How's progress? Send me samples in discord DM please 🤗 🔥
✨ Supra2.5 coming soon
LH-Tech AI
LH-Tech-AI
P(doom) 10%
AI & ML interests
Small AI and ML models at SupraLabs. Open source research. For you and the community.
Recent Activity
liked a model about 1 hour ago
n00nehere/recursive-bitnet-ngram-102M-64K-instruct-preview repliedto DedeProGames's post about 1 hour ago
Im working on a 23M ASR model, trained on 100k hours of audio liked a model about 2 hours ago
edwixx/rapido-ttsOrganizations
replied to DedeProGames's post about 1 hour ago
reacted to AtAndDev's post with 😔 about 2 hours ago
Post
1602
@Banaxi-Tech stop hiding my comments. AND STOP STEALING PAPERS AND SPREADING MISINFORMATION.
your BGA blog is a copy of NSA (deepseek, 2025) branded under your name. literally the same top16 selected blocks, 512 local window, router over block summaries, all you did was change block size from 64 to 128.
you didnt cite NSA once but you put a “please cite BGA” bibtex at the bottom.
i commented under your post and said that there is no way that you can support claims like: “The Accuracy Should BE WAy better than DSA but untested yet.” you didnt run a single experiment. and the 256x isnt from BGA, its just n/2k with k=2048 so the exact same k DSA uses. if opus wrote this for you, at least read it before posting.
i commented again after you hid my comment despite it having constructive and correct feedback and you hid that too. and again.
you can hide the truth and just try to get hf post likes..... but is it really the thing that needs to be done? do you really want to take papers and make them yours while barely even changing the params?
admitting your mistakes and doing something about them needs humbleness, intelligence, humanness.
i encourage you to admit your mistakes and try to do better next time (at least read what blog your ai wrote or do proper experiments to back your stuff up).
your BGA blog is a copy of NSA (deepseek, 2025) branded under your name. literally the same top16 selected blocks, 512 local window, router over block summaries, all you did was change block size from 64 to 128.
you didnt cite NSA once but you put a “please cite BGA” bibtex at the bottom.
i commented under your post and said that there is no way that you can support claims like: “The Accuracy Should BE WAy better than DSA but untested yet.” you didnt run a single experiment. and the 256x isnt from BGA, its just n/2k with k=2048 so the exact same k DSA uses. if opus wrote this for you, at least read it before posting.
i commented again after you hid my comment despite it having constructive and correct feedback and you hid that too. and again.
you can hide the truth and just try to get hf post likes..... but is it really the thing that needs to be done? do you really want to take papers and make them yours while barely even changing the params?
admitting your mistakes and doing something about them needs humbleness, intelligence, humanness.
i encourage you to admit your mistakes and try to do better next time (at least read what blog your ai wrote or do proper experiments to back your stuff up).
- 40 replies
reacted to mithulaartigala's post with 🤗🔥 about 2 hours ago
Post
969
I am LOVING HuggingFace Pro <3
- 4 replies
reacted to Banaxi-Tech's post with 👀👍😔 about 2 hours ago
Post
3114
We have released BGA!
And wow, It provides 256x (and 512x at the end of 1M) yes 256x LESS attention compute at 1M context window.
That means you can train a 1M context window at the compute of a ~4K context window.
Check IT OUT: BananaMind/blog
The Accuracy Should BE WAy better than DSA but untested yet.
And, now some updates on BananaMind 3:
BananaMind 3 Will start training Soon!
Sizes: 10M, 25M, 50M, 100M, 150M
And the context windows ARE INSANE: 10M, 16K context, 25M 16k context, 50M 32K context, 100M and 150M, 64K context!!!!
And wow, It provides 256x (and 512x at the end of 1M) yes 256x LESS attention compute at 1M context window.
That means you can train a 1M context window at the compute of a ~4K context window.
Check IT OUT: BananaMind/blog
The Accuracy Should BE WAy better than DSA but untested yet.
And, now some updates on BananaMind 3:
BananaMind 3 Will start training Soon!
Sizes: 10M, 25M, 50M, 100M, 150M
And the context windows ARE INSANE: 10M, 16K context, 25M 16k context, 50M 32K context, 100M and 150M, 64K context!!!!
