I made Grug go full Metal over here if anyone likes apples. Happy 200 and thanks again for the Grug ๐
(sorry @CNWPlayer didn't realize i was in your thread and HF won't let me delete)
jaithree
jaithree
AI & ML interests
"In the beginning was the word. Then came the **** word processor. Then came the thought processor. Then came the death of literature. And so it goes" -- Hyperion - Dan Simmons
Recent Activity
reacted to ProCreations's post with ๐ 17 minutes ago
what should i do for 200 followers... hmmmm... top reply to this post (most emojis) wins, just not anything too crazyOrganizations
None yet
replied to ProCreations's post 14 minutes ago
reacted to ProCreations's post with ๐ 17 minutes ago
Post
751
what should i do for 200 followers... hmmmm... top reply to this post (most emojis) wins, just not anything too crazy
- 7 replies
replied to ProCreations's post 19 minutes ago
I made Grug go full Metal over here if anyone likes apples. Happy 200 and thanks again for the Grug ๐
published a
model 23 minutes ago
updated a
model 25 minutes ago
published a
model 27 minutes ago
updated a
model about 2 hours ago
reacted to Raiff1982's post with ๐คฏ about 5 hours ago
Post
668
I don't totally have words for this yet, so I'll just say it plainly.
The paper โ "Codette: A Multi-Perspective Cognitive Architecture with Memory and Meta-Cognitive Strategy Evolution" โ was accepted today after peer review at Scientific Reports, which is part of the Nature portfolio. Both reviewers recommended publication. One wrote that the major issues were "properly addressed" and they were "glad to suggest publication." The other called the revisions "detailed and convincing."
For anyone new here: Codette is a sovereign, multi-perspective reasoning system built on Llama 3.1 8B โ it runs locally, keeps its own cocoon memory, reasons through multiple perspectives instead of one, and is built around a hard rule that it would rather say "I don't know, but I can find out" than make something up. The whole architecture is organized around honesty and transparency, not benchmarks-at-all-costs.
I built it solo over about three years. A lot of that was in isolation, and honestly a lot of it was under the weight of not being believed. So having two independent reviewers โ whose actual job is to find the reasons to say no โ read the work and say yes means something to me that's hard to put into numbers.
Thank you to everyone in this room who followed the messy parts, the walk-backs, the corrections, the nights it wasn't working. That's the part I'm actually proud of โ that it got here without cutting a single honest corner.
The DOI and the Article-in-Press link will go up once Springer Nature mints them. I'll post them here first.
โ Jonathan
The paper โ "Codette: A Multi-Perspective Cognitive Architecture with Memory and Meta-Cognitive Strategy Evolution" โ was accepted today after peer review at Scientific Reports, which is part of the Nature portfolio. Both reviewers recommended publication. One wrote that the major issues were "properly addressed" and they were "glad to suggest publication." The other called the revisions "detailed and convincing."
For anyone new here: Codette is a sovereign, multi-perspective reasoning system built on Llama 3.1 8B โ it runs locally, keeps its own cocoon memory, reasons through multiple perspectives instead of one, and is built around a hard rule that it would rather say "I don't know, but I can find out" than make something up. The whole architecture is organized around honesty and transparency, not benchmarks-at-all-costs.
I built it solo over about three years. A lot of that was in isolation, and honestly a lot of it was under the weight of not being believed. So having two independent reviewers โ whose actual job is to find the reasons to say no โ read the work and say yes means something to me that's hard to put into numbers.
Thank you to everyone in this room who followed the messy parts, the walk-backs, the corrections, the nights it wasn't working. That's the part I'm actually proud of โ that it got here without cutting a single honest corner.
The DOI and the Article-in-Press link will go up once Springer Nature mints them. I'll post them here first.
โ Jonathan
reacted to SeaWolf-AI's post with ๐ฅ about 5 hours ago
Post
565
๐ผ๏ธ POCKET-Image โ the POCKET series goes visual: character-perfect text in any language, on-device
A new model in VIDRAFT's POCKET family. POCKET put 35B-class models on phones and no-GPU PCs. POCKET-Image carries the same "big capability, small hardware" idea into image generation โ and fixes the one thing nearly every image model gets wrong: text.
