Ram @ram_chandalada
Joined January 2015-
Tweets73
-
Followers118
-
Following194
-
Likes205
10M instructions could have been used to finetune LLaMa-3 instruct So here are ~2.2M (>1B tokens) high quality instructions curated from ~30 popular datasets (~6B tokens) across various tasks including function-calling 8.8M to go 🚀 huggingface.co/datasets/0-her…
Llama 3 has been my focus since joining the Llama team last summer. Together, we've been tackling challenges across pre-training and human data, pre-training scaling, long context, post-training, and evaluations. It's been a rigorous yet thrilling journey: 🔹Our largest models…
This release has flown under the radar. Very excited about HuggingChat on my phone! @huggingface
MT Bench Results for the latest Mixtral-8x22B-Instruct-v0.1
Well we had a nice run, humans.
Cool & hard benchmark: OSWorld. Where you have to fill tasks on ubuntu that requires multiple steps planning, and potentially search over internet to solve them. os-world.github.io
Being able to interpret an #ML model’s hidden representations is key to understanding its behavior. Today we introduce Patchscopes, an approach that trains #LLMs to provide natural language explanations of their own hidden representations. Learn more → goo.gle/4aS5epd
HumanEval benchmark results for the new Mixtral 8x22B, DBRX-base, Qwen1.5-72B, and Mixtral 8x7B:
A small experimental run with Mixture-of-Depths (from @iamgrigorev) & Bitnet. Used 4 types of OLMo (50M) on Dolma dataset for 100k steps - OLMo-50M -> model - OLMo-50M-bitlinear -> bitnet model - OLMo-50M-mod -> mixture-of-depths model - OLMo-50M-mod-bitlinear ->…
Our 12 scaling laws (for LLM knowledge capacity) are out: arxiv.org/abs/2404.05405. Took me 4mos to submit 50,000 jobs; took Meta 1mo for legal review; FAIR sponsored 4,200,000 GPU hrs. Hope this is a new direction to study scaling laws + help practitioners make informed decisions
Trying to recognise text from speech with Mistral-7B. Because, why not A small experiment inspired by "DOOM-Mistral" from the Mistral Hackathon 1. Convert Speech Audio to Waveform 2. Waveform charts to ASCII Art 3. Finetune 7B to predict text from waveform ASCII art Experiment…
Took a look at @databricks's new open source 132 billion model called DBRX! 1) Merged attention QKV clamped betw (-8, 8) 2) Not RMS Layernorm - now has mean removal unlike Llama 3) 4 active experts / 16. Mixtral 2/8 experts. 4) @OpenAI's TikToken tokenizer 100K. Llama splits…
RewardBench updates: * Reduced weight of prior sets to 50% so people can't easily gamify by training on Anthropic HH / Summarize etc * Models: Archangel KTO suite from @ContextualAI and @ethayarajh, showing scaling of Llama 1 and Pythia suite on constant data. The first DPO like…
Open source AI is a net win for developers, businesses, and humanity.
Large DPO dataset with >200k rows 1. Score popular datasets with "Self-Alignment with Instruction Backtranslation" prompt out of 5 2. Generate accepted_pairs (score 5) for rows with scores 1,2,3 using gpt-4-0125-preview 3. Generate rejected_pairs (score 2,1) for rows with score…
From Gemma Technical Report
From Gemma Technical Report https://t.co/V8cWp9Nrec
Holly Kennedy @HollyKenne99757
109 Followers 3K Following
Theasisoos @theasisoos46761
0 Followers 190 Following
Laughing Elf @elf_laughi67259
22 Followers 254 Following
Markus Junginger @greenrobot_de
1K Followers 369 Following Distributed and on-device data/AI. Cofounder/CTO @objectbox_io.
Aexyn @Aexyn
0 Followers 1K Following
David Garay @DavidGa81060671
5 Followers 37 Following
Totatoo @Totatoo380268
2 Followers 392 Following My hobbies are reading, food and sports. I like cats😘 I like to meet new friends while traveling🎉🎉🎉
McFrore @McFrore5lWWFYz
2 Followers 169 Following
Ron Williams @McclaneDet
928 Followers 604 Following Founder Kindo (Usable Machines), LP First Close. CSO 3X at Bird, Clover Health, Riot Games, & Founder Zeevex. Veteran.
