reacted to mrfakename's post with 🔥 12 months ago
Post
6,562
Trained a model for emotion-controllable TTS based on MiMo audio on LAION's dataset.
Still very early and does have an issue with hallucinating but results seem pretty good so far, given that it is very early into the training run.
Will probably kick off a new run later with some settings tweaked.
Put up a demo here: https://huggingface.co/spaces/mrfakename/EmoAct-MiMo
(Turn 🔊 on to hear audio samples)
Still very early and does have an issue with hallucinating but results seem pretty good so far, given that it is very early into the training run.
Will probably kick off a new run later with some settings tweaked.
Put up a demo here: https://huggingface.co/spaces/mrfakename/EmoAct-MiMo
(Turn 🔊 on to hear audio samples)
- 5 replies