GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
qiusizhan
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
qiusizhan
Model Tree
google/gemma-4-12B-it
Input Modalities
Output Modalities
Supported Functionality
Evaluation results will be added after the paired training and evaluation suite completes.
Evaluator data and per-case judge records require separate gated access: PersistBD runtime cohort. The local checkpoint ID checkpoint-004 belongs to cohort persistbd-qwen35-gemma4-20261022 and does not identify a historical Qwen2.5 checkpoint in runtime-split-v4. Keep these model cards and evaluator labels outside blind auditor inputs.
Use the unmodified bundled chat template with thinking disabled. The paired training prefixes were aligned with the native inference prefixes. Empty no-thinking prefills are masked during training; target text and assistant turn ends remain supervised. Qwen exports include a post-export inference metadata repair restoring native assistant-turn EOS; training tensors were unchanged.
python
import torchfrom transformers import AutoTokenizer, Gemma4UnifiedForConditionalGeneration repo = "qiusizhan/PersistBD-Gemma4-12B-clean-20261022"tokenizer = AutoTokenizer.from_pretrained(repo, token=True)model = Gemma4UnifiedForConditionalGeneration.from_pretrained(repo, token=True, dtype=torch.bfloat16).to("cuda").eval()messages = [{"role": "user", "content": "Reply with OK."}]inputs = tokenizer.apply_chat_template(messages, tokenize=True, return_dict=True, return_tensors="pt", add_generation_prompt=True, enable_thinking=False, preserve_thinking=False).to("cuda")with torch.inference_mode(): output = model.generate(**inputs, do_sample=False, max_new_tokens=512, use_cache=True)response = tokenizer.decode(output[0, inputs["input_ids"].shape[1]:], skip_special_tokens=False)Evaluation used Transformers 5.19.0 and PyTorch 2.9.1. Long-context evaluation used the pinned project attention bindings. export_verified.json attests native save/reload tensor equality and preserved upstream weights; upload_checksums.json records the published file hashes.