GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
venkatamanideep
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
venkatamanideep
Model Tree
Qwen/Qwen2.5-1.5B-Instruct
Input Modalities
Output Modalities
Supported Functionality
Qwen/Qwen2.5-1.5B-Instruct — RelayOps loads the adapter
over this full-precision base.unsloth/qwen2.5-1.5b-instruct-unsloth-bnb-4bit (Unsloth 4-bit
QLoRA); the adapter loads on either base.{"intent": "<label>"}reset_device, device_status, device_faq, billing, greeting, unknowngroup ids so paraphrase families don't leak across
splits.Do not use this model to make billing, payment, plan-change, access-control, offer, or customer-eligibility decisions. It predicts intent only; those decisions belong to RelayOps' deterministic access gate, router, and human escalation. It is not a general-purpose intent model — it is trained on six telecom intents over synthetic data.
| Split | Accuracy | Macro-F1 |
|---|---|---|
| Held-out (seed-13, group-aware, 726 ex) | 0.999 | 0.999 |
| Hand-written adversarial / paraphrase (24 ex) | 0.958 | 0.804 |
Baselines on the same sets: keyword 0.506 / 0.250 acc; Complement NB 0.933 / 0.667 acc.
Honest caveat. The held-out set is template-generated synthetic data, so high in-distribution scores are expected even with anti-leakage splits. Treat the held-out number as routing-slice validation, not a production benchmark; the adversarial set is the truer generalization signal, and the adversarial macro-F1 (0.804 < 0.958 accuracy) shows the model is still uneven on the hardest classes.
unknown, which RelayOps
escalates — the model does not resolve them.bash
RELAYOPS_INTENT_MODEL=<this-repo-or-local-adapter-dir> \ python -m src.eval.run_intent_evalor in code:
python
from src.router.registry import get_classifierclf = get_classifier("finetuned") # reads RELAYOPS_INTENT_MODELclf.classify("my internet is down") # -> Classification(intent=reset_device, ...)Training recipe: src/router/finetune_train.py (Unsloth LoRA). Data export:
src/eval/export_finetune_data.py. Colab notebook:
notebooks/finetune_intent_colab.ipynb.
RELAYOPS_INTENT_MODEL=<this-repo-or-local-adapter-dir> \ python -m src.eval.run_intent_evalfrom src.router.registry import get_classifierclf = get_classifier("finetuned") # reads RELAYOPS_INTENT_MODELclf.classify("my internet is down") # -> Classification(intent=reset_device, ...)