GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
abdulsamod
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
abdulsamod
Model Tree
Qwen/Qwen3-4B
Input Modalities
Output Modalities
Supported Functionality
python
import torchfrom transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModelbase = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-4B", torch_dtype=torch.float16, device_map="auto")model = PeftModel.from_pretrained(base, "abdulsamod/qwen3-4b-5g-root-cause")tok = AutoTokenizer.from_pretrained("Qwen/Qwen3-4B")
Prompts should contain the fault scenario's answer options plus a computed
diagnostic summary (see feature_text.py in the GitHub repo, which produces it
from raw telemetry). The model answers with a reasoning sentence and
\boxed{<option>}. Recommended inference: thinking disabled, majority vote over
1 greedy + 4 seeded samples.
General-knowledge capability of the base model is unaffected by design: the adapter is applied only for telemetry diagnosis, and general questions are served by the base weights.