Or Tal

PhD candidate @HebrewU; Research Assistant @MetaAI (FAIR)

[Or Tal @Or__Tal]

Which modeling to choose for text-to-music generation?
We run a head-to-head comparison to figure it out.
Same data, same architecture - AR vs FM.
👇 If you care about fidelity, speed, control, or editing see this thread.
🔗https://t.co/FBWu7ThspC
📄https://t.co/Dp1co1esvd
1/6

[Noam Issachar @IssacharNoam]

1/6 Diffusion models are scaling up, but deploying a massive, monolithic network uniformly across the entire generative timeline is inherently inefficient.
Introducing Complexity-Balanced Splitting (CBS): a principled framework that allocates capacity exactly where needed!👇🧵

[Eliahu Horwitz @EliahuHorwitz]

🚨 New paper alert! 🚨

Millions of neural networks now populate public repositories like Hugging Face 🤗, but most lack documentation. So, we decided to build an Atlas 🗺️

Project: https://t.co/1JpsC6dCeg
Demo: https://t.co/4Xy7yLdIZY

🧵👇🏻 Here's what we found:

[Eliahu Horwitz @EliahuHorwitz]

Excited to share this has now been accepted at #NeurIPS2025 as a position paper (<6% acceptance)!🎉

We advocate for systematically studying entire model populations via weight-space learning, and argue that this requires charting them in a Model Atlas.
@NeurIPSConf #NeurIPS
🧵👇

[Heli Ben-Hamu @helibenhamu]

Excited to share our work Set Block Decoding!

A new paradigm combining next-token-prediction and masked (or discrete diffusion) models, allowing parallel decoding without any architectural changes and with exact KV cache.

Arguably one of the simplest ways to accelerate LLMs!

[Or Tal @Or__Tal]

@jesseengel very cool!

[Or Tal @Or__Tal]

@__Rafail__ @NadavHarTuv I believe it should. If the audio is very long you may need to parse it in chunks but thats also true to all other audio representation models

[Or Tal @Or__Tal]

@__Rafail__ @NadavHarTuv Nope, PAST has ~180M params, with the streamable version having ~125M params. This should run on a standard gpu. For speech LM training we used 2 a100 gpus but it could be done with less

[Gallil Maimon @GallilMaimon]

Many modern SpeechLMs are trained with Speech-Text interleaving. How does this impact scaling trends?

In our new paper, we train several dozen SLMs, and show - quite a lot! So there is room for optimism 😊

Key insights, code, models, full paper 👇🏻

[Gallil Maimon @GallilMaimon]

🎉Thrilled that our paper on "scaling analysis of interleaved speech-text LMs" was accepted to #CoLM2025
It gives room for optimism when scaling SpeechLMs *right* - with large TextLMs (in place of more data), interleaving, and synth training data💪

[Ron Yosef @ron_yosef]

Happy to announce that our paper “EditInspector: A Benchmark for Evaluation of Text-Guided Image Edits” was accepted to #ACL2025 🎉

📄 https://t.co/mwugXz1H5q
🌐 https://t.co/NMmdzi3nBn

[Or Tal @Or__Tal]

💣Introducing PAST: a speech tokenizer that jointly model phonetics and acoustics (No SSL involved).
Past demonstrates great reconstruction as well as semantic capabilities in the form of ABX and sWUGGY.
🤗 https://t.co/teQQ9s5whr
Check out Nadav's post👇@NadavHarTuv @adiyossLC

[נדב הר-טוב @NadavHarTuv]

🚨 New paper alert!
PAST: phonetic-acoustic speech tokenizer – just got accepted to Interspeech 2025 🎉
It learns phonetic + acoustic tokens jointly, with no SSL babysitter or external vocoder.

🔗https://t.co/yGypWO6YpM
👇 If you’re into speech LMs, keep reading!

[Audio and Speech Processing Papers @AudioAndSpeech]

Auto-Regressive vs Flow-Matching: a Comparative Study of Modeling Paradigms for Text-to-Music Generation. https://t.co/pbwzHsgUGj

[Gallil Maimon @GallilMaimon]

🎵💬 If you are interested in Audio Tokenisers, you should check out our new work!
We empirically analysed existing tokenisers from every way - reconstruction, downstream, LMs and more.

Grab yourself a ☕/🍺 and sit down for a read!

[Niv Eckhaus @niveckhaus]

🚨 New Paper: "Time to Talk"! 🕵️

We built an LLM agent that doesn't just decide WHAT to say, but also WHEN to say it!

Introducing "Time to Talk" - LLM agents for asynchronous group communication, tested in real Mafia games with human players.

🌐https://t.co/HdNUwlvF2F
🧵1/7

[Or Tal @Or__Tal]

Read the full paper!
🔗https://t.co/FBWu7TgUA4
📄https://t.co/Dp1co1dUFF

@FelixKreuk @adiyossLC

[Or Tal @Or__Tal]

Which modeling to choose for text-to-music generation?
We run a head-to-head comparison to figure it out.
Same data, same architecture - AR vs FM.
👇 If you care about fidelity, speed, control, or editing see this thread.
🔗https://t.co/FBWu7ThspC
📄https://t.co/Dp1co1esvd
1/6

[Or Tal @Or__Tal]

What if training steps are capped at 500k?
FM reaches near-topline quality with small batches.
It’s compute-efficient and forgiving.
AR needs larger batch sizes to recover performance.
It benefits more from large-scale training.
See📉 below by model duration + batch size:
6/6

[Felix Kreuk @FelixKreuk]

We’ve been exploring the trade-offs between Autoregressive and Flow-Matching models for music generation. We share our findings in this latest paper led by @Or__Tal. Many interesting take-aways and practical advice on training generative models for music! 🎶🧠