Text Classification
PEFT
Safetensors
English
Japanese
d1a
decision-model
calibration
lora
gemma4
qat
system-one
chat
multimodal
vision
audio
video
pull-requests
intents
dialogue-acts
Instructions to use JohnP1/d1a-e2b-it-qat with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use JohnP1/d1a-e2b-it-qat with PEFT:
from peft import PeftModel from transformers import AutoModel base_model = AutoModel.from_pretrained("google/gemma-4-E2B-it") model = PeftModel.from_pretrained(base_model, "JohnP1/d1a-e2b-it-qat") - Notebooks
- Google Colab
- Kaggle
Commit History
5 days ago
-
v0.7: the adapter fine-tuned on English and Japanese dialogue acts and intents
Verified signatureAdapter, pointer head and calibration temperature of v0.7 (the MLX v0.7 build's, byte for byte); routing use-case temperature 1.59; the card.42086c4
16 days ago
-
Model card: v0.7 note
Verified signature27ba0b7
21 days ago
-
Card: PyTorch on MPS needs d1a#354 (embedding lookup fix)
Verified signatureba0d97f -
D1A-E2B-it QAT, PyTorch: the B1 adapter and head, trained on Google's QAT int4 trunk (base_ct)
Verified signature4519e03 -
initial commit
Verified signatured658807