Max` @Max292618236199
5 Followers 36 Following
anniewu @anniewu1214
0 Followers 126 Following
Alpay Ariyak @AlpayAriyak
1K Followers 2K Following AI @RunPod_io | Lead: @OpenChatDev (600k+ downloads on HuggingFace🤗)
NOVASPARK @NOVASPARKXX
117 Followers 2K Following
huansong @huansong514
10 Followers 173 Following
Bernard Wanyama @bmwanyama
1K Followers 4K Following Life is short, play more.... and make a difference while you are at it! Cyber Security, Cloud, IoT, People Development. SYNTECH, ISACA, Toastmasters
Kyle Mistele 🏴�.. @0xblacklight
613 Followers 780 Following Product // Engineering // OSCP // My other computer is your computer // Opinions are not my employer's.
Muharrem AYICI @MuAyI
64 Followers 911 Following
EP @EP225654
167 Followers 5K Following
Redjojovic @redjojovic
81 Followers 207 Following
Junyang Lin @JustinLin610
5K Followers 1K Following Chief Evangelist Officer of Qwen Team & OpenDevin, building LLM and LMM. Now @Alibaba_Qwen . Previously @PKU1898 LANCO group. ❤️ 🍵 ☕️ 🍷 🥃
Knut Jägersberg @JagersbergKnut
6K Followers 5K Following Content Strategy & AI @[email protected] https://t.co/xnBUK02hWS
Yash @Yash11386432
3 Followers 53 Following
Ashutosh Mehra @ashutoshmehra
2K Followers 5K Following Senior Principal Scientist at Adobe. Working on Acrobat AI Assistant, LLMs, and document ML.
Ahmed Morsi @eramax
29 Followers 656 Following
gemitarmy @kume0011
348 Followers 1K Following
John Stalcup @John_Stalcup_
295 Followers 776 Following Senior @Google Engineer, Compiler Developer, Android Location, AI enthusiast, hobby coding addict
Cory Watts @worycatts
132 Followers 105 Following
Darshan @DarshanJay
4 Followers 132 Following
Tuyen Huynh @hntuyen
48 Followers 1K Following
Geronimo @Geronimo_AI
803 Followers 399 Following LLM enthusiast 🚀 failing fast, learning fast. sharing it all on X and Medium
Chris T. N. @chris_t_ng
118 Followers 1K Following #nlp #deeplearning @Microsoft, previously @Samsung , @_skyhive
zirui @zirui3
42 Followers 976 Following
Eva Louise Marie Gabr.. @e681554349
8 Followers 3K Following
Kadir Nar @kadirnar_ai
3K Followers 2K Following 👨💻 Generative AI Engineer @AdCreativeai 🤗 https://t.co/xbyPJqUJWl
Philipp Schmid @_philschmid
16K Followers 656 Following Tech Lead and LLMs at @huggingface 👨🏻💻 🤗 AWS ML Hero 🦸🏻 | Cloud & ML enthusiast | 📍Nuremberg | 🇩🇪 https://t.co/l1ppq3q3hk
qnguyen3 @stablequan
3K Followers 1K Following Multimodal | Synthetic Data | Multimodal Lead at Ontocord AI
Omar Sanseviero @osanseviero
32K Followers 2K Following Chief Llama Officer @huggingface 🦙 Founder @AI_Learners. Xoogler (SWE @Google Assistant, 20% PM TF Graphics). 100% Hacker Llama🇵🇪🇲🇽
Alpay Ariyak @AlpayAriyak
1K Followers 2K Following AI @RunPod_io | Lead: @OpenChatDev (600k+ downloads on HuggingFace🤗)
Knut Jägersberg @JagersbergKnut
6K Followers 5K Following Content Strategy & AI @[email protected] https://t.co/xnBUK02hWS
Dwarkesh Patel @dwarkesh_sp
55K Followers 703 Following Being pretrained Host of Dwarkesh Podcast https://t.co/3SXlu7fy6N https://t.co/rEhnfYywXY https://t.co/hQfIWdM1Un
Ahmad Al-Dahle @Ahmad_Al_Dahle
4K Followers 53 Following #Girldad of twins. Leading GenAI @ Meta (llama, imagine, meta ai and more)
Tony Z. Zhao @tonyzzhao
13K Followers 784 Following CS PhD student @Stanford. Aspiring full-stack roboticist. Prev Deepmind, Tesla, GoogleX, Berkeley.
