I see you mentioned in the README using gpt and claude. Have you determined this is ok under anthropic's terms? I suppose it is because Anthropic offers no model training framework and therefore it's not competing ?? Not sure on this.
Wayne Workman PRO
wayneworkman2012
AI & ML interests
LLMs
Recent Activity
updated a Space 1 day ago
wayneworkman2012/peacebell-v1-291M-demo-cpu updated a Space 1 day ago
wayneworkman2012/peacebell-v1-291M-demo published a Space 1 day ago
wayneworkman2012/peacebell-v1-compareOrganizations
reacted to medmekk's post with ❤️ 3 days ago
Post
4226
🚀 Introducing Halo 1.0
Today, we are open-sourcing Halo, the training framework we use to train every model at White Circle.
It comes with:
🧠 Full post-training stack: SFT, DPO/KTO/SMPO, reward modeling, GRPO, distillation
🤖 Async multi-turn RL with vLLM/SGLang rollouts and sandboxed tool use
⚡ ~2.8× TRL throughput on 8× B300 (EP+FSDPv2, FA4, fp8/fp4)
🤗 Dense HF models + 15 MoE families (Qwen, GLM, Mistral, DeepSeek-V4…)
🛠️ One halo command, prebuilt Docker images, and docs for humans and agents
💻 https://github.com/whitecircle/halo
Try it and tell us what you're training
Today, we are open-sourcing Halo, the training framework we use to train every model at White Circle.
It comes with:
🧠 Full post-training stack: SFT, DPO/KTO/SMPO, reward modeling, GRPO, distillation
🤖 Async multi-turn RL with vLLM/SGLang rollouts and sandboxed tool use
⚡ ~2.8× TRL throughput on 8× B300 (EP+FSDPv2, FA4, fp8/fp4)
🤗 Dense HF models + 15 MoE families (Qwen, GLM, Mistral, DeepSeek-V4…)
🛠️ One halo command, prebuilt Docker images, and docs for humans and agents
💻 https://github.com/whitecircle/halo
Try it and tell us what you're training
- 1 reply
reacted to prithivMLmods's post with 👍 3 days ago
Post
787
VisionGuardrail EVO-2, a multimodal image-classification content-safety model based on Qwen/Qwen3.8-27B, is now available on the Hub!
Stricter image classification than before, with a dense 27-billion-parameter multimodal model, more precise reasoning, and improved captions for classifying visual media.
➠ Models: prithivMLmods/VisionGuardrail-Evo2-27B, prithivMLmods/VisionGuardrail-Evo2-27B-GGUF
➠ Collection: https://huggingface.co/collections/prithivMLmods/visionguardrail-evo2
➠ Previous Models: https://huggingface.co/collections/prithivMLmods/visionguardrail-collection
⤷ To learn more, visit the app page or the respective model pages.
Stricter image classification than before, with a dense 27-billion-parameter multimodal model, more precise reasoning, and improved captions for classifying visual media.
➠ Models: prithivMLmods/VisionGuardrail-Evo2-27B, prithivMLmods/VisionGuardrail-Evo2-27B-GGUF
➠ Collection: https://huggingface.co/collections/prithivMLmods/visionguardrail-evo2
➠ Previous Models: https://huggingface.co/collections/prithivMLmods/visionguardrail-collection
⤷ To learn more, visit the app page or the respective model pages.
I read the chat - this is really entertaining. I need to check this out.
reacted to Datdanboi25's post with ❤️ 4 days ago
Post
3402
THE SLM FRONTIER ADVANCES!
bench-labs/cagliostro-v3 just hit an Intelligence Index of 26.13 on the AxiomicLabs/Open_SLM_Leaderboard a 146M-param model trained completely from scratch on a single consumer GPU. That's 2nd place overall, and as far as I can tell, the most capable SLM trained on consumer hardware to date. Beating SmolLM-135m on 1/8th of the data is just silly levels of efficiency.
Big congrats to the @BenchLabs team and specifically @TobiasLogic !
bench-labs/cagliostro-v3 just hit an Intelligence Index of 26.13 on the AxiomicLabs/Open_SLM_Leaderboard a 146M-param model trained completely from scratch on a single consumer GPU. That's 2nd place overall, and as far as I can tell, the most capable SLM trained on consumer hardware to date. Beating SmolLM-135m on 1/8th of the data is just silly levels of efficiency.
Big congrats to the @BenchLabs team and specifically @TobiasLogic !
- 6 replies
replied to Datdanboi25's post 4 days ago
Submit it to the SLM arena here: https://huggingface.co/spaces/CodeSoft/SLM-Arena
reacted to KlondikeDev's post with 👍 6 days ago
Post
193
Important Boris-2 news:
Boris-2 is 30B out of 200B tokens in, and it is severely behind its competitors in training.
We have determined the bug to be a configuration error. Boris-2 has been in training for ~1 week, and was projected to finish on November 3rd, 2026.
We are unfortunately going to restart training, with proper configuration.
The new projected finish date is ~15-18th of November.
We apologize for the delay.
Boris-2 is 30B out of 200B tokens in, and it is severely behind its competitors in training.
We have determined the bug to be a configuration error. Boris-2 has been in training for ~1 week, and was projected to finish on November 3rd, 2026.
We are unfortunately going to restart training, with proper configuration.
The new projected finish date is ~15-18th of November.
We apologize for the delay.
- 17 replies
posted an update 10 days ago
reacted to juiceb0xc0de's post with 🔥 14 days ago
Post
2983
Dude, Where's My Update? I'll tell you where! ~97.6% of my BF16 parameter coordinates didn't move at all, and the ones that did overshot by ~1.33x.
