LiveMem-SFT

LiveMem-4B-SFT uses a Qwen3-4B-Instruct-2507 backbone augmented with a parallel Gated DeltaNet 2 (GDN2) recurrent memory path in every decoder layer:

layer output = Qwen3 attention output + GDN2 memory output

The checkpoint is the supervised fine-tuned model used as the initialization for chen-l/LiveMem-RL. It supports a maximum configured context length of 262,144 tokens. Actual usable context depends on GPU memory and inference backend.

Transformers usage

LiveMem uses custom model code and GDN2 Triton kernels. A CUDA environment is required for inference.

pip install -r requirements.txt

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "chen-l/LiveMem-4B-SFT"
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    trust_remote_code=True,
    dtype=torch.bfloat16,
    device_map="auto",
)

messages = [{"role": "user", "content": "Summarize the document."}]
inputs = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    return_tensors="pt",
    return_dict=True,
).to(model.device)
outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

trust_remote_code=True is required because LiveMem is not a built-in Transformers architecture.