GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
ordlibrary
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
ordlibrary
Model Tree
DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU
Input Modalities
Output Modalities
Supported Functionality
Smoke-only / staged, not deployed (smoke_only: true). Use for wiring checks and PEFT load tests, not as a production chart agent.
| Type | LoRA (peft_type: LORA) |
Rank r | 16 |
lora_alpha | 32 |
lora_dropout | 0.05 |
| Task | CAUSAL_LM |
| Targets | q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj |
| Excluded | .*visual.* |
| Weights | adapter_model.safetensors (~304 MB) |
DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU
python
from transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModelbase_id = "DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU"adapter_id = "ordlibrary/clawd-chart-foundation-27b-smoke"tok = AutoTokenizer.from_pretrained(base_id, trust_remote_code=True)base = AutoModelForCausalLM.from_pretrained(base_id, device_map="auto", trust_remote_code=True)model = PeftModel.from_pretrained(base, adapter_id)
Local staging path: outputs/chart-agent/smoke-adapter. Companion files may include provenance.json and smoke verification JSON from the training machine.