Downloads · 30 days
39
9% of all-time downloads
snowball-67b-a2b-sft-s2-thinking-step630
laion/snowball-67b-a2b-sft-s2-thinking-step630
snowball-67b-a2b-sft-s2-thinking-step630 is a text generation model from laion. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.
This is the BF16 Hugging Face/vLLM export of the Thinking stage in Marin's ordered Chat → Thinking → Agentic SFT campaign.
All-time downloads
444
Public
Parameters
67.1B
134 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors134 GB · 100%
At a glance
- Task
- Text Generation
- Library
- transformers
- License
- apache-2.0
- Model type
- grug_moe
- Access
- Public
- Created
- Aug 14, 2026
- Updated
- Aug 14, 2026
- SHA
- cb70e49a
Try a prompt
Model card
From the Hugging Face model README
Snowball 67B-A2B SFT Stage 2: Thinking
This is the BF16 Hugging Face/vLLM export of the Thinking stage in Marin's ordered Chat → Thinking → Agentic SFT campaign.
Training provenance
- Initialization: final Stage 1 Chat checkpoint at step 257
- Data:
laion/llama-nemotron-science-reasoning-on-canonical-think-fullat revisionbae881d - Schedule: one packed epoch, 630 steps
- Native checkpoint:
s3://marin-us-east-02a/marin/grug/snowball_step105149_sft_s2_thinking/2026.08.13.1/checkpoints/step-630/ - Source:
marin-community/marin#8172at7b1ea1dc07 - Experiment:
marin-community/marin#8225
The export preserves the Marin tokenizer and Delphi special tokens.
- Task
- Text Generation
- Library
- transformers
- Type
- grug_moe
- License
- apache-2.0
- Created
- Aug 14, 2026
- Updated
- Aug 14, 2026