When implemented into a robotic arm, for example, it helped swap two toy fruits across three circular mats, a minimal example of a family of long-horizon tasks that require memories. Despite starting from random positions and seeing distractions like a shopping bag blocking the
RESEARCH
-
CSAIL Researchers Introduce Diffusion Forcing for Sequence Models
By
–
To combine the strengths of both models, CSAIL researchers created a sequence model training technique called "Diffusion Forcing." The name comes from "Teacher Forcing," the conventional training scheme that breaks down full sequence generation into the smaller, easier steps of
-
Diffusion Forcing Bridges Diffusion Models and Teacher Forcing
By
–
Diffusion Forcing found common ground between diffusion models & teacher forcing: they both use training schemes that involve predicting masked (noisy) tokens from unmasked ones. In the case of diffusion models, they gradually add noise to data, which can be viewed as fractional
-
MIT’s Diffusion Forcing: Combining Token Prediction and Video Diffusion
By
–
Sequence models have skyrocketed in popularity for their ability to analyze data & predict what to do next.
— MIT CSAIL (@MIT_CSAIL) 17 octobre 2024
MIT’s "Diffusion Forcing" method combines the strengths of next-token prediction (like w/ChatGPT) & video diffusion (like w/Sora), training neural networks to handle… pic.twitter.com/u9CPIGq0tjSequence models have skyrocketed in popularity for their ability to analyze data & predict what to do next. MIT’s "Diffusion Forcing" method combines the strengths of next-token prediction (like w/ChatGPT) & video diffusion (like w/Sora), training neural networks to handle
-
Honest Model Evaluation Over Hype in AI Research
By
–
This is the opposite imo: they claimed good scores on three benchmarks and that's it. No grandiose claims in the model card or the paper. The team made a valuable contribution, let's just not overhype it.
-
Llama 3.1 Nemotron reveals benchmark evaluation disparities
By
–
Llama-3.1-Nemotron-70B is a good reminder that chat capabilities evaluated by Arena Hard, AlpacaEval, and MT-Bench correlate poorly with benchmarks like MMLU and GPQA. They also provide a useful but narrow view of human preferences. Blame the benchmarks, not the models
-
Mycelium Graph Visualization Library for AI Applications
By
–
Mycelium • Graph visualization library https://
bit.ly/4dtRYIo
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
Leading AI Adoption: Bold Strategy and Learning Culture
By
–
What does it take to lead in AI? Louise Herring from McKinsey dives into the keys to AI adoption—embracing bold, strategic decisions while balancing innovation with responsibility. In her conversation with Cohere, she emphasizes fostering a learning culture to unlock AI's full… pic.twitter.com/QeXO5XPBVS
— Cohere (@cohere) 17 octobre 2024What does it take to lead in AI? Louise Herring from McKinsey dives into the keys to AI adoption—embracing bold, strategic decisions while balancing innovation with responsibility. In her conversation with Cohere, she emphasizes fostering a learning culture to unlock AI's full
-
Aymericroucher shares Open LLM Leaderboard Comparator link
By
–
Link here: https://
huggingface.co/spaces/open-ll
m-leaderboard/comparator
…
cc @avillanovamoral -
Open LLM Leaderboard releases visualizer to compare two models
By
–
> Coolest release today:
— m_ric (@AymericRoucher) 17 octobre 2024
The Open LLM Leaderboard, aka the best benchmark suite for comparing open LLMs, just released a visualizer to compare any two models together! 🔎
Example below: how does the new nvidia Llama-3.1-Nemotron-70B that we've heard so much about compare with… pic.twitter.com/OG7GslMraM> Coolest release today:
The Open LLM Leaderboard, aka the best benchmark suite for comparing open LLMs, just released a visualizer to compare any two models together! Example below: how does the new nvidia Llama-3.1-Nemotron-70B that we've heard so much about compare with