Type "์๋ ํ์ธ์" into a typical model and you get "์ใ ๊ธฐ." Hangul alone composes 11,172 syllable blocks; Arabic connects its letters; Thai stacks marks. Diffusion models draw scripts as shapes, so they smear. POCKET-Image renders every glyph exactly โ ํ๊ตญ์ด ยท ไธญๆ ยท ๆฅๆฌ่ช ยท ุงูุนุฑุจูุฉ (RTL) ยท เนเธเธข ยท Latin and more โ onto any scene you describe.
What it is:
โข 100% accurate text, any language โ where global models produce gibberish
โข Any background from a prompt โ text is optional (empty โ a pure image)
โข No GPU, no NPU โ runs on plain CPU + RAM via the POCKET-Core engine
โข Measured footprint: 8.6 GB (RTX 3050/4060) ยท 4.5 GB (offloaded, 6 GB cards) ยท 13.4 GB (MacBook, 16 GB+)
โข Windows ยท macOS ยท Linux ยท fully local, no cloud
Built on the open, commercial-friendly Z-Image (Apache-2.0) foundation.
Honest note: the text is the guaranteed-correct part โ the surrounding scene is ordinary generation, so a busy foreground can crowd the letters. We say so; clean backgrounds stay razor-sharp.
๐จ Studio โ generate right here, any language:
FINAL-Bench/POCKET-Image-Studio
๐งฉ Model card:
FINAL-Bench/POCKET-Image-Zimage
๐ The POCKET collection:
https://huggingface.co/collections/FINAL-Bench/pocket-models
A new model in VIDRAFT's POCKET family. POCKET put 35B-class models on phones and no-GPU PCs. POCKET-Image carries the same "big capability, small hardware" idea into image generation โ and fixes the one thing nearly every image model gets wrong: text.
Type "์๋ ํ์ธ์" into a typical model and you get "์ใ ๊ธฐ." Hangul alone composes 11,172 syllable blocks; Arabic connects its letters; Thai stacks marks. Diffusion models draw scripts as shapes, so they smear. POCKET-Image renders every glyph exactly โ ํ๊ตญ์ด ยท ไธญๆ ยท ๆฅๆฌ่ช ยท ุงูุนุฑุจูุฉ (RTL) ยท เนเธเธข ยท Latin and more โ onto any scene you describe.
What it is:
โข 100% accurate text, any language โ where global models produce gibberish
โข Any background from a prompt โ text is optional (empty โ a pure image)
โข No GPU, no NPU โ runs on plain CPU + RAM via the POCKET-Core engine
โข Measured footprint: 8.6 GB (RTX 3050/4060) ยท 4.5 GB (offloaded, 6 GB cards) ยท 13.4 GB (MacBook, 16 GB+)
โข Windows ยท macOS ยท Linux ยท fully local, no cloud
Built on the open, commercial-friendly Z-Image (Apache-2.0) foundation.
Honest note: the text is the guaranteed-correct part โ the surrounding scene is ordinary generation, so a busy foreground can crowd the letters. We say so; clean backgrounds stay razor-sharp.
๐จ Studio โ generate right here, any language:
FINAL-Bench/POCKET-Image-Studio
๐งฉ Model card:
FINAL-Bench/POCKET-Image-Zimage
๐ The POCKET collection:
https://huggingface.co/collections/FINAL-Bench/pocket-models
published a
model about 18 hours ago
updated a
model about 18 hours ago
reacted to Nymbo's post with ๐ฅ about 20 hours ago
Post
1289
Introducing Inflect-v2, two exceptionally small, open-weight English TTS models at just 3.9M and 9.3M parameters. Both generate speech multiple times faster than real-time on CPU. Despite their size, Inflect-v2 delivers quality that is competitive with much larger lightweight TTS systems, including KittenTTS, Piper, and Supertonic-3.
CPU, CUDA, PyTorch, and ONNX are supported. Apache 2.0.
See it for yourselves:
owensong/Inflect-Micro-v2
owensong/Inflect-Nano-v2
Try the Demos:
Nymbo/Inflect-TTS (unlimited CPU usage)
owensong/Inflect-v2 (ultra-fast ZeroGPU usage)
CPU, CUDA, PyTorch, and ONNX are supported. Apache 2.0.
