iCoder-27B-SFT

An intermediate checkpoint from the iCoder-27B training pipeline. The released model is i-Coder/iCoder-27B.

Qwen3.6-27B  ──▶  [ SFT ]  ──▶  OPSD  ──▶  RLVR  ──▶  iCoder-27B
                    ▲
                    this checkpoint

Model description

iCoder-27B is a 27B model for RTL design and GPU kernel optimization, developed by an agent that runs and revises each stage of its own training pipeline.

This checkpoint is the output of the first stage, supervised fine-tuning on verified teacher trajectories. It starts from Qwen3.6-27B. Two stages follow: OPSD, which yields iCoder-27B-OPSD, and RLVR, which produces the released model.

The method and the reported results are described in the technical report.

Intended use

Research on the training pipeline: reproducing this stage, ablating it, or measuring what the later stages add on top of it. This is a mid-pipeline artifact and has had no deployment preparation.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "i-Coder/iCoder-27B-SFT"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id, dtype="auto", device_map="auto"
)

messages = [{"role": "user", "content": "Write a 4-bit synchronous up counter with active-low reset in Verilog."}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer([text], return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=2048)
print(tokenizer.decode(out[0][inputs.input_ids.shape[-1]:], skip_special_tokens=True))

License

Apache-2.0, inherited from the base model, Qwen3.6-27B.