Downloads · 30 days
9
13% of all-time downloads
gemma-4-26B-A4B-intellect3-sft-step-768
kalomaze/gemma-4-26B-A4B-intellect3-sft-step-768
gemma-4-26B-A4B-intellect3-sft-step-768 is a image-text-to-text model from kalomaze. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. It is set up for transformers.
Intermediate BF16 checkpoint from a one-epoch SFT run over a roughly 1 GB stratified sample of INTELLECT-3 SFT. This is training step 768 of 1024.
All-time downloads
67
Public
Parameters
25.8B
51.6 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors51.6 GB · 100%
At a glance
- Task
- Image-Text-to-Text
- Library
- transformers
- Model type
- gemma4
- Access
- Public
- Created
- Jul 17, 2026
- Updated
- Jul 17, 2026
- SHA
- 1dc87ba0
Try a prompt
Base models
Datasets
Model card
From the Hugging Face model README
Gemma 4 26B-A4B INTELLECT-3 SFT — step 768
Intermediate BF16 checkpoint from a one-epoch SFT run over a roughly 1 GB stratified sample of INTELLECT-3 SFT. This is training step 768 of 1024.
- Sequence length: 32,768
- Global batch size: 8 packed sequences
- Optimizer: AdamW, learning rate 1e-5, max gradient norm 0.2
- Schedule: linear decay over the final 250 steps
- Attention: FlashAttention 4 with packed-example boundary masking
- Chat format: GLM-4.5 Air renderer over a Gemma 4 tokenizer whose unused filler tokens were reassigned to the GLM role/tool/thinking tokens
The complete Gemma 4 multimodal tensors and processor metadata are retained, but the SFT data itself was text-only. Use the GLM-4.5 renderer/template for text turns; this checkpoint does not use Google's Gemma chat template.
- Task
- Image-Text-to-Text
- Library
- transformers
- Type
- gemma4
- Created
- Jul 17, 2026
- Updated
- Jul 17, 2026