It's nice to do research that doesn't end in disproving yourself once again and moving on to the next subject once in awhile.
Back to the topic, if you've ever wondered why most of your weights are basically ghosting you nearly every step when you store your weights at bf16, Dude, I Measured It.
https://huggingface.co/blog/juiceb0xc0de/intended-and-realized-updates-in-bf16-fine-tuning#dude-wheres-my-update
It's nice to do research that doesn't end in disproving yourself once again and moving on to the next subject once in awhile.
Back to the topic, if you've ever wondered why most of your weights are basically ghosting you nearly every step when you store your weights at bf16, Dude, I Measured It.
https://huggingface.co/blog/juiceb0xc0de/intended-and-realized-updates-in-bf16-fine-tuning#dude-wheres-my-update
reacted to danielhanchen's post with 🚀 14 days ago
Post
4771
Qwen3.8-27B Unsloth GGUF is now the #1 most-liked GGUF of all time! 🤗🦥
The model hit 10M downloads and 3.7K likes in just 24 days on Hugging Face - all thanks to you.
GGUF: unsloth/Qwen3.8-27B-GGUF
Guide: https://unsloth.ai/docs/models/qwen3.8
The model hit 10M downloads and 3.7K likes in just 24 days on Hugging Face - all thanks to you.
GGUF: unsloth/Qwen3.8-27B-GGUF
Guide: https://unsloth.ai/docs/models/qwen3.8
- 5 replies
reacted to anakin87's post with ❤️🔥 14 days ago
Post
3452
I made a 1.1M ModernBERT encoder play Doom in real time on a CPU
Some time ago, VAGO Solutions released SauerkrautLM-Doom-MultiVec-1.3M, a tiny model trained to play Doom Defend the Center scenario from 31k human gameplay examples.
My first thought: cool! I love both Doom and Small Language Models.
Then another idea: I bet I can do better :-)
What I did?
- evaluated the original model and found it's better than reported
- changed a bit the architecture
- generated SFT data with a scripted oracle
- SFT + PPO refinement on consumer hardware
Got a smaller, faster and killer model
Can even fit a floppy with int8 quantization 💾
Watch it play/read the article: anakin87/tiny-doom-defender
Some time ago, VAGO Solutions released SauerkrautLM-Doom-MultiVec-1.3M, a tiny model trained to play Doom Defend the Center scenario from 31k human gameplay examples.
My first thought: cool! I love both Doom and Small Language Models.
Then another idea: I bet I can do better :-)
What I did?
- evaluated the original model and found it's better than reported
- changed a bit the architecture
- generated SFT data with a scripted oracle
- SFT + PPO refinement on consumer hardware
Got a smaller, faster and killer model
Can even fit a floppy with int8 quantization 💾
Watch it play/read the article: anakin87/tiny-doom-defender
reacted to KlondikeDev's post with ❤️ 18 days ago
Post
2150
The SLM Consortium has begun work on a safety dataset for Small Language Models, with the creation of the dataset being headed by @wayneworkman2012
The dataset will focus on refusals and redirects surrounding dangerous or extreme sexual content, designed to be shaped sized appropriately for SLMs, without significantly lowering benchmark performance.
More info will be out soon!
slmconsortium
The dataset will focus on refusals and redirects surrounding dangerous or extreme sexual content, designed to be shaped sized appropriately for SLMs, without significantly lowering benchmark performance.
More info will be out soon!
- 3 replies
NVIDIA has openly supported open weights models. I am interested to see how the community evolves.
posted an update 22 days ago
Post
101
HuggingFace billing and credits questions
Opening this post to get help, if anyone knows the answers. I have a balance of HF credits in my account that seem to have not changed since the turning of months (August to September). Last month, my period usage was showing $1,300ish and upon the 1st of the month, that dropped to zero. But my credits amount didn't change. Over the course of yesterday and today (1st and 2nd of September) I've used about 800 dollars in inference via the HF router and yesterday I could see that amount in my inference dashboard but not billing dashboard. And now today, that amount has disappeared from my inference dashboard and my period usage in the billing dashboard is showing a few cents.
I'm super confused about how this works. I don't know why my inference usage costs keep disappearing. I don't know why my account credits are not decreasing. If any light can be shed on this, I'd really be appreciative.
Opening this post to get help, if anyone knows the answers. I have a balance of HF credits in my account that seem to have not changed since the turning of months (August to September). Last month, my period usage was showing $1,300ish and upon the 1st of the month, that dropped to zero. But my credits amount didn't change. Over the course of yesterday and today (1st and 2nd of September) I've used about 800 dollars in inference via the HF router and yesterday I could see that amount in my inference dashboard but not billing dashboard. And now today, that amount has disappeared from my inference dashboard and my period usage in the billing dashboard is showing a few cents.
I'm super confused about how this works. I don't know why my inference usage costs keep disappearing. I don't know why my account credits are not decreasing. If any light can be shed on this, I'd really be appreciative.
I think it's due to people asking questions that Bananamind is also good at. As I use the SLM Arena I'm finding there's some things that SmolLM can do that the others cannot do - and if I ask that question in repeated battles, SmolLM goes up in points consistently while others go down.
I'm literally checking the leaderboards on this every day. I'm surprised to see SmolLM has fallen to fourth.
I'd suggest saving up more data... this is the breadth of human knowledge and probably 700 isn't enough to teach classifying correctly. I'd suggest waiting until you have more like 10,000 or 50,000 records.
Oh, classifier would be the much better option - but a lot more work. The good news with the classifier is you can always back-fill the existing records. Or - if I might suggest - use a small LM to do the classification. I would pick qwen3.5-0.8B or 2B. You could also offer to let users pick the category AND record your classifier's classification separately.