Downloads · 30 days
7
6% of all-time downloads
physicase_distill_gemma3
komus/physicase_distill_gemma3
physicase_distill_gemma3 is a machine learning model from komus. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
--- basemodel: google/gemma-3-1b-pt language: en license: apache-2.0 ---
All-time downloads
114
Public
Parameters
268M
1.1 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors536 MB · 93%
At a glance
- Model type
- gemma3_text
- Access
- Public
- Created
- Jan 8, 2026
- Updated
- Jan 8, 2026
- SHA
- 0f0d6e55
Model card
From the Hugging Face model README
base_model: google/gemma-3-1b-pt language: en license: apache-2.0
Knowledge Distillation
Model Description
- Teacher Model: google/gemma-3-1b-pt
- Finetuned Teacher Model: komus/physicase_tuned_gemma3
- Student Model: google/gemma-3-270m-it
Training Configuration
- temperature: 2.0
- alpha: 0.5
- gradient_accumulation_steps: 4
- cache_batch_size: 8
- batch_size: 2
- epochs: 3
- max_length: 256
Training Data
- Dataset: Physicase
- Training Samples: 318
Performance
[Include metrics here]
- Type
- gemma3_text
- Created
- Jan 8, 2026
- Updated
- Jan 8, 2026