See it for yourselves:
owensong/Inflect-Micro-v2
owensong/Inflect-Nano-v2
Try the Demos:
Nymbo/Inflect-TTS (unlimited CPU usage)
owensong/Inflect-v2 (ultra-fast ZeroGPU usage)
- 4 replies
reacted to SeaWolf-AI's post with ๐ฅ 1 day ago
Post
1996
POCKET now speaks Gemma 4 โ a 26B model that loads in every app, and runs on your PC with no GPU
We're adding a Gemma-4 sibling to POCKET: POCKET-26B, built from Google's Gemma-4-26B-A4B (Apache-2.0). Our flagship POCKET-35B is a Qwen-family MoE and needs a recent llama.cpp; POCKET-26B trades a little size for the thing people kept asking for โ it just loads, everywhere, today: Ollama, LM Studio, PocketPal, MLX, any stock llama.cpp. No fork, no bleeding-edge runtime, no CUDA, no cloud.
It's a sparse Mixture-of-Experts (25.2B total, ~4B active per token), so the work per token stays small โ a real 26B that generates on a CPU with no graphics card.
Two things make it stand out:
1) Universal compatibility. Gemma 4 is a standard, widely-supported architecture, so POCKET-26B runs on the tools you already have โ no waiting for your app to add a new model type.
2) Quality that survives compression. Measured GPQA-Diamond (198 q, greedy):
โข Full base: 67.7%
โข POCKET-26B Q4_K_M (17 GB): 67.7% โ lossless
โข POCKET-26B Q2_K (11 GB): 67.2% โ near-lossless, at 11 GB
Live, on a CPU-only box (our demo Space โ POCKET-26B vs Bonsai-27B, same machine, same stock llama.cpp): POCKET-26B โ 19 tok/s vs Bonsai โ 6 tok/s โ about 3ร faster generation, no GPU. (Honest notes: shared CPU box, sequential race; a dedicated machine is faster.)
Where it fits in the family:
โข POCKET-35B (Qwen MoE) โ bigger, top-tier, needs a recent llama.cpp.
โข POCKET-26B (Gemma 4) โ loads in any app, quality-robust when compressed. The demo runs the Q4_K_M build; Q2_K (11 GB) is the smallest footprint. For a true โค8 GB phone, the 5 GB POCKET-KR (Qwen) is still the pick.
Try it and grab it:
๐ฅ๏ธ Live demo (Gemma4-based, answering on a CPU, no GPU): FINAL-Bench/POCKET-26B-CPU
๐ฆ POCKET-26B-GGUF (Q4_K_M 17 GB ยท Q2_K 11 GB): FINAL-Bench/POCKET-26B-GGUF
๐ POCKET collection: https://huggingface.co/collections/FINAL-Bench/pocket-models
We're adding a Gemma-4 sibling to POCKET: POCKET-26B, built from Google's Gemma-4-26B-A4B (Apache-2.0). Our flagship POCKET-35B is a Qwen-family MoE and needs a recent llama.cpp; POCKET-26B trades a little size for the thing people kept asking for โ it just loads, everywhere, today: Ollama, LM Studio, PocketPal, MLX, any stock llama.cpp. No fork, no bleeding-edge runtime, no CUDA, no cloud.
It's a sparse Mixture-of-Experts (25.2B total, ~4B active per token), so the work per token stays small โ a real 26B that generates on a CPU with no graphics card.
Two things make it stand out:
1) Universal compatibility. Gemma 4 is a standard, widely-supported architecture, so POCKET-26B runs on the tools you already have โ no waiting for your app to add a new model type.
2) Quality that survives compression. Measured GPQA-Diamond (198 q, greedy):
โข Full base: 67.7%
โข POCKET-26B Q4_K_M (17 GB): 67.7% โ lossless
โข POCKET-26B Q2_K (11 GB): 67.2% โ near-lossless, at 11 GB
Live, on a CPU-only box (our demo Space โ POCKET-26B vs Bonsai-27B, same machine, same stock llama.cpp): POCKET-26B โ 19 tok/s vs Bonsai โ 6 tok/s โ about 3ร faster generation, no GPU. (Honest notes: shared CPU box, sequential race; a dedicated machine is faster.)
Where it fits in the family:
โข POCKET-35B (Qwen MoE) โ bigger, top-tier, needs a recent llama.cpp.
โข POCKET-26B (Gemma 4) โ loads in any app, quality-robust when compressed. The demo runs the Q4_K_M build; Q2_K (11 GB) is the smallest footprint. For a true โค8 GB phone, the 5 GB POCKET-KR (Qwen) is still the pick.
