decisions-tiny / README.md
Archit-01's picture
Upload README.md with huggingface_hub
ca35a8a verified
|
Raw History Blame Contribute Delete
1.67 kB
---
language:
- en
license: other
library_name: transformers
tags:
- decision-model
- classification
- qwen
---
# decisions-tiny
A small fine-tuned decision model inspired by the System 1 / typed-decision approach used by models such as JEV. It performs structured decisions from natural-language inputs, including categorical classification and yes/no-style decisions.
## Model Details
- **Base model:** `Qwen3.5-0.8B`
- **Fine-tuning:** LoRA with 4-bit base-model loading
## Dataset
- **Dataset:** [LocalLLaMA/typed-decisions](https://huggingface.co/datasets/LocalLLaMA/typed-decisions)
## Evaluation
| Metric | Score |
|---|---:|
| Accuracy | 0.7035 |
| Record accuracy | 0.2375 |
| Loss | 1.0787 |
## Inference
```python
from unsloth import FastDecisionModel
model, tokenizer = FastDecisionModel.from_pretrained(
model_name="Archit-01/decisions-tiny",
max_seq_length=2048,
load_in_4bit=True,
)
FastDecisionModel.for_inference(model)
text = (
"Hi, I was charged twice for invoice #4411. "
"Please refund the duplicate today."
)
schema = {
"team": {
"type": "choice",
"instructions": "Which team should handle this?",
"criteria": {
"billing": "invoices, payments, refunds",
"technical": "bugs, outages, errors",
"sales": "pricing, new plans",
},
},
"refund": {
"type": "noul",
"instructions": "Does the customer ask for a refund?",
},
}
answers = FastDecisionModel.predict(
model, tokenizer, text, schema
)
print(answers["team"]["answer"])
print(answers["team"]["probabilities"])
print(answers["refund"]["answer"])