MRAG-Bench: Vision-Centric Evaluation for Retrieval-Augmented Multimodal Models
Paper • 2410.08182 • Published
AI & ML interests
Contributors who are invited to beta-test our next big feature! Contact us if you want to join this team :-)
Recent Activity
View all activity
mohsenfayyaz
authored 3
papers 4 days ago
AlekseyKorshuk
authored a
paper 3 months ago
HannaAbiAkl
authored a
paper 3 months ago
AlekseyKorshuk
authored a
paper 4 months ago
monsoon-nlp
posted an update 5 months ago
Post
475
PatchDNA, a DNA foundation model based on Meta's BLT tokenization strategy https://www.biorxiv.org/content/10.1101/2025.11.28.691095v1
SivilTaram
authored a
paper 6 months ago
nouamanetazi
posted an update 6 months ago
Post
4789
After training 𝐒𝐦𝐨𝐥𝐋𝐌𝟑 on 𝟑𝟖𝟒 𝐇𝟏𝟎𝟎𝐬 for nearly a month, I've come to realize something most people overlook: 𝐢𝐧𝐟𝐫𝐚𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞 𝐢𝐬 𝐭𝐡𝐞 𝐦𝐚𝐤𝐞-𝐨𝐫-𝐛𝐫𝐞𝐚𝐤 𝐟𝐚𝐜𝐭𝐨𝐫 𝐢𝐧 𝐋𝐋𝐌 𝐭𝐫𝐚𝐢𝐧𝐢𝐧𝐠. 🔥
Everyone talks about model architecture and data quality. And yes, those matter immensely. But here's what nobody tells you: when your training run fails at 2 AM because of mysterious 𝐍𝐂𝐂𝐋 𝐞𝐫𝐫𝐨𝐫𝐬, or when your expensive GPU cluster is running at 𝟔𝟎% 𝐞𝐟𝐟𝐢𝐜𝐢𝐞𝐧𝐜𝐲, the problem isn't your model. It's most probably a 𝐦𝐢𝐬𝐮𝐬𝐞 𝐨𝐟 𝐭𝐡𝐞 𝐡𝐚𝐫𝐝𝐰𝐚𝐫𝐞. 🛠️
Questions that seemed simple but had no clear answers: Why is 𝐌𝐨𝐄 𝐭𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝐬𝐥𝐨𝐰𝐞𝐫 𝐭𝐡𝐚𝐧 𝐝𝐞𝐧𝐬𝐞 𝐦𝐨𝐝𝐞𝐥𝐬? Which 𝐍𝐂𝐂𝐋 𝐟𝐥𝐚𝐠𝐬 should we actually set? How often should we checkpoint without killing throughput?
That's why we built 𝐓𝐡𝐞 𝐒𝐦𝐨𝐥 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝐏𝐥𝐚𝐲𝐛𝐨𝐨𝐤 📖: a complete guide covering everything from model architecture and data curation to the SmolLM3 training marathon, post-training techniques, and crucially, the 𝐢𝐧𝐟𝐫𝐚𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞 𝐥𝐚𝐲𝐞𝐫 that most teams get wrong.
We validated real vs theoretical bandwidth across the entire stack: 𝐇𝐁𝐌𝟑 𝐡𝐢𝐭𝐭𝐢𝐧𝐠 𝟑 𝐓𝐁/𝐬, 𝐍𝐕𝐋𝐢𝐧𝐤 𝟒.𝟎 𝐫𝐞𝐚𝐜𝐡𝐢𝐧𝐠 𝟕𝟖𝟔 𝐆𝐁/𝐬, 𝐏𝐂𝐈𝐞 𝐆𝐞𝐧𝟒 𝐚𝐭 𝟏𝟒.𝟐 𝐆𝐁/𝐬. Then we ran collective operations across 𝟏𝟐𝟖 𝐆𝐏𝐔𝐬 (16 nodes, 8xH100s each) and measured how performance degrades at scale: all-reduce drops from 𝟒𝟖𝟎 𝐆𝐁/𝐬 on a single node to 𝟑𝟐𝟎-𝟑𝟓𝟎 𝐆𝐁/𝐬 across 16 nodes.
If you've ever wondered why your training runs are slower than they should be, or you're planning to scale up and want to avoid expensive mistakes, this guide might save you weeks of debugging.
𝐓𝐡𝐞 𝐒𝐦𝐨𝐥 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝐏𝐥𝐚𝐲𝐛𝐨𝐨𝐤: https://lnkd.in/e5MKXUHS
Shared with ❤️ by the HuggingFace team
Everyone talks about model architecture and data quality. And yes, those matter immensely. But here's what nobody tells you: when your training run fails at 2 AM because of mysterious 𝐍𝐂𝐂𝐋 𝐞𝐫𝐫𝐨𝐫𝐬, or when your expensive GPU cluster is running at 𝟔𝟎% 𝐞𝐟𝐟𝐢𝐜𝐢𝐞𝐧𝐜𝐲, the problem isn't your model. It's most probably a 𝐦𝐢𝐬𝐮𝐬𝐞 𝐨𝐟 𝐭𝐡𝐞 𝐡𝐚𝐫𝐝𝐰𝐚𝐫𝐞. 🛠️
Questions that seemed simple but had no clear answers: Why is 𝐌𝐨𝐄 𝐭𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝐬𝐥𝐨𝐰𝐞𝐫 𝐭𝐡𝐚𝐧 𝐝𝐞𝐧𝐬𝐞 𝐦𝐨𝐝𝐞𝐥𝐬? Which 𝐍𝐂𝐂𝐋 𝐟𝐥𝐚𝐠𝐬 should we actually set? How often should we checkpoint without killing throughput?
