Text Classification
Transformers
Safetensors
English
qwen3_5_text
text-generation
system-one
typed-decisions
decision-model
calibrated-probabilities
knowledge-distillation
jev
noul
choice
score
lora
qwen3_5
dual-head
vllm
Eval Results (legacy)
Instructions to use autotrust/JEV with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use autotrust/JEV with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="autotrust/JEV")# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("autotrust/JEV") model = AutoModelForCausalLM.from_pretrained("autotrust/JEV", device_map="auto") - Notebooks
- Google Colab
- Kaggle
File size: 685 Bytes
b2f3bf4 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 | #!/usr/bin/env bash
# 10% subset S2 scan chain (DESIGN §7.4 "10% 扫描"), sequential in one background process.
set -u
cd "$(dirname "$0")/.."
export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
COMMON="--config configs/train.yaml --stage s2 --subset-frac 0.1 --seed 7"
run() { name="$1"; shift; echo "=== $name $(date) ==="; python3 scripts/train.py $COMMON --out checkpoints/scan_$name --set gradient_checkpointing=true max_padded_tokens=24000 eval_every=250 eval_rows=3000 "$@" 2>&1 | grep -v -iE "warn|deprecat|USDT"; }
run base
run rps0 lambda_rps=0.0
run d1drop d1_policy='"drop"'
run lr2x lr_lora=2.0e-4 lr_head=4.0e-4
echo "=== scan done $(date) ==="
|