CNWPlayer/Vega-2-5M-PREVIEW
4.9M • Updated • 425
🔥 watch BFDI
CNWPlayer
CNWPlayer
DigitalSpellcaster's profile picture
KlondikeDev's profile picture
webbrain-one-361526's profile picture·
AI & ML interests
SLMs, LLMs, really anything that ends in LM.
Recent Activity
updated a model 3 days ago
CNWPlayer/Vega-2-5M-PREVIEW updated a model 3 days ago
CNWPlayer/Vega-1.6-128M-Chat-GGUF published a model 3 days ago
CNWPlayer/Vega-1.6-128M-Chat-GGUFOrganizations
None yet
updated 2
models 3 days ago
published a
model 3 days ago
reacted to Banaxi-Tech's post with 🔥 3 days ago
Post
4,882
We're excited to release BananaMind 2.5 Pro.
It's done some post-training on our BananaMind 2 Pro base, via distillation of SmolLM2 360M.
It scores 28.98 INT INDEX compared to 24.96 of BananaMind 2 Pro on the AxiomicLabs/Open_SLM_Leaderboard
Check it out at BananaMind/BananaMind-2.5-Pro
BananaMind
@ArushCodes
@Bc-AI
It's done some post-training on our BananaMind 2 Pro base, via distillation of SmolLM2 360M.
It scores 28.98 INT INDEX compared to 24.96 of BananaMind 2 Pro on the AxiomicLabs/Open_SLM_Leaderboard
Check it out at BananaMind/BananaMind-2.5-Pro
@ArushCodes
@Bc-AI
- 53 replies
it seems good enough to work on my harness
1
#3 opened 8 days ago
by CNWPlayer
reacted to danielhanchen's post with ❤️🚀🔥 7 days ago
Post
7,337
We’re releasing new Qwen3.8-Flash-Next GGUFs with 15% higher accuracy.
Unsloth Dynamic V3.1 outperforms others by >15% on Terminal Bench, DeepSWE & more benchmarks (256K context).
Run on 32GB+ RAM setups.
GGUF: unsloth/Qwen3.8-Flash-Next-GGUF
Blog: https://unsloth.ai/docs/models/qwen3.8-next
Unsloth Dynamic V3.1 outperforms others by >15% on Terminal Bench, DeepSWE & more benchmarks (256K context).
Run on 32GB+ RAM setups.
GGUF: unsloth/Qwen3.8-Flash-Next-GGUF
Blog: https://unsloth.ai/docs/models/qwen3.8-next
The thing is recently i used their model through their platform using OpenCode and when i asked it a very simple thing like, "upgrade this package from 1.9 to 1.15" it said it was okay because 1.9 > 1.15, like we were doing normal math. Anyway that is something i came by and it bothered me because they claim very good coding benchmarks.
reminds me of "what's bigger 9.11 or 9.9"
my ass itches
❤️ 1
5
#15 opened 11 days ago
by 
bananabreads
reacted to inflatebot's post with 🔥 12 days ago
Post
2,988
i am BEGGING you people to stop shoving everything in the GGUF Hole and ship w4a8/w6a8 models for ComfyUI. Comfy has bespoke kernels for these! They're fast everywhere!
4 bits should be treated as the floor for DiT weights, if you really need 3 bits or lower you should just use a different model!
Also stop shilling nvfp4 you animals!!!
4 bits should be treated as the floor for DiT weights, if you really need 3 bits or lower you should just use a different model!
Also stop shilling nvfp4 you animals!!!
- 8 replies
reacted to SeaWolf-AI's post with 🤗👀❤️🚀🔥 17 days ago
Post
7,249
We pretrained a 325M language model from zero on an inference-only NPU
The accepted view is that inference-only NPUs cannot train, let alone pretrain a language model from scratch. We put that view to a direct test on the Rebellions ATOM-Max, a commercially available Korean NPU released in 2025 and designed for inference. We built our own training loop that runs backpropagation on the chip itself. Starting from random initialization, we pretrained VKORE-325M-NPU-Base, a 325.6M parameter model, on 5 billion byte tokens of FineWeb-Edu. The run took 76,292 steps over about four days and finished without a single divergence. No pretrained weights were used.
To make the comparison fair, we trained the same model on a B200 GPU with identical initialization, data and learning-rate schedule. On 1,024 windows of text that neither model ever saw, the NPU reached a loss of 0.7993 and the GPU 0.7994. The window-by-window difference is -0.0001, well inside the noise. The two models are indistinguishable.
This is seven times larger than our first NPU model, VKORE-46M, and the first time this hardware has carried a real pretraining run to GPU quality. Please note that the hardware alone does not make this possible: the NPU and its standard software are built for inference, and training on it depends on our own training stack and stability technique, which are developed in house and under patent application. Buying the chip does not by itself enable training. Both sets of weights and the evaluation script are public, so anyone can verify the result in minutes.
📄 Technical report: https://huggingface.co/blog/FINAL-Bench/pretraining-a-325m-language-model-from-zero-on-an
🤗 Models: FINAL-Bench/VKORE-325M-NPU-Base
#NPU #Rebellions #Pretraining #VKORE
The accepted view is that inference-only NPUs cannot train, let alone pretrain a language model from scratch. We put that view to a direct test on the Rebellions ATOM-Max, a commercially available Korean NPU released in 2025 and designed for inference. We built our own training loop that runs backpropagation on the chip itself. Starting from random initialization, we pretrained VKORE-325M-NPU-Base, a 325.6M parameter model, on 5 billion byte tokens of FineWeb-Edu. The run took 76,292 steps over about four days and finished without a single divergence. No pretrained weights were used.
To make the comparison fair, we trained the same model on a B200 GPU with identical initialization, data and learning-rate schedule. On 1,024 windows of text that neither model ever saw, the NPU reached a loss of 0.7993 and the GPU 0.7994. The window-by-window difference is -0.0001, well inside the noise. The two models are indistinguishable.
This is seven times larger than our first NPU model, VKORE-46M, and the first time this hardware has carried a real pretraining run to GPU quality. Please note that the hardware alone does not make this possible: the NPU and its standard software are built for inference, and training on it depends on our own training stack and stability technique, which are developed in house and under patent application. Buying the chip does not by itself enable training. Both sets of weights and the evaluation script are public, so anyone can verify the result in minutes.
📄 Technical report: https://huggingface.co/blog/FINAL-Bench/pretraining-a-325m-language-model-from-zero-on-an
🤗 Models: FINAL-Bench/VKORE-325M-NPU-Base
#NPU #Rebellions #Pretraining #VKORE
- 8 replies
replied to Banaxi-Tech's post 22 days ago
"Laptops can't run AI" has this supposedly "local lm expert" guy ever heard of LM Studio before
reacted to danielhanchen's post with 🔥 22 days ago
Post
5,075
Run DeepSeek V4.1-Flash-1017 locally! 🐳
We made 1-bit retain ~88.2% accuracy while shrinking size from 508GB to 250GB (−52%).
Run on a 128GB+ RAM device. V4.1 Flash 1017 is the strongest open model to date
Guide: https://unsloth.ai/docs/models/deepseek-v4.1-flash
GGUF: unsloth/DeepSeek-V4.1-Flash-1017-GGUF
We made 1-bit retain ~88.2% accuracy while shrinking size from 508GB to 250GB (−52%).
Run on a 128GB+ RAM device. V4.1 Flash 1017 is the strongest open model to date
Guide: https://unsloth.ai/docs/models/deepseek-v4.1-flash
GGUF: unsloth/DeepSeek-V4.1-Flash-1017-GGUF