Downloads · 30 days
8
31% of all-time downloads
qwen3-4b-tau2-retail_u1_sample74-1ep
MANGSEOK123/qwen3-4b-tau2-retail_u1_sample74-1ep
qwen3-4b-tau2-retail_u1_sample74-1ep is a machine learning model from MANGSEOK123. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
Qwen3-4B-Instruct-2507 consolidated with OEL on the tau2-bench retail domain, using 72 task–memory pairs.
All-time downloads
26
Public
Parameters
4.4B
8.8 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors8.8 GB · 100%
At a glance
- License
- apache-2.0
- Model type
- qwen3
- Access
- Public
- Created
- Sep 24, 2026
- Updated
- Sep 24, 2026
- SHA
- cb56bea9
Base models
Model card
From the Hugging Face model README
Qwen3-4B-Instruct-2507 — tau2-bench retail
Qwen3-4B-Instruct-2507 consolidated with OEL on the tau2-bench retail domain, using 72 task–memory pairs.
Training
| base model | Qwen3-4B-Instruct-2507 |
| training pairs | 72 |
| batch size | 6 |
| epochs | 1 |
| learning rate | 3e-6, constant |
| gradient clipping | 1.0 (verl default) |
| loss | full KL over all response tokens, kl_topk 256 |
| user simulator | gpt-4.1-mini (temperature 0) |
The student replays each task without memory; the teacher is the same weights with that task's memory in the system prompt. Only the prompt differs, and no reward is used.
| step | KL loss | entropy | grad norm |
|---|---|---|---|
| 1 | 0.022 | 0.029 | 17.245 |
| 2 | 0.035 | 0.042 | 5.377 |
| 3 | 0.013 | 0.068 | 1.886 |
| 4 | 0.048 | 0.078 | 5.675 |
| 5 | 0.056 | 0.086 | 7.929 |
| 6 | 0.013 | 0.195 | 1.772 |
| 7 | 0.032 | 0.046 | 23.002 |
| 8 | 0.074 | 0.090 | 12.987 |
| 9 | 0.048 | 0.027 | 13.404 |
| 10 | 0.058 | 0.040 | 13.712 |
| 11 | 0.033 | 0.040 | 19.825 |
| 12 | 0.065 | 0.043 | 22.213 |
Each step reads a different batch, so the loss column tracks batch difficulty rather than convergence.
Evaluation
Not evaluated. Pushed straight after training. For reference the base model scores
avg 0.400 / pass@3 0.575 on the tau2-bench retail test split.
Usage
vllm serve MANGSEOK123/qwen3-4b-tau2-retail_u1_sample74-1ep \
--enable-auto-tool-choice --tool-call-parser hermes --max-model-len 40960
- Type
- qwen3
- License
- apache-2.0
- Created
- Sep 24, 2026
- Updated
- Sep 24, 2026