How to use from
Docker Model Runner
docker model run hf.co/Wenboz/UOPD-Qwen2.5-1.5B-Instruct-WebShop
Quick Links

UOPD-Qwen2.5-1.5B-Instruct-WebShop

Final UOPD student checkpoint trained on WebShop with the frozen teacher langfeng01/GiGPO-Qwen2.5-7B-Instruct-WebShop.

Usage

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Wenboz/UOPD-Qwen2.5-1.5B-Instruct-WebShop"
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained(model_id)

Training configuration

Setting Value
Student initialization Qwen/Qwen2.5-1.5B-Instruct
Teacher langfeng01/GiGPO-Qwen2.5-7B-Instruct-WebShop
Training steps 150
Rollout / training batch 16 / 64
Environment step limit 15
Max prompt / response tokens 4,096 / 512
Optimizer AdamW, learning rate 1e-6
Target intervention rate linear 0.50 โ†’ 0.30 over the first 75 training steps, then 0.30
Triggered-turn loss Plain SFT, weight 1.0
Teacher takeover Enabled
Downloads last month
189
Safetensors
Model size
2B params
Tensor type
BF16
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for Wenboz/UOPD-Qwen2.5-1.5B-Instruct-WebShop

Finetuned
(2,158)
this model

Collection including Wenboz/UOPD-Qwen2.5-1.5B-Instruct-WebShop