Yi Tay @YiTayML
29K Followers 97 Following chief scientist / cofounder @RekaAILabs 🫠 past: research scientist @google brain 🤯 currently learning to be a dad 🍼
George Grigorev @iamgrigorev
2K Followers 538 Following formerly generative ml @ snap, global talent interested in llms
udio @udiomusic
29K Followers 0 Following
mrfakename @realmrfakename
840 Followers 70 Following LLMs, TTS, & Open Source https://t.co/PIhamCNjhp
Grant Sanderson @3blue1brown
366K Followers 362 Following Pi creature caretaker. Contact/faq: https://t.co/brZwdQfdif
Thomas Wolf @Thom_Wolf
69K Followers 4K Following Co-founder and CSO @HuggingFace - open-source and open-science
Daniel van Strien @vanstriendaniel
3K Followers 1K Following Machine Learning Librarian @huggingface 🤗 | Championing Open Science & ML | Sharing the latest ML datasets 🌟 | Tips for mastering the HF Hub
Maxime Labonne @maximelabonne
13K Followers 440 Following Staff ML Scientist @LiquidAI_ • Author of Hands-On Graph Neural Networks https://t.co/Q8victWUmR
vLLM @vllm_project
856 Followers 11 Following A high-throughput and memory-efficient inference and serving engine for LLMs
Wei-Lin Chiang @infwinston
3K Followers 852 Following CS PhD student at UC Berkeley. co-lead of Chatbot Arena @lmsysorg
Chip Huyen @chipro
92K Followers 442 Following Data processing on GPUs @VoltronData Designing ML Systems: https://t.co/G81hL2dWmr @designmlsys #AI x #GPU
Lex Fridman @lexfridman
3.5M Followers 126 Following Host of Lex Fridman Podcast. Interested in robots and humans.
Lianmin Zheng @lm_zheng
4K Followers 439 Following CS Ph.D. @ UC Berkeley. Creator of Alpa, Vicuna, and Chatbot Arena. @lmsysorg
Exploding Topics @explodingtopics
32K Followers 199 Following Discover rapidly growing trends before they take off.
Daniel Han @danielhanchen
7K Followers 945 Following Building @UnslothAI. Finetune LLMs 30x faster https://t.co/aRyAAgKOR7. Prev ML at NVIDIA. Hyperlearn used by NASA. I like maths, making code go fast
Kyle Corbitt @corbtt
6K Followers 135 Following Currently building @OpenPipeAI. Formerly @ycombinator, @google. I am always down to go on a quest.
Blume Ventures @BlumeVentures
53K Followers 444 Following Backing the next wave of revolutionary ventures! Newsletters : https://t.co/1FSQSIFrz5 and https://t.co/HVtA6HTWM4
Richard Socher @RichardSocher
101K Followers 970 Following CEO @youSearchEngine Investing at @aixventuresHQ Before: Stanford Adj Prof in AI/NLP, Chief Scientist at Salesforce, MetaMind
Nikhil Kamath @nikhilkamathcio
335K Followers 392 Following Co-Founder, Zerodha | True Beacon | Gruhas
Pratik Desai @chheplo
9K Followers 708 Following 🌾 KissanAI - Pioneering Vernacular AgriCoPilot Platform with Agri Vertical Model (Dhenu) for Climate Resilient Agriculture (Kissan=Farmer)
Prayank Swaroop @prayanks
6K Followers 867 Following Human Being. Indian. And startup investor @accel. (Views expressed personal)
Nous Research @NousResearch
19K Followers 29 Following The AI Accelerator Company. https://t.co/vrD0aDJeto
Cognition @cognition_labs
125K Followers 19 Following Makers of Devin, the first AI software engineer. We are an applied AI lab focused on reasoning, and code is just the beginning. Join us: https://t.co/tpfZwEwGiq
Wing Lian (caseus) @winglian
9K Followers 2K Following @axolotl_ai OSS maintainer. Axolotl AI founder. AI/ML tinkerer. Building tools for everyone. ☕ https://t.co/3ni1V4rI9w
Theophile Gervet @theo_gervet
1K Followers 482 Following Accelerating open-source AI @MistralAI. Past: @Meta AI, PhD @SCSatCMU
Devendra Chaplot @ IC.. @dchaplot
8K Followers 365 Following Building next-gen AI at @MistralAI. Past: Research Scientist at Facebook AI Research. Ph.D. @SCSatCMU, BTech @iitbombay CS.