That's why we built 𝐓𝐡𝐞 𝐒𝐦𝐨𝐥 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝐏𝐥𝐚𝐲𝐛𝐨𝐨𝐤 📖: a complete guide covering everything from model architecture and data curation to the SmolLM3 training marathon, post-training techniques, and crucially, the 𝐢𝐧𝐟𝐫𝐚𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞 𝐥𝐚𝐲𝐞𝐫 that most teams get wrong.
We validated real vs theoretical bandwidth across the entire stack: 𝐇𝐁𝐌𝟑 𝐡𝐢𝐭𝐭𝐢𝐧𝐠 𝟑 𝐓𝐁/𝐬, 𝐍𝐕𝐋𝐢𝐧𝐤 𝟒.𝟎 𝐫𝐞𝐚𝐜𝐡𝐢𝐧𝐠 𝟕𝟖𝟔 𝐆𝐁/𝐬, 𝐏𝐂𝐈𝐞 𝐆𝐞𝐧𝟒 𝐚𝐭 𝟏𝟒.𝟐 𝐆𝐁/𝐬. Then we ran collective operations across 𝟏𝟐𝟖 𝐆𝐏𝐔𝐬 (16 nodes, 8xH100s each) and measured how performance degrades at scale: all-reduce drops from 𝟒𝟖𝟎 𝐆𝐁/𝐬 on a single node to 𝟑𝟐𝟎-𝟑𝟓𝟎 𝐆𝐁/𝐬 across 16 nodes.
If you've ever wondered why your training runs are slower than they should be, or you're planning to scale up and want to avoid expensive mistakes, this guide might save you weeks of debugging.
𝐓𝐡𝐞 𝐒𝐦𝐨𝐥 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝐏𝐥𝐚𝐲𝐛𝐨𝐨𝐤: https://lnkd.in/e5MKXUHS
Shared with ❤️ by the HuggingFace team
monsoon-nlp
posted an update 7 months ago
Post
468
Bio LLMs train on many genomes, but can we encode differences within a species? TomatoTomato adds pangenome tokens to represent a domestic tomato and a wild tomato in one sequence 🍅 🧬
monsoon-nlp/tomatotomato-gLM2-150M-v0.1
monsoon-nlp/tomatotomato-gLM2-150M-v0.1
SivilTaram
authored a
paper 8 months ago
SivilTaram
authored 3
papers 10 months ago
SWE-Perf: Can Language Models Optimize Code Performance on Real-World Repositories?
Paper • 2507.12415 • Published • 43
First Return, Entropy-Eliciting Explore
Paper • 2507.07017 • Published • 24
ZeCO: Zero Communication Overhead Sequence Parallelism for Linear Attention
Paper • 2507.01004 • Published • 10
Post
4563
🚀 ZeroGPU now supports PyTorch native quantization via
While it hasn’t been battle-tested yet,
Let us know if you run into any issues — and we’re excited to see what the community will build!
torchaoWhile it hasn’t been battle-tested yet,
Int8WeightOnlyConfig is already working flawlessly in our tests.Let us know if you run into any issues — and we’re excited to see what the community will build!
import spaces
from diffusers import FluxPipeline
from torchao.quantization.quant_api import Int8WeightOnlyConfig, quantize_
pipeline = FluxPipeline.from_pretrained(...).to('cuda')
quantize_(pipeline.transformer, Int8WeightOnlyConfig()) # Or any other component(s)
@spaces.GPU
def generate(prompt: str):
return pipeline(prompt).images[0]
- 5 replies
SivilTaram
authored a
paper 12 months ago
drscotthawley
authored a
paper 12 months ago
Post
6168
🚀 ZeroGPU
Nothing too fancy for now—ZeroGPU Spaces still default to
- 💰 size-based quotas / pricing (
- 🦣 the upcoming
You can as of now control GPU size via a Space variable. Accepted values:
-
-
-
The auto mode checks total CUDA tensor size during startup:
- More than 30GB →
- Otherwise →
medium size is now available as a power-user featureNothing too fancy for now—ZeroGPU Spaces still default to
large (70GB VRAM)—but this paves the way for:- 💰 size-based quotas / pricing (
medium will offer significantly more usage than large)- 🦣 the upcoming
xlarge size (141GB VRAM)You can as of now control GPU size via a Space variable. Accepted values:
-
auto (future default)-
medium-
large (current default)The auto mode checks total CUDA tensor size during startup:
- More than 30GB →
large- Otherwise →
medium 
- 3 replies
nouamanetazi
authored a
paper about 1 year ago
monsoon-nlp
posted an update about 1 year ago
Post
1828
I was curious about the Block Diffusion hybrid model and tried retraining it on a DNA tokenizer + dataset 🧬. Too early to evaluate, but it generates sequences (AAATGG TTATTG CAAATC...) and was improving on the validation set during training
Model: monsoon-nlp/dna-blockdiff-papaya
Original paper: Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models (2503.09573)
Model: monsoon-nlp/dna-blockdiff-papaya
Original paper: Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models (2503.09573)