Try it and grab it:
๐ฅ๏ธ Live demo (Gemma4-based, answering on a CPU, no GPU): FINAL-Bench/POCKET-26B-CPU
๐ฆ POCKET-26B-GGUF (Q4_K_M 17 GB ยท Q2_K 11 GB): FINAL-Bench/POCKET-26B-GGUF
๐ POCKET collection: https://huggingface.co/collections/FINAL-Bench/pocket-models
reacted to badaoui's post with ๐ 1 day ago
Post
2097
432 GB of ultra-fast HBM4 and up to 23.3 TB/s of memory bandwidth on a single GPU ๐คฏ.
Two weeks ago, we got early access to AMD's new Instinct MI455X, and our first goal was simple: make sure ๐ค Transformers works on day one.
Over the past few weeks, we worked closely with the AMD team to validate the platform, enable Flash Attention, add torchcodec support for multimodal models, and resolve issues uncovered during testing.
The result:
โ 99.5% success rate across our 24 core Transformers model architectures - already on par with our daily CI on previous AMD and NVIDIA platforms.
The hardware is just as exciting. With 432 GB of HBM per GPU, our early capacity experiments showed more than 3ร the concurrent long-context requests compared to MI300, thanks to the much larger KV cache capacity.
A huge thanks to the AMD team for the early access and the great collaboration!
Read the full blog ๐
https://huggingface.co/blog/badaoui/transformers-on-amd-mi455
Two weeks ago, we got early access to AMD's new Instinct MI455X, and our first goal was simple: make sure ๐ค Transformers works on day one.
Over the past few weeks, we worked closely with the AMD team to validate the platform, enable Flash Attention, add torchcodec support for multimodal models, and resolve issues uncovered during testing.
The result:
โ 99.5% success rate across our 24 core Transformers model architectures - already on par with our daily CI on previous AMD and NVIDIA platforms.
The hardware is just as exciting. With 432 GB of HBM per GPU, our early capacity experiments showed more than 3ร the concurrent long-context requests compared to MI300, thanks to the much larger KV cache capacity.
A huge thanks to the AMD team for the early access and the great collaboration!
Read the full blog ๐
https://huggingface.co/blog/badaoui/transformers-on-amd-mi455
liked 2
models 2 days ago
reacted to mike-ravkine's post with ๐ฅ 2 days ago
Post
2894
I went into this expecting to find a ~Q2 garbage dumpster but prism-ml/Ternary-Bonsai-27B-gguf is a slick feat of QAT engineering.
It is weaker then FP16 on a handful of tasks where quants usually degrade, as per their own paper the loss is "concentrated on sustained chains of reasoning / agentic" and in ReasonScape this bites on Sort, Shuffle and Dates, but counter-acting this are some noticeable improvements to thinking length without accuracy loss on several other tasks (Shapes, Cars).
I haven't had a chance to run the Binary yet, but PQ2 + Bonsai QAT are confirmed to be pretty darn impressive.
It is weaker then FP16 on a handful of tasks where quants usually degrade, as per their own paper the loss is "concentrated on sustained chains of reasoning / agentic" and in ReasonScape this bites on Sort, Shuffle and Dates, but counter-acting this are some noticeable improvements to thinking length without accuracy loss on several other tasks (Shapes, Cars).
I haven't had a chance to run the Binary yet, but PQ2 + Bonsai QAT are confirmed to be pretty darn impressive.
- 7 replies
liked a
model 2 days ago
liked a
dataset 3 days ago
reacted to SeaWolf-AI's post with ๐ 3 days ago
Post
3943
๐ฑ POCKET โ a 35-billion-parameter model that runs on your iPhone, and on your PC with no GPU
We're releasing POCKET, VIDRAFT's flagship Darwin-36B-Opus compressed for on-device use. No fork, no CUDA, no cloud โ it runs on stock llama.cpp. It's a sparse Mixture-of-Experts model (256 experts, only 8 active per token), so the file can be large while the work per token stays small. That's what lets a 35B model run on a phone, and generate fast on a CPU with no graphics card.