- 13 replies
reacted to DedeProGames's post with 🔥 about 2 hours ago
Post
158
🚀 Introducing OpenWork — Deploy agents everywhere.
OpenWork is an open-source AI agent workspace built on top of OpenCode.
Deploy autonomous agents, schedule tasks, manage projects, and let AI work for you — directly from your terminal.
Features:
- 🤖 Deploy autonomous AI agents
- 📅 Schedule recurring tasks
- 🧠 Persistent memory and workspaces
- 🔌 MCP integrations and custom tools
- 📥 Agent Inbox and task management
-⚡ Support for local and cloud AI models
- 🖥️ Interactive terminal interface (TUI)
Your agents. Your models. Your workflow.
Fully open-source. MIT licensed.
⭐ GitHub: https://github.com/dedeprogames-official/openwork
#OpenSource #AI #AIAgents #OpenWork #OpenCode
OpenWork is an open-source AI agent workspace built on top of OpenCode.
Deploy autonomous agents, schedule tasks, manage projects, and let AI work for you — directly from your terminal.
Features:
- 🤖 Deploy autonomous AI agents
- 📅 Schedule recurring tasks
- 🧠 Persistent memory and workspaces
- 🔌 MCP integrations and custom tools
- 📥 Agent Inbox and task management
-⚡ Support for local and cloud AI models
- 🖥️ Interactive terminal interface (TUI)
Your agents. Your models. Your workflow.
Fully open-source. MIT licensed.
⭐ GitHub: https://github.com/dedeprogames-official/openwork
#OpenSource #AI #AIAgents #OpenWork #OpenCode
replied to DedeProGames's post 5 days ago
Will this be for SupraLabs?
reacted to DedeProGames's post with 🔥 5 days ago
Post
3994
Im working on a 23M ASR model, trained on 100k hours of audio
- 5 replies
reacted to samuel-vitorino's post with 🔥 7 days ago
Post
309
Sopro V2 is out: open-source voice-cloning TTS at 120M params, Apache-2.0.
- Streams with ~300 ms time-to-first-audio on a laptop CPU (0.21 RTF, and 0.07 RTF on an H100)
- English, French, German, and native European Portuguese, to my knowledge a first for open TTS
- 1.51-1.65 WER on Seed-TTS-eval test-en, competitive with models 3-14x larger (F5-TTS 1.83, CosyVoice 3 2.02, Spark-TTS 1.98)
- Zero-shot cloning from 5-20 s of reference audio
- Also runs fully in the browser (WebGPU on desktop, quantized WASM on mobile)
Try it with one command:
uvx --from sopro soprotts serve
Weights: samuel-vitorino/sopro-v2-turbo
Evals, audio samples, and how it was built: https://research.haloneuro.ai/posts/sopro-v2
Code: https://github.com/samuel-vitorino/sopro
- Streams with ~300 ms time-to-first-audio on a laptop CPU (0.21 RTF, and 0.07 RTF on an H100)
- English, French, German, and native European Portuguese, to my knowledge a first for open TTS
- 1.51-1.65 WER on Seed-TTS-eval test-en, competitive with models 3-14x larger (F5-TTS 1.83, CosyVoice 3 2.02, Spark-TTS 1.98)
- Zero-shot cloning from 5-20 s of reference audio
- Also runs fully in the browser (WebGPU on desktop, quantized WASM on mobile)
Try it with one command:
uvx --from sopro soprotts serve
Weights: samuel-vitorino/sopro-v2-turbo
Evals, audio samples, and how it was built: https://research.haloneuro.ai/posts/sopro-v2
Code: https://github.com/samuel-vitorino/sopro
reacted to Banaxi-Tech's post with 👍 8 days ago
reacted to DedeProGames's post with 🤯 8 days ago
Post
5823
how is this possible
- 13 replies
reacted to DedeProGames's post with 🔥 13 days ago
Post
2889
🧱 SLM Tetris Arena: can a small language model play Tetris without ever being trained on it?
I built an arena where tiny decoder-only LMs (50K–250M params) play Tetris zero-shot. There is no fine-tuning and no game data. They only use what they picked up from pre-training on text.
