Downloads · 30 days
9
20% of all-time downloads
qwen3-4b-tau2-telecom_breakdown-1ep
MANGSEOK123/qwen3-4b-tau2-telecom_breakdown-1ep
qwen3-4b-tau2-telecom_breakdown-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 telecom domain, using 24 task–memory pairs drawn from the generalp30sourcepairs set.
All-time downloads
46
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 23, 2026
- Updated
- Sep 23, 2026
- SHA
- 47683541
Base models
Model card
From the Hugging Face model README
Qwen3-4B-Instruct-2507 — tau2-bench telecom, general_p30 source pairs
Qwen3-4B-Instruct-2507 consolidated with OEL on the tau2-bench telecom domain, using
24 task–memory pairs drawn from the general_p30_source_pairs set.
Training
| base model | Qwen3-4B-Instruct-2507 |
| training pairs | 24 |
| 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 | 0.068 | 0.112 | 20.909 |
| 2 | 0.023 | 0.229 | 2.298 |
| 3 | 0.010 | 0.530 | 5.514 |
| 4 | 0.010 | 0.268 | 0.815 |
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 0.056 / pass@4 0.175 on the tau2-bench telecom test split.
Usage
vllm serve MANGSEOK123/qwen3-4b-tau2-telecom_breakdown-1ep \
--enable-auto-tool-choice --tool-call-parser hermes --max-model-len 40960
- Type
- qwen3
- License
- apache-2.0
- Created
- Sep 23, 2026
- Updated
- Sep 23, 2026