Measured (POCKET-35B IQ1_M vs Bonsai-27B Q1_0):
โข CPU generate (Xeon, 16 threads): 27.0 vs 10.1 tok/s โ 2.69ร faster
โข GPU generate (H100): 197 vs 89 tok/s โ 2.22ร faster
โข GPU prompt processing (H100): 753 vs 1816 โ 0.41ร (Bonsai wins this one โ MoE prefill wakes every expert, so sparsity stops helping there. We say so.)
โข Quality (HellaSwag, 400 q): 61.0% vs 60.0% โ a tie (confidence intervals overlap)
On a real consumer laptop โ MacBook M3 Pro (18 GB) โ POCKET wins every axis, prompt processing included:
โข Metal generate: 25.4 vs 12.8 โ 1.99ร
โข CPU generate: 13.8 vs 4.4 โ 3.13ร
โข Metal prompt: 240.7 vs 73.4 โ 3.28ร
One more quiet fact: the same-size, quality-oriented rival Ternary-Bonsai-27B (7.2 GB) fails to load in upstream llama.cpp at all โ it needs the PrismML fork. POCKET runs on the tools you already have: LM Studio, Ollama, PocketPal, MLX.
๐ Full story (tech, measurements, recipes): https://huggingface.co/blog/FINAL-Bench/pocket
Models:
๐ฆ POCKET-35B-GGUF (PC / server, no GPU): FINAL-Bench/POCKET-35B-GGUF
๐ฐ๐ท POCKET-KR-GGUF (Android): FINAL-Bench/POCKET-KR-GGUF
๐ POCKET-KR-MLX (iPhone / Mac): FINAL-Bench/POCKET-KR-MLX
๐ POCKET-EN-GGUF (English phone / PC): FINAL-Bench/POCKET-EN-GGUF
๐ฅ๏ธ Live demo (answering on a CPU, no GPU): FINAL-Bench/POCKET-35B-CPU
๐ Collection: FINAL-Bench/pocket-models-6a618ee5d23eafb7e185a5c6
We're releasing POCKET, VIDRAFT's flagship Darwin-36B-Opus compressed for on-device use. No fork, no CUDA, no cloud โ it runs on stock llama.cpp. It's a sparse Mixture-of-Experts model (256 experts, only 8 active per token), so the file can be large while the work per token stays small. That's what lets a 35B model run on a phone, and generate fast on a CPU with no graphics card.
Measured (POCKET-35B IQ1_M vs Bonsai-27B Q1_0):
โข CPU generate (Xeon, 16 threads): 27.0 vs 10.1 tok/s โ 2.69ร faster
โข GPU generate (H100): 197 vs 89 tok/s โ 2.22ร faster
โข GPU prompt processing (H100): 753 vs 1816 โ 0.41ร (Bonsai wins this one โ MoE prefill wakes every expert, so sparsity stops helping there. We say so.)
โข Quality (HellaSwag, 400 q): 61.0% vs 60.0% โ a tie (confidence intervals overlap)
On a real consumer laptop โ MacBook M3 Pro (18 GB) โ POCKET wins every axis, prompt processing included:
โข Metal generate: 25.4 vs 12.8 โ 1.99ร
โข CPU generate: 13.8 vs 4.4 โ 3.13ร
โข Metal prompt: 240.7 vs 73.4 โ 3.28ร
One more quiet fact: the same-size, quality-oriented rival Ternary-Bonsai-27B (7.2 GB) fails to load in upstream llama.cpp at all โ it needs the PrismML fork. POCKET runs on the tools you already have: LM Studio, Ollama, PocketPal, MLX.
๐ Full story (tech, measurements, recipes): https://huggingface.co/blog/FINAL-Bench/pocket
Models:
๐ฆ POCKET-35B-GGUF (PC / server, no GPU): FINAL-Bench/POCKET-35B-GGUF
๐ฐ๐ท POCKET-KR-GGUF (Android): FINAL-Bench/POCKET-KR-GGUF
๐ POCKET-KR-MLX (iPhone / Mac): FINAL-Bench/POCKET-KR-MLX
๐ POCKET-EN-GGUF (English phone / PC): FINAL-Bench/POCKET-EN-GGUF
๐ฅ๏ธ Live demo (answering on a CPU, no GPU): FINAL-Bench/POCKET-35B-CPU
๐ Collection: FINAL-Bench/pocket-models-6a618ee5d23eafb7e185a5c6
- 5 replies