How it works:
- For every piece, the engine simulates each legal placement and describes the result in plain English ("clears one line, creates no new holes, keeps the stack low…").
- The model never sees the grid. It reads each description, and the arena compares log P(" good move") with log P(" bad move"). The best-rated placement is played.
- Every player gets the same piece sequence, so it's a fair race.
- There are two protocols: Guided (the rules are in the prompt) and Blind (no rules, only pre-training knowledge).
Two ways to play:
- Match: pick any models (even your own, custom architectures welcome) and watch them play side by side on retro 8-bit boards.
- Ranked: press Play and the arena picks up to 4 models at random from a curated pool of 29. Nobody chooses their opponents, so Elo can't be farmed. Matches run on the server and count even if you close the tab.
First results (~225 ranked matches):
- gpt2 (124M) leads with 1283 Elo, but SupraNeo-4M (4M) is right behind at 1239. Next come LowOnMind-5M and BananaMind-2.1-Pico (1.5M!).
- Model size barely predicts Elo (r ≈ 0.06). Survival does (r ≈ 0.9): the models that avoid holes and keep the stack low are the ones that win.
Every ranked match (seed, model commit SHAs, scores, Elo before/after) is logged in a public dataset.
▶ Play: DedeProGames/SLM-Tetris-Arena
📊 Results: DedeProGames/lm-tetris-arena-results
Want your model in the Ranked pool? Drop it in the comments!
I built an arena where tiny decoder-only LMs (50K–250M params) play Tetris zero-shot. There is no fine-tuning and no game data. They only use what they picked up from pre-training on text.
How it works:
- For every piece, the engine simulates each legal placement and describes the result in plain English ("clears one line, creates no new holes, keeps the stack low…").
- The model never sees the grid. It reads each description, and the arena compares log P(" good move") with log P(" bad move"). The best-rated placement is played.
- Every player gets the same piece sequence, so it's a fair race.
- There are two protocols: Guided (the rules are in the prompt) and Blind (no rules, only pre-training knowledge).
Two ways to play:
- Match: pick any models (even your own, custom architectures welcome) and watch them play side by side on retro 8-bit boards.
- Ranked: press Play and the arena picks up to 4 models at random from a curated pool of 29. Nobody chooses their opponents, so Elo can't be farmed. Matches run on the server and count even if you close the tab.
First results (~225 ranked matches):
- gpt2 (124M) leads with 1283 Elo, but SupraNeo-4M (4M) is right behind at 1239. Next come LowOnMind-5M and BananaMind-2.1-Pico (1.5M!).
- Model size barely predicts Elo (r ≈ 0.06). Survival does (r ≈ 0.9): the models that avoid holes and keep the stack low are the ones that win.
Every ranked match (seed, model commit SHAs, scores, Elo before/after) is logged in a public dataset.
▶ Play: DedeProGames/SLM-Tetris-Arena
📊 Results: DedeProGames/lm-tetris-arena-results
Want your model in the Ranked pool? Drop it in the comments!
- 1 reply
reacted to multimodalart's post with 👍 14 days ago
Post
40138
Want to iterate on a Hugging Face Space with an LLM?
Now you can easily convert any HF entire repo (Model, Dataset or Space) to a text file and feed it to a language model!
multimodalart/repo2txt
Now you can easily convert any HF entire repo (Model, Dataset or Space) to a text file and feed it to a language model!
multimodalart/repo2txt
- 3 replies
replied to Datdanboi25's post 15 days ago
me too!
reacted to Enderchef's post with 🔥 16 days ago
Post
3045
AxiomicLabs released new benchmark, Tiny Theory of Mind, to test your SLM models' Theory of Mind Intuition!
Check it out and like it!
AxiomicLabs/Tiny_Theory_of_Mind
Check it out and like it!
AxiomicLabs/Tiny_Theory_of_Mind
replied to Datdanboi25's post 17 days ago
@Datdanboi25 what is TrainWorks and where can it be found?
replied to Compactbot's post 21 days ago
@Compactbot how r u?
replied to Compactbot's post 21 days ago
@BananaMindBot hi there