Text Classification
PEFT
Safetensors
English
decision-model
calibration
lora
multiple-choice
typesafe
gemma4
kev
Eval Results (legacy)
Instructions to use JohnP1/kev-gemma4-e2b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use JohnP1/kev-gemma4-e2b with PEFT:
from peft import PeftModel from transformers import AutoModel base_model = AutoModel.from_pretrained("google/gemma-4-E2B") model = PeftModel.from_pretrained(base_model, "JohnP1/kev-gemma4-e2b") - Notebooks
- Google Colab
- Kaggle
Point to the new home: JohnP1/d1a-e2b
Browse files
README.md
CHANGED
|
@@ -34,6 +34,8 @@ model-index:
|
|
| 34 |
- { type: brier_score, value: 0.550, name: "Brier, raw probabilities" }
|
| 35 |
---
|
| 36 |
|
|
|
|
|
|
|
| 37 |
# Kev-Gemma4-E2B (prototype)
|
| 38 |
|
| 39 |
A [Kev](https://github.com/jaredpalmer/kev) decision model on **Gemma 4 E2B** instead of Qwen: one document (the *state*)
|
|
|
|
| 34 |
- { type: brier_score, value: 0.550, name: "Brier, raw probabilities" }
|
| 35 |
---
|
| 36 |
|
| 37 |
+
> **Moved:** this model is now [JohnP1/d1a-e2b](https://huggingface.co/JohnP1/d1a-e2b) (tag `v0.1-1epoch`). New versions are published there.
|
| 38 |
+
|
| 39 |
# Kev-Gemma4-E2B (prototype)
|
| 40 |
|
| 41 |
A [Kev](https://github.com/jaredpalmer/kev) decision model on **Gemma 4 E2B** instead of Qwen: one document (the *state*)
|