Sandeep @sandeep1337
2K Followers 594 Following Researcher @MistralAI | Previously @NVIDIA | PhD @MILAMontreal | Masters @SCSatCMU | Intern at @MSFTResearch @facebookai @element_ai @DescriptApp
the tiny corp @__tinygrad__
33K Followers 61 Following We make tinygrad. Our mission is to commoditize the petaflop.
Nathan Lambert @natolambert
25K Followers 693 Following Figuring out AI @allen_ai, "rl boi" DM me papers. Writes @interconnectsai, talks @retortai Has phd and some credentials
Logan Kilpatrick @OfficialLoganK
92K Followers 2K Following Lead product for @Google AI Studio and working on the Gemini API, helping developers build with AI, my views!
Marc Andreessen 🇺�.. @pmarca
1.4M Followers 24K Following Techno-optimist. E/acc. Technology brother. Move Fast and Make Things. p(Doom) = 0; p(“1984”) = not 0.
OpenAI Developers @OpenAIDevs
73K Followers 0 Following Official @OpenAI account for anyone building on our APIs. Join us in building the future of AI. We ❤️ developers!
Nikita Bier @nikitabier
323K Followers 2K Following I make apps grow really fast. founder @gasappteam (acq by discord), ex-founder @thetbhapp (acq by facebook), ex-new products @meta
Vinod Khosla @vkhosla
633K Followers 575 Following entrepreneurship zealot, grounded technology possibilist, believer in the power of ideas, passionate about sustainability & impact
Ravi Theja @ravithejads
3K Followers 672 Following Developer Advocate Engineer at @llama_index (LlamaIndex)
Munish Kumar @kumar_munish_
58 Followers 238 Following
Important Update: WhiteRabbitNeo & Kindo AI Today, I’m happy to announce that @KindoAI has acquired @WhiteRabbitNeos. When I introduced the 1st AI model for WhiteRabbitNeo in December 2023, I mentioned that someone needed to build an AI-focused on offensive cybersecurity and…
For anyone looking for pointers, the biggest trick for me was to get the prompt template right, and the learning rate. I used @ram_chandalada's PR on Axolotl to sort out the prompt template: github.com/OpenAccess-AI-…
OpenAI is a Nvidia wrapper Nvidia is a TSMC wrapper TSMC is an ASML wrapper ASML is a Zeiss wrapper Congratulations everyone you just discovered how a technologically advanced economy operates.
128k context? Almost. The released 64k already has the rope theta set to 8m. I ran the previous test at 2M. Here's the needle in a haystack for rope_theta=8M. Not sure why it is failing at 90% depth, but my hypothesis is that it's related to data distribution. I think if we…
I'm up to 96k context for Llama 3 8B. Using PoSE, we did continued pre-training of the base model w 300M tokens to extend the context length to 64k. From there we increased the RoPE theta to further attempt to extend the context length. 🧵
Money can't buy happiness. Just like an H100. H100 = happiness.
Feel free to try this Qwen1.5-110B model preview! I hope you enjoy it! We will release the model weights soon! huggingface.co/spaces/Qwen/Qw…
i’m assembling a team
@osanseviero We chose to release yesterday only because of your post :D
@mattshumer_ Hey Matt, appreciate you bringing this to our attention. We haven't modified any of the Claude 3 models since we launched them. On claude.ai, there's currently two layers that may contribute to perceived model performance: our T&S measures (standard mechanisms…
Well we had a nice run, humans.
fume scores a staggering 18.3% on SWE-bench. unassisted. this means fume can successfully solve almost a fifth of real life issues from various open source project.
Trends for United States
You might like
